Chapter 3: Selection and Installation of Control and Monitoring Points

3.1. Introduction. 2

3.2. Point Selection. 2

3.2.1. Temperature Between Each Cooling or Heating Element3

3.2.2. Motor Command Signal and the Motor Status. 3

3.2.3. Control Loop Execution Over the Network. 5

3.2.4. Network Card to Obtain Monitoring Points. 6

3.2.5. Monitoring and Diagnostic Points. 6

3.2.6. Additional Points for Advanced Control Strategies. 8

3.3. Sensor Accuracy. 8

3.3.1. Sensor Error9

3.3.2. Sensor Wiring Errors. 9

3.3.3. Transmitter Error13

3.3.4. Controller Error13

3.3.5. Network Architecture Effects. 16

3.3.6. Workstation Effects. 16

3.3.7. Installation Errors. 16

3.3.8. Persistence of Accuracy. 17

3.4. Installation Guidelines. 18

3.4.1. Safety Points. 18

3.4.2. Manual Override. 26

3.4.3. Calibration Test Ports. 27

3.4.4. Calibrating Analog Inputs. 27

3.4.5. Calibrating Analog Outputs. 29

3.4.6. Calibrating Binary Inputs and Outputs. 30

3.5. Sensor Selection and Installation Guidelines. 30

3.5.1. Temperature. 30

3.5.2. Humidity. 35

3.5.3. Pressure and Flow.. 36

3.5.4. Electrical44

3.6. Point Structure and Interface at the BAS.. 45

3.6.1. Point Naming Conventions. 45

3.6.2. Virtual Points. 47

3.6.3. BAS Settings. 48

3.6.4. Programmable Alarms. 48

3.6.5. BAS Graphical User Interface. 49

3.6.6. Point Trending. 51

3.6.7. System Back-ups. 51

 

Figures

Figure 3.1 Resistor Terminal Block. 14

Figure 3.2 Duct and Damper Damage due to Air Hammer 20

Figure 3.3 Limit Switch Installation for a Permissive interlock. 22

Figure 3.4 Typical Multi-speed Motor Interlocks. 24

Figure 3.5 Typical Function Module. 33

Figure 3.6 Typical 5-Valve Manifold and Schematic. 41

Figure 3.7 Pressure Gauge Snubber Styles. 42

Figure 3.8 Flow Sensor Designed for the Fan Wheel Inlet 43

 

Tables

Table 3.1 Temperature Measurement Technologies. 31

Table 3.2 Humidity Measurement Technologies. 35

Table 3.3 Pressure Measurement Technologies. 37

Table 3.4 Differential Pressure-based Flow Measurement Technologies. 38

Table 3.5 Non DP-based Flow Measurement Technologies. 39

Table 3.6 Electrical Measurement Technologies. 44

 

3.1. Introduction

Selecting control system points not only requires knowledge of the system design intent, but an understanding of how the system will be commissioned and operated for the life of the building. Including the commissioning and building operations perspective when selecting control and monitoring points will go a long way toward achieving a more easily operated building. This chapter draws upon commissioning field experience to help designers, commissioning providers, and contractors install appropriate control system points for well-functioning HVAC systems.

Chapter 3 is organized from the general to the specific. Recommendations for selecting input (i.e., sensors) and output points (i.e., commands to actuators) are presented along with practical advice for installation. Understanding sources of sensor error, from the sensor to the control system workstation, helps designers avoid these errors and helps commissioning providers interpret inaccuracies in the field measurements. Additional issues such as safety points and calibration guide installation. Next, sensor selection and installation guidelines identify sensor requirements based on application. For example, the temperature sensor requirements can often be very different for space temperature and chilled water supply temperature applications. With a limited budget for control and monitoring points, the sensor selection guidelines steer efforts toward high-impact control and monitoring decisions. Recommendations for interfacing points to the building automation system (BAS) include discussion of the details central to making the BAS workstation a useful tool for operators.

Specific issues regarding damper selection and control are covered in the Functional Testing Guide for Air Handling Systems (Fuctional Testing Guide), Chapter 5: Economizer and Mixed Air, Section 5.6.1 Dampers. This chapter is also the educational component behind the point lists in Chapter 5: System Configurations.

3.2. Point Selection

Through careful control and monitoring point selection, designers can enable efficient control, measure system performance, and perform diagnostics. This section gives tips for selecting points that result in more easily controlled and maintained HVAC systems.

3.2.1. Temperature Between Each Cooling or Heating Element

While all temperature points may not be critical for basic control, having temperature measurements between each cooling or heating element helps diagnose operational problems such as leaking control valves, simultaneous heating and cooling, and poorly calibrated sensors.

In the following example, adding a preheat temperature sensor illustrates the usefulness of points for diagnostics. Consider the case in which a designer includes a hot water coil in the HVAC system purely for a warm-up cycle since the outdoor air requirements did not result in a need to preheat. As a result, the control sequence simply controls this hot water coil based on a space temperature during the warm-up cycle then shuts the coil down at all other times. A coil discharge air temperature sensor is not included because it is not need for control. However, including the sensor as a part of the base design would provide the following immediate and long-term benefits.

·       The commissioning provider could trend the point during the commissioning process to verify that coil was not being activated inappropriately.

·       If functional testing revealed that the performance of the control loop under the warm-up cycle was unsatisfactory based on discharge temperature due to system time constants or other unanticipated issues, the sensor in the coil discharge would provide an easily implemented alternative control loop input.

·       The discharge temperature sensor can indicate a temperature rise across the preheat coil when the preheat valve should be closed. Operators could be notified of this problem through a BAS alarm that evaluates preheat discharge temperature, outside air temperature, and outside air damper position. Alarms that involve comparison or calculations are often called “smart alarms”, which are described in more detail in Section 3.6.4 Programmable Alarms.

·       If a change in the owner’s use of the space occurred that affected the census in the space, the systems minimum outdoor air requirements could be increased. The coil might need to perform some modest preheat function during extreme weather. Having an existing sensor in the coil discharge could be used in the preheat control loop without hardware additions and their associated wiring and sensor costs.

·       Instead of fully opening the hot water valve for freeze protection, the preheat discharge temperature sensor can be used to modulate the hot water valve to keep the air handling unit above freezing when the unit is not operating. Fully opening the hot water and chilled water valves for freeze protection has adverse effects, which are described in the Functional Testing Guide in Section 9.5.4.1 Freezestat Control Sequences.

3.2.2. Motor Command Signal and the Motor Status

The digital output that commands a motor to start and stop and the digital input that shows a motor’s on/off status should both be available at the operator’s interface to the control system. The digital input status point is not required to operate the system, but is useful as a troubleshooting tool because it proves that the motor has correctly responded to the command by the digital output. If the value of the command point does not correlate with the value of the status point (for instance the fan is commanded on, but the status point says that the fan is not running) then the drive or motor starter may have been switched to a manual mode. For VFD diagnostics, the motor speed command and speed feedback from the motor can be compared. Smart alarms can alert the operator when the motor speed does not vary over a significant period of time, signaling a potential problem with the VFD. Operator responses and smart alarms need to take into account that there is often a time delay between when a command is sent and when the workstation displays the command or feedback.

To pick up the proof of operation information, a number of techniques can be used. Contrast the following options for a status point for a supply fan.

·       Option 1: Motor Starter Auxiliary Contact Monitoring the motor starter auxiliary contact is probably the least expensive approach to obtaining a motor status point, but it is only an indirect indicator of motor status. The point does not prove that the fan is running, but only that the starter engaged. If the belt or drive coupling to the fan has failed, this failure would go undetected until some other parameter (like loss of control of space temperature) became evident. Even then, the starter auxiliary contact would not be a direct indication of a drive failure since there are often other things that could cause problems.

·       Option 2: Fan Differential Pressure Switch Monitoring the fan differential pressure switch is slightly more expensive than Option 1, but it has the advantage of proving that the fan wheel is actually rotating, thus the drive system is intact. However, this point does not necessarily prove that the fan is moving air since a fan will produce pressure without necessarily producing flow. If the fire dampers in the system closed or the smoke isolation dampers or intake dampers did not open, the fan could start and produce a pressure difference which would be seen by the differential pressure switch and interpreted as proving the fan’s operation. In this case, the fan is operating, but the intended system function of moving air is not being provided.

·       Option 3: Differential Pressure Switch Across Several AHU Elements If the differential pressure switch associated with Option 2 is piped across several AHU components that only produce a pressure drop if there is flow, then the switch will not only prove that the fan has started, but will also prove that there is flow in the system. Adjustment of this switch can be critical on VAV systems since the pressure drop will vary with the square of the flow rate. For instance, when the system unloads and reduces its flow rate to 50% of design flow, the pressure drop through the air handling unit components will only be 25% of the design value. If the switch was only piped across one AHU element, the pressure drop at low loads could be difficult to detect reliably and could result in the switch erroneously indicating that the supply fan was not running. Piping the switch across several elements provides a larger signal to work with.

·       Option 4: Motor Current Switch Motor current switches provide a fan proof of operation point similar Option 3 since the motor current will also vary with load. The first cost will be lower than the differential pressure switch because the current switch can be located in the starter. This location also reduces wiring costs compared to the differential pressure switch because the switch is installed at the same location as the start-stop command and can be picked up by running a second cable to this location.

Properly adjusting these switches on VAV systems can offer similar challenges to those mentioned in Option 3. At 50% of design flow, the horsepower requirement is 1/8 of the design value, since the motor load varies as the cube of flow. A low current at a low load condition should not be mistaken for no supply fan operation. In addition, other parameters that vary with the load also affect motor current, like motor efficiency and power factor. As a result, finding the correct setting for a current switch on a VAV system can take some experimentation. Often, it is easiest to determine the correct current switch setting during the belt re-tensioning effort[1] at start-up. The motor current should be measured at no load with the belts off and at minimum turn-down. Then adjust the switch to prove operation only when the current is exceeds this value. This process can be time consuming and may not be worth the expense on smaller systems.

3.2.3. Control Loop Execution Over the Network

The input to and output from a single control loop should be hardwired to the controller where the loop statement resides. As a general rule, a control loop should not be executed over a communications network.

Consider this scenario for a fan discharge static pressure control loop. The input duct static pressure is wired to a remote terminal unit and sent to the air handler controller over the communications network. The output of the PID loop at the air handler controller is then sent over the network to control a VSD. What would happen to system control if the network crashed? A network failure puts the system in an open loop mode. Information would not be available to the operators, and the system would be controlled by either the last data that the controller received or a default value. To avoid these problems, all inputs and outputs for a control loop should be hardwired to the controller that performs the loop statement.

Another reason to avoid sending control loop inputs and outputs over the network is to prevent the network communication rate from interfering with the control loop. When a network slows down due to network traffic, tuning control loops becomes extremely difficult, since the time constant for loop response becomes inconsistent. Evidence of variable network performance may be apparent in the variable time required to update workstation graphics.[2]

Consider a loop that controls the building relief dampers where the relief damper command resides in the controller, but the loop must run at a network controller because the application specific controller does not support building pressure based control of the relief dampers. Upon a network communications failure, if the system was configured to alarm and hold the relief dampers at the last setting, then the approach may be acceptable. On the other hand, if the loop that resides on the network controlled the supply fan speed based on duct static pressure, an open loop situation due to a network failure could place the duct system in danger of over-pressurization and an inability to meet the load. In this case, an application specific controller (with the control loop executed over the network) may not be an acceptable alternative for the system.

In older systems, the loop statement is often located at the host computer, since older controllers lack the computing power to perform loop calculations. As a result, both the input and the output data for the loop is transmitted over the network. In newer systems, the control loop can reside at the air handler controller, with the input hardwired to the controller and the output hardwired to the actuator or drive. And on some current technology networks, the loop can reside in the VFD drive controller. If this approach is used, then the control loop that controls the drive should be located in the drive controller card, and the input that is used by the controller should be wired to the drive controller card where it can be used directly by the control loop without network interaction.

Sending control variables over the network may be inevitable with the use of application specific controllers with control algorithms that are not available pre-programmed into the controller (see sidebar). In this case, the control logic must reside on a higher-level network controller, with the controlled variable sent over the network. This approach should be used with caution and evaluated with regard to what happens if the network were to fail. In some cases, the result would not be critical to building performance or system integrity and could be tolerated as long as the failure was detected and corrected.

Sending interlock commands over a network should be considered on a case-by-case basis. In some instances, using software interlocks can provide economy in terms of first cost by substituting programming for hard wiring with little loss in system integrity, performance or safety. But, in other cases, a network problem could impose severe operating and safety penalties on the system if the interlocks are accomplished in software. In these situations, hardwired interlocks should be implemented.

3.2.4. Network Card to Obtain Monitoring Points

For a piece of equipment with many output parameters, like a VSD, fire alarm panel, or chiller panel, using a network card that interfaces with the control system is an economical way to obtain operational information.

While the input to and output from a single control loop should be hardwired to the controller where the loop statement resides, other points such as the start/stop command and proof of status feedback can sent across the network since they are generally not continuously varying processes and are not part of the control loop.

For example, the following VSD points should be available through the network card:

·       Proof of Operation

·       Speed feedback This point is not absolutely mandatory, but is highly desirable to ensure that the drive is responding appropriately

·       Selector switch status contact (hand/auto/off, and inverter/bypass) This point is not absolutely mandatory, but also highly desirable. This point allows an alarm when the piece of equipment has been put into a mode where the drive is not functioning. One point with a general alarm can be used for to monitor both switches.

For the cost of a network card and twisted shielded pair, the numerous parameters that go along with the drive are available, including start/stop, feedback proof of speed, status contact, current, kW, faults (programmable), and diagnostic points. Since these points can be obtained by the network card instead of hardwiring, more controller I/O capacity is available for other points. A network card can access many points and drive parameters for less money than would be spent hardwiring. The catch is that the network card must allow the VSD and control system to communicate. The communications card, VSD, and control system should be compatible with a common network connection protocol such as LONworks, BACNet, or IEEE standards. As an alternative, many drive manufacturers offer special drivers or cards to interface their equipment to the proprietary communications protocols used by the various control systems manufacturers.

Similar considerations apply when networking other equipment such as chiller or boiler control panels, lighting control panels, security or access control panels, or fire alarm panels. Fire alarm and security panels may require additional considerations related to maintaining the integrity and security of the system. These considerations may make a network interface to these devices impossible, impractical, or undesirable despite the technical benefits that could be achieved.

3.2.5. Monitoring and Diagnostic Points

Monitoring and diagnostic points serve three main purposes: to understand system performance, to add flexibility to the system, and as a tool for tracking maintenance needs. Monitoring points are often utilized for manual or automated system diagnostic methods to detect problems. Smart alarms can notify the operators of these problems (see Section 3.6.4 Programmable Alarms for more on smart alarms).

Examples of monitoring and diagnostic points include: fan amps, total HVAC power, supply fan flow, cooling coil capacity, and ton-hours. A full list of recommended monitoring and diagnostic points is included with the points lists in Chapter 5. Generally, it will be less expensive to install monitoring points during the initial construction work rather than later, and the costs can be carried by the project’s construction budget. If the point must be added during the start-up phase in response to a commissioning need, the operations staff may need to find alternative funding outside of the project budget, which is not always an easy task.

Systems will last longer than the current design conditions will exist, so it can be beneficial to design systems to be adaptable to changes in operation. Due to overestimations in design loads, most systems have some margin to handle load changes that may occur over time. The trick is designing a system that can adequately serve the peaks and future load requirements while efficiently serving the present and future load conditions.

Monitoring points also provide valuable information for maintenance. For example, points that notify the operator of a high filter pressure drop are especially important for detecting dirty filters in VAV systems since the filter pressure drop varies with flow. If visual monitoring is done, the filter pressure could look normal in the morning at low flow conditions when the operator made his rounds, but have pressure drops in excess of the filter structural rating at peak load in the afternoon when nobody was there to notice it. As a result of excessive pressure drops, some filters could fail. Air would bypass the filter banks and lower the net pressure drop on the filter bank. In this condition, the filters could be mistaken to be clean if the filters were not visible. These problems could be avoided with continuous monitoring of filter pressure drop at the workstation. More examples of monitoring points for maintenance are listed below:

·       Accumulated run hours for pieces of equipment to help schedule preventive maintenance

·       Data to calculate the operating cost for a system

·       Proof of operation input current switches to alert maintenance staff to motor problems. The current switch should be able to tell the difference between a fan running at low load and a broken belt or coupling.

·       High and low water level alarms for evaporative coolers to indicate loss of make-up water and overflow conditions.

·       Make-up and blow down rate for evaporative systems. Owners can to reduce their sewer charges by the amount of water that was evaporated in utility districts where the sewer charge is based directly on the water consumption.

·       Pressure relationships in critical applications like healthcare and clean rooms. While this monitoring can be accomplished manually via testing with hand-held equipment, continuous monitoring will catch a problem sooner and frees staff up for othertasks.

Be careful not to overdo it – adding monitoring and diagnostic points is not always cost effective. When prioritizing the importance of monitoring points, think about how often the data will be used, what the data will be used for, and the implications of detecting problems.

Advanced control strategies such as temperature or pressure reset can save energy and improve overall system performance. However, they must be applied with caution, considering the needs of specific the system. A “one size fits all” application of advanced strategies leads to problems:

Reset strategies combined with PI or PID loops can make systems extremely unstable. If the setpoint is continuously optimized based on other system parameters, then adding Integral control to eliminate proportional error may not be necessary since the reset routine may correct for the proportional error automatically. Making the loop Proportional-only can lower first cost and make the system easier to operate and maintain.

Discharge temperature reset routines can make a variable volume system act like a constant volume reheat system if not properly implemented. Resetting to a higher leaving air temperature can cause critical zones to demand more flow, resulting in a reduction of the controlled static pressure, and a corresponding demand for the fan speed to increase, in some cases to 100% speed. Sophisticated control strategies applied as “canned” routines can increase operating costs and reduce efficiency – the opposite of their intended effect.

3.2.6. Additional Points for Advanced Control Strategies

Additional points may be necessary to implement advanced control strategies. Examples of advanced control strategies associated with air handling equipment include building pressure control and demand control ventilation (based on occupancy status of room or CO2 levels).

Some buildings are beginning to employ strategies that limit demand during peak times. These strategies may require monitoring of end-use power (to understand where demand reductions have occurred), zone temperatures (to measure the impact on occupant comfort), and zone air flows (to verify that indoor air quality is maintained). In some cases, these demand reduction strategies are being implemented to engage automatically during high prices or requests by the utility for demand reduction. The interface of the control system with the pricing information is not an easy task.

3.3. Sensor Accuracy

Specifying the desired accuracy of sensors is not as straightforward as it sounds. Sensor accuracy should be specified with an understanding of the different ways in which manufacturers can present the accuracy as well as with an understanding of the requirements of the sensor application. The following information about sensor accuracy describes potential sources of error in readings and how accuracy can degrade over time. There is potential for error at each step of a control or monitoring measurement – at the sensor, transmitter, wiring, controller, and operator workstation.

Consider all the steps to read an accurate temperature measurement from an RTD with a transmitter. First, the RTD must accurately vary its resistance in response to temperature. The sensor resistance is converted to a voltage, and this voltage is converted to a 4-20 mA current signal by the transmitter. Then, the current travels down a length of wire to the controller where it is converted back into a voltage signal, typically by using a scaling resistor. The controller routes the signal through an analog-to-digital converter with a certain resolution (8- bit, 12-bit, etc.) and sends the digital signal via a network protocol to the host computer. The analog-to-digital converter and all of the other electronics on the circuit board have tolerances associated with them that can impact the signal quality. The way the data is handled and truncated to send it from the controller to the workstation is also important to prevent minor changes from creating a flurry of activity on the communications network. The bottom line is that at each mechanical or electronic interface, there are a number of opportunities for error.

A good understanding of sensor accuracy issues serves both designers and commissioning providers well. By understanding the many ways accuracy can be affected, designers can write clear specifications for sensor selection and installation that result in reliable measurements. Commissioning providers inevitably find erroneous data in the field, and an understanding of measurement error gives valuable insight into troubleshooting these errors. When interpreting test results, commissioning providers should keep in mind that the data may not reflect the true conditions for a variety of reasons related to sensor accuracy and the data handling characteristics of the system. Going one step further, the commissioning provider should test the systems with these possibilities for error in mind. The sources of measurement error presented in this section, from the sensor to the DDC workstation, will help users better understand the complexity of sensor measurements in control systems.

3.3.1. Sensor Error

Both the RTD element and the transmitter electronics package have accuracy issues related to a number of factors. Sensors are manufactured with a given measurement tolerance and sensitivity. Some sensor measurement errors are due to a lack of persistence of calibration, while other errors are intrinsic to the sensor and the way in which it was installed.

Precision field calibration of sensors can be difficult, time consuming, and costly. One way to solve this problem is to have the manufacturer supply calibration certificates for each sensor that demonstrate that the sensor has been appropriately calibrated at the factory. Section 3.4.4.1 gives more information on using factory-calibrated sensors.

Sensor accuracy takes into account all deviations between a measured value and the actual value. Three common ways of measuring sensor accuracy are:

·       Percent of a sensor span (i.e., ±0.5% of the 0°F -100°F)

·       Percent of sensor reading (i.e., ±0.5% of the 50°F reading)

·       Absolute accuracy (i.e., ±0.5°F).

For example, ±0.5% of the reading for a 0°F -100°F transmitter is more accurate than ±0.5% of span. An accuracy of ±0.5% of reading is ±0.175ºF at 35ºF and ±0.475ºF at 95ºF, while an accuracy of ±0.5% of span is ±0.5ºF at all temperatures. By reducing the span, accuracy can be increased. Sensor accuracy can be specified for a sensor alone, a sensor-transmitter assembly, or all inclusively from the sensor to the workstation. If a 100W platinum RTD with a 4-20 mA transmitter was rated for ±0.5°F as an assembly, then you should expect that any sensor and transmitter manufactured to that specification will have an accuracy of at least ±0.5°F.

A sensor’s accuracy may be affected by:

·       Non-linearity

·       Hysteresis (the difference in a measured value when approached from above and below).

·       Mounting location effects (heat and vibration)

·       Thermal drift (thermal cycles that degrade electronics accuracy over time).

·       Transmitter components calibration errors related to their tolerances

·       Self-heating of resistive sensing elements. The measuring current required for the sensor output signal heats the element itself. The error depends on the heat-shedding properties of the sensor’s materials, construction, and the temperature of the environment. (Reference: See Minco Resistance Thermometry, Application Aid #18.)

3.3.2. Sensor Wiring Errors

The way in which a sensor is wired to the controller can also reduce accuracy. Overcoming lead wire resistance using current loops can be a more robust method for getting information from the field device to the controller than measuring voltage (with little current flow in the input wires). Properly shielding and twisting the input wire results in more reliable measurements by minimizing the potential for electrical noise.

3.3.2.1. Lead wire resistance

The added resistance from the length of wire connecting the sensor to the controller is called lead wire resistance. Lead wire resistance results in errors for resistance type sensors. The higher the resistance of the RTD sensor, the less the lead wire length affects the RTD accuracy, since the lead wire resistance itself accounts for a smaller fraction of the total resistance.

There are three components to the lead wire resistance problem. One is the resistance added to the circuit at any fixed temperature simply due to the length of the leads. This effect can easily add a degree or more of equivalent resistance to the circuit that has nothing to do with the measured value. Given time and sufficient instrumentation, this error could be field calibrated out of the circuit using scaling factors in the system software since it is purely a function of the physical length and gauge of the wiring run.

The other components of the lead resistance problem are much harder to deal with. Since the resistance of most metals varies with temperature, (and therefore the metals are used to measure temperature) the resistance of the leads will vary with temperature. A variable error will be added or subtracted from the input resistance as the ambient temperature changes. In equipment rooms that see large temperature swings, several tenths of a degree to a degree or more of variability can be introduced into the measurement. The quality of the sensing connections is the third factor in the overall input resistance equation. Poor or loose connections will introduce inaccuracies that are difficult to predict, vary with temperature and degrade over time.

The following calculation illustrates the impact of lead resistance problems for a common sensing technology often applied without lead length compensation or a transmitter.                      

The specification for an automation system to optimize the performance of multiple chiller plants on a onepipe chilled water loop called for 100W platinum RTDs installed for an overall accuracy of ±1/10°F — a very tight spec that will not be achieved without a transmitter with a limited range or using lead length compensation. Instead of 100W platinum RTDs, 1000W copper RTDs without lead length compensation had been installed. The impact of the lead resistance made it impossible to meet the accuracy in the specifications by an order of magnitude. As a result, the system could not make needed temperature-based decisions about starting and stopping the chillers, making the incapable of meeting its design intent.

Effect of length of lead wire on RTD measurements

·       Distance to the sensor - 100 ft.

·       Wire size - 22 AWG

·       Specific resistance - 0.0165ohms per foot at 25°C

·       Total lead length in series with the RTD – 200 ft.

The added resistance from the length of wire is 3.3 ohms. With an average RTF sensitivity of 4.7880 ohms per°C, the equivalent temperature associated with lead resistance is 1.24 °F.

Effect of temperature change on RTD measurements on a rooftop unit where the conduit is run outdoors in the Midwest

·       Minimum temperature - minus 20°F

·       Maximum temperature - 105°F (assuming no solar effects)

·       Temperature change - 125°F or 69°C

·       Resistance temperature coefficient for copper - 0.0043 ohms per ohm per °C

For the temperature variability stated, the corresponding change in RTD resistance is 0.98 ohms. This change in resistance translates to a 0.37°F change in temperature due to the outdoor temperature.

Overcoming Lead Wire Resistance

Two technologies help overcome susceptibility to lead wire resistance error. Lead length compensating wiring configurations can be used, or transmitters can be installed near the sensing element.

There are several lead length compensating wiring configurations, but three and four wire circuits are the most common. Four wire circuits provide the most accuracy. In general, these four wire circuits use one pair of wires to carry the current that excites the resistive sensor. This current creates the voltage drop that is measured by a high impedance voltage measuring device with the remaining two wires. The voltage measuring device accurately measures the voltage drop across the resistor without lead length effects because it draws very little current.

Another approach to minimizing the lead wire resistance effect is to use a transmitter. The transmitter converts the low level input into a higher level signal (1-5V, 2-10V, or 4-20 mA are common examples) and sends this signal to the controller. Even the highest quality transmitter will degrade the input signal to some extent, so a RTD with lead length compensation (4-wire RTD) can be more accurate than a transmitter if properly implemented.

3.3.2.2. Electrical Noise

Voltage and current loops A sensor measurement can be sent to the controller as a voltage or current loop. As the name implies, a current loop is simply an approach where the flow of current in a wiring circuit is varied in direct proportion to the measured variable. Current loops are often used since the signal is immune to both voltage spikes and degradation due to transmitting over a distance. As long as there is enough voltage available to drive the 4-20 ma signal through the total resistance of the circuit, then the wire can run as far as necessary. For example, a standard supply voltage of 24 vdc will drive 20 ma through a resistance of 1200 ohms.

Some transmitters require a separate power supply to drive their electronics and are commonly called four-wire transmitters (not to be confused with the 4-wire lead length compensation circuits described above). Others transmitters that simply take their power from the current loop itself are commonly called two-wire or self-powered transmitters.

Current loops use scaling resistors to generate a voltage signal at the controller. The scaling resistor at the controller accounts for either 250 or 500[3] ohms, which leaves 950 ohms for wires and transmitter electronics. The manner in which the current loop is terminated and the quality of the scaling resistor can impact measurement accuracy. Poor connections and low accuracy resistors will degrade signal quality. The biggest factor is the scaling resistor, which is predictable and controllable. A well done, secure connection will eliminate potential erratic problems. Many newer systems have the scaling resistors built directly into their circuit boards eliminating potential termination problems and locking down the quality of the scaling resistor to factory specifications. The scaling resistor installation is discussed in more detail under Section 3.3.4 Controller Error.

Voltage Signal Interference If a voltage signal is used instead of a current signal, the wire must be shielded to guard against voltage spikes and noise. Nearby cabling, conduit, or walkie-talkies can result in signal interference that reduces the accuracy of electronic measurements. Twisted shielded pair wiring (TSP) is made of two wires twisted and surrounded by a foil or braided metal shield. The twisted wires carry the signal from the sensing element to the controller. The shield prevents noise from external electromagnetic and electrostatic sources from affecting the signal on the twisted pair, while twisting the wires cancels out electromagnetic interactions between the pair of conductors that contribute to signal noise.

In connecting the shield to ground, the charges, voltages and eddy currents are dissipated. To aid in grounding the shield, a bare wire is wrapped into the shield and grounded. Since the shield is metallic, it has the potential to become a current carrying conductor itself. If this were to happen, then many of the benefits provided by the shield would be negated, and the shield itself could induce noise into the signal conductors. To prevent this conduction of current, the shield grounding wire must be grounded on only one end.

One of the functions of grounding is to establish a uniform voltage reference for setting voltage values in electrical circuits. In most buildings, this reference is established by connecting the grounded conductors to the earth or ground (hence the name) or to conducting structures in contact with the earth (like water lines or building structural steel that has been connected to a buried grid of conductors). This reference is sometimes referred to as the grounding plane, and we tend to think that the ground reference voltage is equal at all points in the grounding plane. But in fact, the ground voltage is not always equal due to the resistance of the elements that comprise the grounding plane and variations in the resistivity of the soil. As a result, small differences in voltage exist at various points along the grounding plane. If a conductor connects these points, a small current will flow. The voltages that induce currents between points in the grounding plane are often called common mode voltages, and the problems that the currents create are often called common mode voltage problems.

Most electronics packages are designed to deal with common mode voltage problems and often contain a specification rating termed common mode rejection, usually expressed in terms of decibels or db. A higher rating is better than a lower rating. With current technology, there will seldom be a problem if the equipment has common mode rejection capability, the field wiring is shielded, and the shield grounding is handled correctly. When a problem does occur, it will typically occur on a very low level signaling system (milli-volt inputs from direct wired RTDs for example) or on the communications wiring on a network.

RTDs and Noise Of all the inputs commonly found on HVAC systems, RTDs are probably the most subject to noise problems. Generally, this is because of their low resistance change per unit temperature change (fractions of an ohm per degree) and the millivolt signals associated with the bridge systems used to read them. In theory, these issues can be handled with proper shielding techniques, which is not as easy as it sounds in a real construction environment because incorrect shielding techniques are difficult to track down. Adding a transmitter to an RTD solves both problems with lead length resistance noise, especially if a current loop is used.

A DDC controller’s hardware is physical components, including circuit boards and terminal strips. The programming in a DDC controller is the software, which in turn has two components of its own, the point database and the operating logic. Both of these are typically stored in the controller’s programmable memory to allow them to be custom configured to the needs of the system. The point database contains information about both physical and virtual points. The information typically includes the type of point, scaling factors, name, descriptor, engineering units, alarm states, and the physical terminals or memory location associated with it. The operating logic software tells the controller what to do with the points defined by the point database.

The controller’s firmware ties the operating logic and point database together with the controller’s microprocessor. Essentially, it is the operating system of the controller. The firmware is typically stored in programmable read-only memory, which can be modified by the factory to add improvements or fix bugs but cannot be modified by routine programming in the field. Due to memory limitations, some systems store part of the point database information and non-critical programming at a location other than the controller. Information like point descriptors, comment lines, and other parameters that are not essential to run the program are stored in other devices, such as the network supervisory controllers or the host computer’s hard disk. Someone looking at the system through the operators console window can view all the information seamlessly, but if a programming tool were attached to the controller itself, some of the information may not be available.

3.3.3. Transmitter Error

Limiting transmitter span is a technique that can be used to provide greater measurement resolution and accuracy. For example, a 12-bit A to D converter at the controller resolves an input into 4,096 counts (212) , so the 16 ma span of a 4-20 ma transmitter is divided into 4,096 units. If the 4-20 ma signal represents a temperature change (span) of 100°F, then the transmitter can break the 100°F down into 4,096 parts so that each count represents 0.024°F. The temperature cannot necessarily be read to ±0.024°F accuracy for other reasons like repeatability, hysteresis, drift, and sensor self heating, but the signal that is received can be broken down into an identifiable part that is that small. If the span is reduced to 50°F, each of the 4,096 counts represents a smaller increment, and in this case the resolution is 0.012ºF per count.

3.3.4. Controller Error

A number of measurement errors can occur at the controller due to incorrect scaling parameters, limited analog to digital converter resolution, problems with scaling resistors, and the controller’s change of value limit setting. The description of the controller software, firmware and hardware in the sidebar give background for this discussion of controller errors.

3.3.4.1. Scaling parameters

The scaling parameters for each point must be correctly set in the DDC software at the controller. This information is usually entered as configuration information for the point database when the controller is set up, a function that is independent of developing the software code that executes the control strategy for the system. In some systems, the scaling parameters for a temperature measurement using a 4-20 ma signal can be set directly as ; i.e. 4 ma = 20ºF and 20 ma = 200ºF. In other systems, these scaling parameters are based on the span set at the sensor or transmitter.

3.3.4.2.  A to D Converter Resolution

After the controller receives an analog signal from a sensor, the signal must be converted from analog to digital for use in the controller microprocessor. The resolution of A/D converters is a result of the resolution (8-bit,12-bit, etc.). As stated in the discussion of limiting the transmitter span, a 12-bit A/D converter can resolve a measurement span into 4096 increments. An 8-bit resolution resolves into 256 (28) increments. If the resolution is too low, then the analog sensor reading will be degraded, since the step values in the measurement will be too large.

3.3.4.3. Scaling resistors

Current loops use scaling resistors to generate a voltage signal at the controller. If the scaling resistor is not already incorporated into the controller circuit board, it is necessary to add high precision scaling resistors in the field.

Typically, this scaling resistor is mounted under the same terminal as the twisted shielded pair that is bringing the field signal into the controller. When the solid resistor lead is clamped under the same terminal screw as the stranded field wiring, the clamping action on the stranded field wiring is limited. Over time, the scaling resistor ends up looking like a loose electrical connection and the resistance changes. While this resistance change is not large, the voltage drop it produces is erratic and also significant in magnitude relative to variations in voltage drop that occur at the precision resistor as it converts current changes representing fractions of a degree to voltage. As a result, a variable and unpredictable calibration error is introduced into the system.

Figure 3.1 Resistor Terminal Block

The resistor is circled in red and brazed between the two bus bars (top and bottom). To use it, you wire the current loop so that it flows through the resistor from A to B, and then pick up the voltage drop and take it into the system I/O at C and D.

Fortunately, there are specialized terminal blocks that are made to handle this problem. One such terminal block is illustrated in Figure 3.1. The blocks consist of double deck terminals with a precision resistor soldered between the upper and lower deck. The current loop flows through on one side and the voltage across the resistor can be picked up on the other side. Several firms manufacture these terminal blocks with resistors sized for the two most common standards (250 and 500 high precision). Other resistance values can be special ordered.[4]

Terminal strips and some of the special features that can be built into them can also:

·       Make troubleshooting easier at the controller by allowing the field wiring to be isolated from the controller without lifting wires via a simple, built in switch mechanism

·       Allow standard troubleshooting techniques to be developed that allow operating staff with less experience to diagnose and correct field sensor problems.

·       Provide a consistent and accessible point of identification for the numerous wires associated with the control system.

·       Make future controller upgrades and replacements simpler and faster to accomplish.

·       Serve as boundary between sensor installation and panel control responsibility.

3.3.4.4. Controller change of value limit

The I/O microprocessors in a controller scan the data from its inputs, put the data into memory, and execute programming code using the data. Change of value (COV) limits are often set to control two areas:

·       How much change in the measured value must occur for the controller to process the input values in the programming code.

·       How much change in the I/O points must occur for the controller to send information over the network to the workstation.

If the change of limit parameter is too low, point fluttering will result in many small changes in output and many updates to the central workstation – this high traffic can cause the network to crash. Filtering techniques can also help reduce this occurrence. If the COV limit is set high, the controller may not have good control response to the sensed inputs or the workstation will not be updated with the measured values that the controller is acting upon. It is important to keep in mind that the microprocessor may be working with a different number than what is seen at the workstation screen.

For example, if the COV limit for the network to sense a change is 1 degree and the COV limit for the microprocessor to execute the programming code is 0.1 degree, then the microprocessor could be controlling to a tighter degree than is seen at the workstation. This is especially problematic when tuning control loops, since the workstation may not be displaying the true system response. When a step change is introduced into the control system, a large COV limit for the network makes it easy for the technician to think that the system is sluggish and needs to have its tuning parameters adjusted. The loop may actually be hunting, but the magnitude of the hunt is smaller than the network COV limit, so it is not visible at the workstation. As a result, the technician may overreact and begin to modify the tuning parameters to make the loop more sensitive, which only makes the problem worse and can lead to another problem called wind-up. (See the Functional Testing Guide, Overview and Applications of PID Control for a description of wind-up). As a general rule, a stepped, square-wave system response curve viewed at the operator’s workstation is a good indicator of a large network COV limit, especially if the step changes are uniform in magnitude.

3.3.4.5. Software version

Without careful management of the system’s software, it would be quite easy for a working program to be replaced with previous version of the program. This overwriting could even occur automatically, triggered by a controller recovering from a power failure and downloading bad information. As a result, a well-tuned system can suddenly have problems.

As good operating practice, it is important to make sure that any modifications made to a controller’s software are saved and backed up appropriately. Different systems handle on-line program editing in different ways. In most cases, there will be at least two locations where the program data is stored. One is in the controller, where the program is actually running. The other is on the hard disk of the operator’s workstation, where a copy of the program (or some version of the program) usually resides. Additional copies of the program or edited versions of it may also exist on the laptops of the control technicians working on the system and on back-up tapes and CDs. The back-up method used should provide a standard method for saving versions of the software. These procedures can be addressed during operator training and through documentation to minimize the possibility for corrupting the system’s software by overwriting the latest version of the software with an older, incorrect version (see the sidebar below).

Some systems allow technicians to edit a copy of a program or point database and then download the modified software to the controller, overwriting the existing software with the new information. If the starting point for the modification was not a copy of the current program, an existing, error-free program could be overwritten by a program or with a bug or error. Even if the modification used the working program as a starting point, it is possible that the modification may turn out to be a bad idea and having a backup copy of the working program may be highly desirable.

Other systems allow programs to be modified “on the fly”: Modifications made to a controller’s software will not be reflected in the copy on the hard disk unless the completed modifications are manually uploaded. If proper back-ups are not maintained, then the “good” software residing in the controller may be lost if the controller’s memory is cleared by a power failure, or if the “good” software is over-written by an older version downloaded from the disk by an operator error or in response to some other system event.

3.3.5. Network Architecture Effects

Designers should keep network architecture limitations in mind when designing control systems so that the network does not adversely affect control functions. The way the network packages data for transmission affects the data seen at the workstation. The design of network communications can severely limit the frequency that data is updated at the operator’s workstation as well as increase the control loop response times. These effects, while not errors in the purest sense of the word, can cause data to be misinterpreted.

While a properly implemented control loop should not be configured to run over the network and thus will be immune to the direct effects of slow network communications (see Section 3.2 Point Selection, recommendation #3), the ability to tune the loop based on the observed performance at the operating console can be severely limited by slow data handling. The functionality of the control system graphics and the ability of the operators to command and control the system manually, especially in an emergency, can also be compromised by poor network functionality.

Commissioning providers need to be aware that when tuning a loop or observing the performance of a system, there may appear to be delays in the system’s response due to data handling problems if there are many controllers on a lower speed network. These delays may not actually exist at the controller making the system more responsive than it appears. These delays are a result of the network architecture, with controllers either residing on a network with peer-to-peer communications or on a lower speed network with supervisory (or global) control modules polling the controllers. Section 2.4.3.2 Component Specifications: Controllers and Actuators includes a detailed discussion of network communications.

3.3.6. Workstation Effects

Other opportunities for error beyond the controllers exist – the data that is shown on the workstation graphics may not be the same data that the controller sees. When troubleshooting a questionable reading at the workstation, it may be necessary to check the measurement directly at the controller using a laptop. As long as the microprocessor in the controller is getting good data from the input devices, then it can control well. If the data is not properly conveyed to the workstation, then it may not be apparent that the controller is working.

3.3.7. Installation Errors

A temperature sensor located near a diffuser or computer monitor, or an outdoor air temperature sensor that is not properly shielded from solar radiation can be large and highly variable sources of error. Outdoor air temperature sensors can also be affected by exhaust air. Locating a sensor on the wrong branch of a pipe or duct tee can totally change the data that it is reporting.

The maintenance staff members for a new facility, who had yet to receive control system training, were plagued with cold complaints from the break room. This seemed to be caused by a major calibration problem with the input sensor. Unfortunately, they could not find the sensor since it was not shown on the plans and the location was not obvious when they inspected the space. The “light came on” several weeks later during the control system training when the trainer projected a picture of the temperature sensors used on the project up on the training screen. The sensors were designed in an architecturally pleasing package that looked just like the blank cover plate used to close off an unused electrical device opening in a wall. Just such a cover had been noticed in the troublesome break room, directly behind the large coffee maker. The next day, relocating the coffee maker solved the break room temperature control problems.

Placing sensors in the appropriate location can be difficult. It is not uncommon for the installing tradesmen to misinterpret the location of a sensor shown on a schematic or described only by a note. Showing the sensors on floor plans can help, but floor plans are still a two-dimensional picture, and some interpretation of exactly where the sensor goes is required. Often inches can mean the difference between being in the right spot and being in the wrong spot. An effective way to minimize this problem during the construction process is to locate the sensors during a walk-through of the project conducted after the systems in which they will be installed in are in place. The exact location can be marked on the piping and ducts with indelible markers using the point symbol and point number. This approach provides benefits to all parties involved in the process:

·       The Designer is provided with the chance to be sure that the sensor is installed exactly where it needs to be in order for the design intent of the project to be achieved.

·       The Owner and their Operating Staff are exposed to the location and installation requirements of the sensors controlling the systems they will be charged with operating and maintaining. This can be a valuable training lesson and also can provide insight into the design intent behind some of the sensor requirements.

·       The Control Contractor and the Installing Tradesmen have the opportunity to review the sensor location in the context of the surrounding systems as well as the physical arrangement of the system the sensor is installed in with the designer present. Design and contractual issues that may not be obvious in plan or schematic can be identified, discussed and resolved on the spot in an interactive environment involving all concerned parties in the process.

Along these same lines, during start-up, it may be advisable to go look at a sensor that is providing what appears to be inaccurate questionable data. You may discover that it is in the wrong place and is actually providing the system with very accurate but inappropriate information.

The bottom line in the commissioning and operating arena is to remember that what is happening at a sensor in the field is not necessarily the same as what you see on the workstation. Understanding the different opportunities for error from the sensor to the workstation allows a commissioning provider to anticipate and troubleshoot these problems in the field.

3.3.8. Persistence of Accuracy

Checking that the calibration is within a tolerance and the signal is correctly displayed at the workstation should be a part of a regular maintenance routine. New sensors and transmitters will degrade over time due to factors such as thermal drift, mounting effects, vibration effects, and aging of components. The decay is usually a function of the quality of the transmitter. For a given accuracy, more expensive sensors have higher quality housing construction (i.e., utility box housing vs. explosion proof) and long-term stability of the electronics. Long-term stability and low thermal drift may not be an issue for a space temperature sensor, but it could be absolutely critical for a chilled water sensor, especially if the sensor is used for load calculations by a load-based control algorithm.

A number of persistence issues are described below.

·       Some sensors can be sensitive to the location and manner of installation. As a result of mounting effects, a sensor calibrated for one location will not necessarily remain calibrated if it is moved to another location or if it is installed in a different orientation from that which it was factory calibrated.

·       The connection of the twisted shielded pair to the controller can degrade over time, with resistance developing where the stranded wire (twisted pair) is clamped to the scaling resistor lead. The added resistance results in large errors in measurement. See Section 3.3.4 Controller Error for more details.

3.4. Installation Guidelines

The following section presents guidelines for installation of monitoring and control devices. Installation details for all safety points are critical to safe operation of the building’s systems, Calibration test ports and techniques are emphasized since careful calibration is essential for proper control.

3.4.1. Safety Points

Three general guidelines apply to all safety interlock points.

1   Safety points should be hardwired to the system they serve to shut down the air handling unit regardless of the operating status of the computer system, its network, or the position of the starter or VFD selector switches (Hand or Auto - Local or Remote - Inverter or Bypass). The control system software should not be relied upon to perform critical safety operations, since the software will not shut down the system if the fan starter, output circuit board, or drive selector switches are placed in the hand or bypass positions.

2   The input to the safety circuit should be a different sensor than the input to the control loop. Otherwise, the “fox is guarding the hen house” since the sensor tries to protect the system from a problem that could be caused by its own failure. For instance, an independent freezestat should be provided even if the air handling system has a mixed air temperature controller or sensor. The independent freezestat provides a valuable “second opinion.” A number of potential problems ranging from calibration to computer failures could make the DDC mixed air sensor unavailable or unusable for safety purposes. In addition, the independent devices used for safety controls are by design, fairly simple and rugged pieces of hardware that are immune to programming errors, microprocessor failures, and network communications problems.

3   The safety device should require a manual reset to allow the system to resume normal operation. Making the safety device a manual reset type of switch forces a manual intervention to allow the system to resume normal operation. Hopefully, the person resetting the system will first investigate the cause of the safety trip and correct it before resetting the switch. This is an important training issue and a good target for an operating procedure for owners who use written procedures to direct operators in various situations.

4   Failure modes need to return dampers and valves to safe positions. Hardwired interlock switches piloted by the proof of operation circuit should be provided to return critical dampers and valves to a normal, fail-safe position when the system is not operating. This is important because the system could be shut down by functions outside of the DDC controller, including manual intervention at the starter by selecting “Off” on the selector switch, belt or drive coupling failures, or undetected localized power outages. Items to consider include:

·       Failure to close these outdoor air or relief dampers when the unit was off could cause a coil or plumbing in the building to freeze during cold weather. During warm weather, the open dampers could allow the building to fill with hot, humid air, creating a potential condensation and pull down load problem at restart.

·       Failure to close the chilled water valve on inactive units can waste pumping energy at the central plant if it remains in operation to serve other loads. Leaving the valves open can also cause performance problems with variable flow central plants since the inactive coil with an open control valve creates a hydraulic short circuit that depresses the return temperature while creating a flow condition associated with a high load conditions. On some systems, the wide open valve may cause condensation problems because the stagnant air in the vicinity of the coil will be cooled to very near the chilled water supply temperature rather than the chilled water return temperature, which locally lowers the dew point of the air and the temperature of nearby surfaces below the surrounding dew point.

Description of the operation and guidelines for the application of the following typical safety points are included in the following sections.

·       Freezestat

·       High static pressure switch

·       Proof of Operation Interlocks

·       Air flow switch

·       High temperature limit control

·       Smoke detector

·       Firestat

·       Position and limit switch

·       Multi-Speed Motor Interlocks

·       Variable speed drive programming

·       Motor Overloads

3.4.1.1. Freezestat

The freezestat protects coils from freezing when cold outdoor air enters the mixing box by providing an input that is interlocked to shut down fans, close outside intake dampers, and open heating coil valves. For mixed air applications, freezestats on the entering side of the heating coil and monitoring the heating coil return water temperature may be necessary to fully protect the system, since a reduction in return water temperature occurs and can be detected faster than the air temperature drop. To attempt to avoid the system outage from a freezestat trip, the DDC system can generate a warning alarm ahead of a system shut down by the freezestat.

The freezestat elements should be tie-wrapped with the DDC mixed air temperature sensor element. For large systems, more than one freezestat and more than one mixed air sensor may be required to adequately cover the plenum. A good rule of thumb is to provide 1 foot of element for every 4 square feet of coil face area. In this way, the DDC system and the freezestat see the same temperature. Remember that freezestat elements may look like averaging elements, but they do not average; rather, they are designed to respond to the lowest temperature seen by a specific length of the overall element.

See Section 3.6.3.2 Freezestats in the Functional Testing Guide for a detailed description of operation.

3.4.1.2. High static pressure switch

A supply air duct high static pressure switch set at the duct pressure class rating shuts down and locks out fans to protect the air distribution system from excessive pressures that could damage ducts. This type of switch will protect the equipment from a relatively gradual over-pressurization or continued operation at an over-pressurized condition. This over-pressurization is created by a restriction in airflow that pushes the fan up its operating curve, where the peak on the operating curve is higher than the rated static pressure of the duct system. An example of a situation where a high static pressure switch might be desirable would be a large system that was equipped with smoke isolation dampers in which the potential exists for the fan to start with the isolation dampers closed.

High static pressure switches will not protect a system from the effects of air hammer that can occur when fire dampers or smoke isolation dampers close suddenly with the system in operation. A fan that is moving approximately 25,000 cfm is literally moving a ton (2,000 pounds) of air a minute. For a detailed description of how to protect a design against air hammer, refer to Section 11.4.3 in the Functional Testing Guide.

3.4.1.3. Proof of Operation Interlocks

The supply fan proof of operation interlock should be used to pilot hardwired interlocks to close critical valves and dampers. Interlocks should be provided for the outdoor air damper, relief damper, smoke isolation dampers, minimum outdoor air constant volume regulators, and the chilled water valve when the fan is not in operation. These hardwired interlocks are in addition to providing proof of operation indication to the DDC system. This interlock should function regardless of the state of the DDC system or any selector switches controlling the fan and should be totally independent of the DDC system software.

Figure 3.2 Duct and Damper Damage due to Air Hammer

 

(Image courtesy of the Ruskin Catalogue)

3.4.1.4. Air flow switch

Air flow switches are typically provided with electric reheat coils to shut them down upon loss of air flow to protect the coils from overheating and causing a fire. The National Electric Code requires them along with temperature based limit controls, some of which may require a manual reset when tripped. The air flow switches are usually differential pressure switches arranged to sense the velocity pressure entering the coil.

3.4.1.5. High temperature limit control

High temperature limit controls shut down electric reheat coils if the coil casing temperature exceeds the setpoint. The National Electric Code also requires these devices. Only automatic reset limit switches will typically be installed on smaller (in terms of kW rating, not face area) coils. Larger coils will have an additional limit control that must be manually reset if it is tripped.

3.4.1.6. Smoke detector

Smoke detectors are typically installed in the supply and return air streams to the air handling unit and at smoke separations. They can be hardwired to shut down fans, shut smoke dampers, and perform other fire and smoke management functions directly. However, they are often wired to signal the fire alarm system, and the fire alarm system sends commands to the appropriate dampers and motor control circuits.

There are several important points to remember with regard to smoke detectors in addition to the points already made about safety devices.

1   Because the smoke detectors are often a part of the fire alarm system that is installed in the HVAC duct system and interfaced to by the control system, they require a coordinated effort on the part of several specification sections and trades in order to provide a complete and functioning installation. The designer can arrange their specifications and drawings to make the coordination requirements clear to all parties and define key requirements and responsibilities.

2   Duct mounted smoke detectors usually have sensing tubes that extend into the duct system into the detector, which is typically located in a housing on the outside of the duct. Specific requirements apply to the position, length, and orientation of the sensing tubes in the duct system that must be met in order to ensure the correct operation of the detector and maintain its U.L. listing. These requirements are specific the each manufacturer’s equipment. Thus, while the sensing tubes from Manufacturer A’s detector may physically fit into Manufacturer B’s detector, they probably cannot be substituted for Manufacturer B’s sensing tubes without violating Manufacturer B’s U.L. listing. Similarly, in most instances, the sensing tubes cannot be field cut to make a tube fit an existing condition without violating the U.L. listing of the detector.

3   Very large ducts may require extra detectors to adequately protect them. Local code officials may require this even if there is no specific language in the applicable building code regarding detector requirements as a function of cross sectional area. It is usually a good idea to meet with the local code official sometime during the design process to review issues like this. Most officials will welcome the opportunity to influence the project in the early stages rather than having to fight to get a change made during the final stages of construction. Arranging for such a meeting lets the code official know that the designer is aware of their needs and requirements and can pave the way for a good working relationship throughout the project.

3.4.1.7. Firestat

Firestats are manual reset high temperature limit devices required by code in smaller air handling systems were smoke dampers are not required and in exhaust systems. Generally, firestats are not as common as they were at one time since smoke detectors can provide the required protection along with other features needed for code compliance on a project.

3.4.1.8. Position and limit switch

Fire and smoke dampers are often equipped with limit switches to monitor and indicate blade position. Back draft dampers on exhaust fans and parallel fans are often equipped with limit switches, as well as intake dampers on 100% outdoor air systems. Generally, the limit switches will be two position switches that indicate that the damper is either open or closed. However, it is also possible to obtain analog switches that provide an output that is proportional to the rotation of the damper blade if this type of information is required by the control system or the smoke management system.

Many buildings incorporate fire and smoke management systems that require that a panel be provided to allow a fireman to monitor and manually control the various smoke dampers and fans in the building in a fire situation. Usually this requirement is driven by code requirements, but it can also be driven by an owner’s own internal requirements or by the owner’s insurance underwriter’s requirements. To convey damper position information, systems of this type use the damper position switches for an indicating circuit to light pilot lights at a fireman’s control panel. In most cases, factory installed limit switches provided with the damper will provide adequate service for this type of application.

The damper position and limit switches can also be used to provide a permissive interlock that will not allow or “permit” the fan system to operate until key dampers are in safe positions. These switches will ensure that fan operation does not over or under pressurize the ductwork due to restricted air flow. While factory-installed switches can be used for this function, they often lack the precision and adjustability required to provide the best level of protection and adjust the trip point of the switch to the operating requirements of the system. Some issues to consider when selecting and applying limit switches include:

Figure 3.3 Limit Switch Installation for a Permissive interlock

This standard motion control switch’s sensing arm has been adjusted to respond to blade rotation (red circle). Note also how the Unistrut™ mounting support arrangement coupled with the flexible conduit connection (green circle) will allow the switch to be easily shifted up and down and in and out to fine tune the trip point.

·       Arrange the switch to sense blade position, not linkage or shaft position. Shafts and linkages can come loose from the blades they drive.

·       Multiple section dampers may require multiple limit switches. In general, the goal should be to prove that enough of the damper assembly is open to allow the fan to move the design flow of air without collapsing the intake system or blowing apart the discharge system.

·       The switch mounting arrangement should hold the switch securely, but allow it to be easily adjusted vertically and horizontally to fine tune the point at which it trips. Unistrut™ and similar steel channel type framing systems are provide a variety of options for achieving this using standard fittings and clamping bolts that slide inside of C channels. (see Figure 3.3).

·       The limit switch should be adjusted so that it does not sense that the damper is open until the damper is truly open. In addition, the differential should be tight enough that, on the return stoke, the switch indicates that the damper is no longer safely open at nearly the same point. When the damper is commanded open, the switch should be adjusted so that it does not indicate that a fan start would be safe until the damper is open far enough that the pressure drop through it with design flow would not create an unsafe pressure downstream of the damper. Typically, this will be at 85° or more of blade rotation. Similarly, when the damper starts to close, the switch should open and indicate that the damper is no longer safely open at nearly the same point, in the case of this example, 85° of blade rotation. Obviously, any real switch will have some hysteresis and switching differential, so the make and break points will not be identical. Lower quality switches often have wider switching differentials and more hysteresis than higher quality switches.

Additional information regarding this type of interlock that is peculiar to economizers can be found in the Functional Testing Guide, Section 3.5.2: Operational Interlocks.

3.4.1.9. Multi-Speed Motor Interlocks

While variable speed drives are being used for many applications previously served by multi-speed air handling equipment, multi-speed motors are still used on projects where only two distinct operating points are required, and where the complexity associated with a VFD is not desired by the Owner. Multi-speed motors require some special interlocks above and beyond those provided as standard for a typical motor interlock circuit to protect the motor and equipment. Generally, there are three areas of concern:

1   The interlocks must be arranged to prevent the high-speed motor winding from being engaged or energized at the same time as the low speed motor winding.

To provide two-speed control with a DDC system, it is typically necessary to have two outputs. There are several ways to use these outputs to produce the speed change. The manner in which the outputs are implemented can enhance the interlocking safety designed to protect the system from simultaneously energizing both motor windings. The most obvious approach is to use one command to control the high-speed starter coil and one command to control the low-speed starter coil. This approach allows the system’s software to be used to provide the speed command.

Using a relay with a Form C contact (Figure 3.3) to perform the speed change command can provide an added measure of safety because the physical arrangement of the contact makes it virtually impossible to have both the normally open and the normally closed terminals energized at the same time.

Regardless of how the control system command points are arranged, an additional hardwired interlock should be provided in the starter as a final measure of insurance against simultaneous energizing of the high and low speed windings. This interlock can be accomplished by wiring a normally closed auxiliary contact[5] from the high-speed motor starter in series with the low-speed motor starter. While somewhat redundant with the interlocking features described in the preceding paragraphs, this hardwired interlock provides added insurance in the event of a manual override of the control logic at the starter selector switches for little additional cost.

2   All other safety interlocks must be arranged to shut down the motor and machinery regardless of which motor speed is active and regardless of the position of any selector switches.

3   The interlocks need to be arranged to protect the drive system that transmits the power from the motor to the driven machinery. There are two components to this.

a)   Some drive systems have minimum speeds that they must operate at or above to ensure proper lubrication of all the internal components (gear boxes are typical of this). This is a design requirement that is often addressed by the equipment supplier in a manner that is nearly transparent to the designer. But for an equipment substitution or a repair to an existing system, this area can become a commissioning issue.

b)  When the motor changes to a lower speed, the utility system imposes a significant and nearly instantaneous braking effect on the rotating fan wheel, drive system, and rotor. This nearly instantaneous change in speed and dissipation of energy places a huge stress on the drive train components and can destroy gearboxes, shear shafts, and keys and cause belts to be thrown. This problem can be averted by providing a time delay in the control logic to delay energizing the low speed winding for an interval of time after the high speed winding are de-energized. The delay should be long enough to allow the air movement loads on the fan wheel and drive and bearing friction to slow the wheel down to a speed at or below that associated with low speed operation so that when the low speed windings are energized the fan is already running at the lower speed or requires a slight acceleration to bring it up to the lower speed. It is tempting to simply use the DDC programming capabilities to provide the time delays and eliminate an independent time delay relay from the control circuit. This approach should be used with caution because if someone elected to manually make the speed change, they could damage the drive system if they did not wait for a sufficient interval of time between de-energizing the high-speed windings and energizing the low speed windings.

Figure 3.4 illustrates a typical multi-speed motor interlock circuit and illustrates some of the concepts discussed in the preceding paragraphs.

Figure 3.4 Typical Multi-speed Motor Interlocks

AutoCAD users can double click to open the image.

3.4.1.10. Variable speed drive programming

Most current technology variable speed drives incorporate microprocessors that allow them to perform a wide variety of functions that can be used to tailor their application. There are some parameters related to the safety of the drive that are usually arranged to prevent them from being placed in a state that would jeopardize the drive. Other parameters pose no threat to the drive in any of the possible states, but could pose a threat to the HVAC system or the loads served by them if not properly programmed. In some ways, this programming is more of a commissioning issue than a design issue but designers need to include provisions in their drive specifications for drive start-up, programming, and training by a person knowledgeable with all aspects of the drives. It is helpful to include a list of HVAC-oriented requirements that relate to typical drive parameters to guide this process.

Many designers simply specify a factory start-up of the drive. This is highly desirable, but if it is not provided, then someone familiar with the drive should be required to perform the following work at a minimum.

·       Disable VFD capability to function regardless of the status of the external safety devices and interlocks wired to its input terminals: The design of some drives allows the drive to be placed in a local control mode that will cause the drive to ignore all external safeties and interlocks, regardless of their status. While useful in the process industry, this feature has the potential for disaster in HVAC because it can easily be misinterpreted by the operators, who think they are simply taking control of the start-stop function when in fact they have also taken away any safety functions. Most drives with this feature have a programming parameter that allows it to be disabled so that an operator cannot place the drive in the external safety override mode.

·       Enable and Properly Set the DC Injection Braking Parameters: Most current technology drives have the capability to perform a braking function by injecting a DC signal onto the motor power circuit. The stationary field created by this DC signal can bring a rotating motor to a stop much more quickly than would occur if it were simply allowed to coast to a stop. It can also stop a motor that is spinning backwards. This feature can be particularly useful in HVAC. A spinning fan connected to a de-energized motor is putting energy into the motor shaft causing the motor to actually function as an unregulated generator. If the variable speed drive is engaged against this unregulated voltage source, the voltages and currents created can damage the drive circuitry. Programming the DC injection braking to bring the motor to a complete stop prior to starting and accelerating it can protect the drive from damage.

·       Coordinate with the Control Contractor and Commissioning Team to Adjust Drive Parameters During Start-Up: Drives have many adjustable parameters that can be modified to match the unit to the requirements of the system. Some of the more crucial parameters related to the voltage and motor characteristics are typically factory set. Others must be configured in the field to tune the drive to the system. For HVAC systems, the key field-tuned parameters typically include:

·       Acceleration and deceleration settings.

·       Minimum and maximum speed settings.

·       PID loop parameters if an on-board PID loop has been provided and is being used for control.

·       Communications settings if the drive is being networked with the automation system via a communications bus.

·       Programmable input and output functions.[6]

·       Input scaling factors.

3.4.1.11. Motor Overloads

The National Electric Code requires motor overload protection for all motors. Smaller motors (fractional horsepower, single phase) provide this protection via internal sensors that simply shut down the motor if it gets too hot. Larger motors typically require independent components external to the motor to serve this function, typically mounted in the starter or drive. Motor overloads can be electro-mechanical or solid state but the primary function is still the same: to shut down the motor if it is operating under a sustained overload condition. The overload condition is different from a short circuit condition. Short circuit protection is provided by other components in the motor power circuit, usually the fuse or circuit breaker. Overloads typically must be adjusted to match the motor nameplate amps and service factor rating so that them motor is shut down if the load on exceeds its rating. When properly sized and selected, they can also provide a measure of protection against “single phasing”, the situation that occurs when a motor experiences a power loss on only one phase. When this happens, the motor will generally continue running but will rapidly overheat. Properly sized overloads can detect this condition and shut down the motor to protect it from damage.

3.4.2. Manual Override

The owner may desire a simple way for the zone occupants to override the operating schedule regardless of whether or not a facility operator is on site. The manual override can be accomplished via a variety of techniques including:

·       An independent input wired to a timer controlling a contact. Spring wound timers are the least expensive approach to this since they require no external power supply. However, some owners find the ticking sound associated with them to be annoying. In these instances, providing a powered timer is necessary, which requires power wiring in addition to the pair of wires that monitor the contact. These switches can be purchased with a feature called a “Hold” setting that locks the switch on. This feature is not desirable for an override application because it allows the tenants to easily defeat the scheduled fan operation by locking the switch on. Key activated switches can be provided if an owner only wants certain parties to have access to the override feature.

·       Some systems provide an auxiliary switch as an option on their zone temperature sensors. One approach may include an independent contact that is simply housed inside the thermostat and requires an independent input. Another approach uses one pair of wires to monitor temperature, switch status, and drive a digital temperature display.

·       Some systems, either as a standard option or via an interface with independent third party equipment, can provide tenant override capability through a phone connection. Generally, this approach provides each person authorized to initiate an override function with an access code. Some systems also allow the start time and number of hours to be entered rather than simply starting a fixed override cycle when the authorization code is entered.

Regardless of the approach used, the system can be programmed to accumulate the hours of override operation for each switch and report and archive the data. This feature allows owners with tenants who lease the space to bill for extra HVAC operation beyond the normal hours provided. Tracking override hours can be a strong energy conservation incentive because it makes the tenants accountable for their use of the system after hours. Owners with multiple departments can employ similar tracking in which the extra hours are billed to the departments.

3.4.3. Calibration Test Ports

Without sensor test ports, it is difficult to verify the calibration of installed sensors. While absolute calibration in the field is difficult, it is still important to have a way to cross-check the indication from an operating sensor with the system online and without removing the sensor from the system or control circuit.

To facilitate installation of calibration test ports, all piping penetrations should be put in place in the construction phase. Putting a weldolet into an empty piping system during initial pipe fitting work takes one pipe fitter about 20-30 minutes or less. Installing the same weldolet after the system is online can take a crew of two or three people a day or more if a welding machine must be set up and a portion of the system must be drained and refilled. This process could require an outage if the sensor is in the main piping. Designers and commissioning providers should review all sensor and test port locations with the contractor early in construction by marking the pipe or duct together and then verifying the locations before filling the system. The test port should be a self-sealing plug rated for the temperature, pressure, and fluid associated with the application. The following bullets set forth guidelines for different types of calibration test ports.

·       Temperature For every temperature sensor well, there should be a second well for a calibration thermometer immediately adjacent to the sensor well. Pete’s plugs should be used when both temperature and pressure measurements are needed. Thermometer wells are more desirable for temperature-only measurement sites because they will never have seal failure.

·       Pressure Every gauge cock should have a second valved connection to allow a calibration gauge to be connected to the system. Pete’s plugs are good for locations where only occasional measurements are required. Differential pressure transmitters should have 3 or 5 valve manifolds to prevent them from being decalibrated or damaged when placed online. Install isolation valves and test ports at each pressure sensor location.

·       Flow Field calibration of a flow meter is very difficult, and providing test ports for calibration is generally not practical. If field calibration is attempted, a calibration device that is as accurate as the flow meter must be installed in series with the flow meter, or the flow meter can be removed and replaced with the calibration instrument.

3.4.4. Calibrating Analog Inputs

This section discusses calibration methods for analog and binary inputs or outputs. In general, all hardware and software calibration information should be documented in a log. The rigor of calibration for any sensor depends on the sensor application. Calibration should be performed using a standard that is as accurate or more accurate than the sensor being calibrated. It is also important to understand exactly what is being calibrated with different procedures.

For instance, unless the current is measured at the 4-20 mA level, the 4-20 mA input is a function of some other parameter. So, if calibration is performed using a 4-20 mA generator, you have verified that the control system will interpret the 4-20mA correctly, but you are relying on the 4-20 mA transmitter to interpret the sensor output correctly. A better calibration approach might be to hook a decade resistance box up in place of the RTD and calibrate to that. Of course, then you are assuming that the RTD is correct. So, the best approach is to put the RTD in a fluid or dry bath and subject it to temperature extremes and reference it against this known standard. The sensor should be subjected to a temperature near the low end of its span to set the zero adjustment on the transmitter. Then the sensor should be subjected to the high end of its span to set the span adjustment on the transmitter. Since some transmitters have interactive span and zero adjustments, (changing one affects the other slightly) repeat the calibration process several times. Higher end transmitters may have non-interacting zero and span, but it is wise to verify the calibration at least once. Calibrating with a reference fluid temperature the most accurate method, but is a very difficult and expensive process to accomplish in the field. A factory-calibrated sensor can solve the problem of field calibration.

3.4.4.1. Factory calibrated sensors

High precision calibration in the field can be time-consuming and costly. Purchasing factory-calibrated sensors with calibration certificates adds some cost to the sensor prices, but typically this cost is much less than calibrating in the field. A factory calibrated sensor and transmitter are assumed to be accurate unless proven otherwise, so the installing contractor should not make adjustments to the zero and span of the transmitter. Factory-calibrated sensors should be checked in the field, but not calibrated. If a calibration problem is identified during the check, the factory must correct the problem since they did the traceable calibration.

3.4.4.2. Calibration at the transmitter or the BAS

Most transmitters have zero and span adjustments for field calibration. Control systems also have zero and span adjustments for each I/O point using slope and intercept values or other database parameters. So where should you make the adjustment of zero and span - in the control system or at the transmitter?

If the sensor does not have a transmitter or device that allows calibration at the sensor (an RTD or a thermistor for instance), then the adjustments must be made at the control system database. If the sensor is replaced, then make sure to look at the database and make changes if necessary. If someone tweaked the slope and intercept at the control system database to field tune the previous sensor, then leaving these old parameters in the database will most likely decalibrate a newly installed sensor.

If possible, calibration should be performed at the transmitter to allow sensors to be interchangeable. Then if a sensor is replaced, it is not necessary to remember to modify the control system database calibration parameters.

3.4.4.3. Single and Multi-point calibration

Single point calibration is accurate enough for sensors that will typically only sense at one point, like space temperature. Single point calibration can also be used for a point that is controlled for a fixed value all of the time when the accuracy of the sensor is not as important as knowing a fixed setpoint. For example, a discharge temperature on a system with no reset can be calibrated at only one point. Documenting and keeping track of how the sensor was calibrated allows you to let others know so that a reading does not mislead them at a point other than the calibration point.

A typical method for single point calibration is to compare the control system display to one measurement from a calibrated instrument. The calibration should be made at a typical value for each sensor, by an instrument at least as accurate as the sensor that is being calibrated. This method can be inaccurate if the span is not adjusted correctly or if the device is not linear.

Multi-point calibration should be done when accuracy over a range of values is important. A typical method for multi-point calibration is to compare the control system display to two or more values measured by a calibrated instrument. For linear devices, calibration points should occur at the high and low limits of typical operating conditions. For non-linear devices, three or more points should be compared to the sensor look up table or curve fit with both rising and falling signals. In HVAC, calibrating non-linear inputs typically only applies to thermistors. When replacing thermistors, the new thermistor must be calibrated because the sensors are not necessarily interchangeable.

3.4.4.4. Calibration verification through energy and mass balance

Calculating energy and mass balances is one way to verify that sensors agree. Verification of calibration is especially important for systems where temperature differences are used for control decisions (i.e., airside economizers, hydronic systems). However, an absolute energy or mass balance may not be achievable due to sensor accuracy issues. Instead of taking data off the control system for the energy or mass balance, take differential readings with the same sensor to cancel out the sensor error. A thermal or mass balance indicates a problem worth investigating if the balance is outside of the worst-case window based on the sensor accuracy rating.

Examples of these checks are listed below:

·       When the outdoor air dampers are fully open, the outdoor air temperature and the mixed air temperature should be equal.

·       When the cooling and heating coils are closed, verify that the mixed air temperature equals the supply air temperature.

·       Compare the supply air flow (outdoor air + return air) to the sum of the VAV box flows.

·       The air side and water side numbers for a coil should create a thermal/mass balance.

3.4.4.5. Matched sensors

The manufacturer can select two sensors with nearly identical output characteristics so that the differential they measure is accurate. Without matched sensors, one sensor could read on the high side of the tolerance (+0.5ºF) and the other on the low side of the tolerance (-0.5ºF), resulting in a differential accuracy of ±1ºF. Having matched sensors can make energy and mass balances a more reliable means of calibration verification. As an alternative, two sensors can be matched through field calibration only if calibration instruments are more accurate than the sensors. Otherwise, you may be decalibrating the accurate sensor to match the inaccurate sensor.

3.4.4.6. Relative Calibration

In many instances the relative calibration of sensors in a system is more critical than their absolute calibration within the constraints of their specified accuracy. For the purposes of our discussion, relative accuracy is defined as the accuracy of the sensors relative to each other and perhaps some field reference standard. Absolute accuracy is the accuracy of the sensor relative to a NBS traceable standard. Two sensors that have been calibrated to the same standard with the same accuracy specification will have the same absolute accuracy.

More detailed information about relative calibration can be accessed in the Functional Testing Guide, Section 18.3.2: Relative Calibration Test. From there, a link to a relative calibration test procedure template is provided.

3.4.5. Calibrating Analog Outputs

Calibrating analog outputs essentially consists of calibrating the analog inputs in reverse. Calibration should verify that the valves and dampers have correctly stroked through their full range. Additionally, the position indicators on the actuators should be verified, and the start and span setting of any positive positioners should be inspected.

One signal may be used to sequence a number of actuators by taking advantage of different spring ranges. This type of sequencing deserves careful calibration to make sure each valve fully strokes at the appropriate time.

For valve and damper actuator design and testing information, see the Functional Testing Guide, Chapter 11: Preheat.

3.4.6. Calibrating Binary Inputs and Outputs

Binary inputs show status, or proof of operation, of equipment. Calibration of binary inputs should be done in conjunction with the binary outputs. This task verifies that the proof of operation point reports the correct status.

For binary outputs, calibration involves verifying that the command from the controller results in the correct action. During calibration, the actuation point (the value at which the sensor switches from ‘off’ to ‘on’) is checked to verify that it is within tolerance. The switching differential, or deadband, may be too small or too large. If the differential is too small, then the output may switch from ‘off’ to ‘on’ so rapidly that the point “flutters”, which can reduce component life. If the differential is too large, then once the switch turns on, it may never switch back.

3.5. Sensor Selection and Installation Guidelines

Correctly specifying and installing a sensor is based upon the importance of each point to system control and efficiency. The accuracy of each sensor should be tailored to the purpose and requirements of the function it serves. To be valuable, some monitoring functions will require better accuracy than some control functions. The reverse case can also be true. The absolute accuracy recommendation presented in this section is not as important as understanding what you are trying to measure and then matching the specification to the measurement requirement.

The following sensor selection guidelines, organized by sensing application, help designers specify sensor accuracy and installation procedures. Accuracy guidelines refer to combined effect numbers that take into account errors from the sensor to the terminals of the controller, including sensor errors, lead wiring effects, and scaling resistor errors.

3.5.1. Temperature

In this section, accuracy, installation, and calibration recommendations are presented for the following HVAC temperature sensor applications:

·       Space temperature

·       Duct temperature

·       Averaging sensor applications

·       Outdoor air temperature

·       High temperature applications

·       Critical control, monitoring, and billing applications

·       Water temperature differentials 20ºF or less

Depending on the function, temperature sensors can have a range of required accuracies in HVAC applications. Inaccurate sensors can lead to energy waste and uncomfortable space conditions. Critical temperature sensors and sensors where errors may cause significant energy waste, loss of performance, or problems with process or production may warrant more accurate standards and more frequent calibration verification. In some processes, relative calibration may be more important that absolute accuracy.

Consider a discharge air temperature sensor that reads two degrees high. The controller achieves a 55°F discharge temperature, but is actually cooling the air to 53°F. For a fairly efficient chiller plant (0.8 system kW/ton) and a 40,000 cfm constant volume reheat air handler, this two degrees of extra cooling translates to 5.8 kW of demand. Based on a typical 2,600 hour office building operating schedule, this will save 15,000 kWh per year which equates to about $1200/yr[7] (for a 24 hour operation, over $4000/yr). To maintain comfortable conditions, the overcooled discharge air is reheated at the zones using hot water coils. For a hot water plant efficiency of 80%, the reheat energy for a 2,600-hour operating year is about 2800 therms, or $1400/yr[8] (for 24 hour operation, over $4700/yr). The moral of this story is clear - selecting temperature sensors with the appropriate accuracy for the application and making sure they are calibrated is essential for energy efficiency.

Table 3.1 Temperature Measurement Technologies

Technology

Function

Advantages

Disadvantages

Thermocouple

Example application:

Boiler flue gas

Applying heat to the junction point of two wires made of dissimilar metals generates a voltage that is temperature-proportionate.

Inexpensive, rugged, good for high temperature applications, no external power supply required

Nonlinear, the lowest accuracy, cold junction reference required, potential calibration errors can be introduced if not properly wired using the correct type of wire that is matched to the thermocouple metals. The highest precision wire accuracy is the larger of ±4% of reading or 2ºF.

Resistance Temperature Detector (RTD)

general purpose

The resistance of a metal (wound wire or thin film) is temperature-sensitive.

Nearly linear over a wide range of temperature (-260 to 650º C). Good long-term stability. Very accurate over a wide range of operating temperatures. Can be utilized as averaging sensor. Little need for recalibration of the sensing element itself, although transmitters associated with them typically require some calibration to correct for thermal drift and other effects.

More expensive than a thermocouple or thermistor. Though often applied in HVAC applications without it, they require lead wire resistance compensation or a transmitter at the RTD to allow the full benefit of the accuracy and stability of the technology to be realized. A transmitter also can add inaccuracy into the signal. Subject to moderate self-heating.

Thermistor

narrow span application like space temp

The resistance of a semiconductor is temperature-sensitive. Typically the relationship is inverse; i.e. the resistance decreases as the temperature increases. Most of the other temperature elements have a direct relationship between their output and temperature.

High sensitivity (-80 to 150ºC). Large resistance compared to a RTD, so lead wire resistance errors are negligible. Low cost. Good for point sensing applications requiring high precision over a limited range.

Non-linear beyond small range. Self-heating due to high resistances can decrease accuracy, not as interchangeable as other technologies. Higher tendency to drift over time than some other technologies.

Integrated circuit

Tend to be OEM applications since not as many systems set up to take it as input

The voltage-current relationship of solid-state devices (diodes, transistors) is temperature-sensitive.

Linear high level output at a low cost. Can be easy to interface with other electronics.

Temperature measurement range is smaller than thermocouples or RTDs, but adequate for most HVAC applications. Subject to self-heating.

 

3.5.1.1. Space temperature

Accuracy

± 1ºF to ±1.5ºF

Installation

Locate the sensor away from electronics and other heat sources. Avoid locations where air diffusion may be stagnant. Also avoid exterior walls and stud cavities that connect to plenum floors since infiltration and surface wall surface temperatures can affect what the sensor is reading. It may be necessary to completely insulate the mounting box or mount the thermostat on an insulating pad. It is also important to make sure that small jets of air created by infiltration or pressurized air from a plenum exiting through the stud cavity and hitting the sensor are eliminated.

Calibration

Calibrate sensor if complaint or troubleshooting calls warrant and/or based on sensor drift specifications from the manufacturer. For instance a lower quality thermistor may show a drift of 0.1°F per year, which could add up over 5-10 years. If a space never seems to come under control, it’s probably worth going and taking a look. You may discover that the sensor is accurately reporting what it is seeing, but its being influenced by a draft, infiltration or cold surface.

 

There may be some spaces that require a closer specification, like an operating room or a clean room. In general, use thermistors or RTDs, with or without transmitters. The lowest cost approach is typically a thermistor without a transmitter, but the long-term stability of an RTD may make it a better choice.

3.5.1.2. Duct temperature

Accuracy

± 0.5ºF to ±1.5ºF

Installation

Higher velocities will improve response time due to improved convective heat transfer characteristics. Averaging sensors need to see uniform mass flow rates to reflect a true average temperature.

Calibration

Sensor error leads to energy waste. Calibrate per the manufacturers recommendations, or at least annually if no standards exist.

 

If it is an averaging sensor, the practical tolerance is 1.5ºF. See the averaging sensors discussion below. On large systems, tighter tolerances are desirable because the energy waste from small errors can be considerable over time.

3.5.1.3. Averaging sensor applications

Accuracy

± 1.5ºF

Installation

For large duct cross sections with the potential for stratification, it is important to use an averaging sensor that fully traverses the duct. Mixed air plenum temperatures can vary by 5-10ºF in well designed, well commissioned plenums and up to 60 ºF in a poorly designed and commissioned system.

Calibration

See calibration discussion on multiple averaging elements below.

 

Averaging sensors are necessary when air is not adequately mixed and stratification occurs. There can be temperature gradients in both directions, although for most systems, one will predominate due to geometry issues. Averaging sensors can consist of a string of thermistors or RTDs wired in series. Averaging RTDs can also be constructed by stringing the resistance wires back and forth inside the wire casing to form a length of resistance. Even when using the longest versions of the averaging sensors (about 25 feet), large ducts may not be adequately traversed, especially if temperature gradients exist in multiple directions. Therefore, it is desirable to average a number of the averaging temperature sensors to increase the accuracy of the reading. As a rule of thumb, 1 foot of averaging sensor should be used per four square feet of duct cross sectional area.

There are three main ways to generate an average signal from multiple temperature sensors:

·       Software averaging A number of RTD or thermistor averaging sensors can be hardwired or transmitted using current loops to a controller and the inputs can be averaged in the controller software. An advantage to this method is that all the input temperatures are displayed separately and shown as an average. Disadvantages include the added cost of transmitters for each averaging sensor and the additional I/O points used on the controller.

Figure 3.5 Typical Function Module

 (Image courtesy of Eurotherm Controls)

·       Independent signal conditioner Function modules exist that allow customized process controls to be built. An averaging function module takes multiple temperature input resistances and outputs one current signal. This strategy can cost less than adding a transmitter for each input.

·       Series/parallel resistance network Averaging can be accomplished by the sensor connection configuration, taking advantage of the way parallel and series resistances add. If four resistance temperature sensors are used, two pairs should be connected in series, then the pairs connected in parallel. The same principle applies to resistor networks of nine (or sixteen) resistors, with either three (four) sets of three (four) resistors in series wired in parallel. This method should only be used for applications where the temperature each resistor measures varies over approximately the same range. The series/parallel resistance network is a low-cost way to average sensors, with only one input to controller. One drawback to this method is that it is not widely used and therefore it may not be straightforward.

Field calibration of averaging sensors:

1   Use a data logger with 4-8 matched sensors for calibration. The sensors can be matched from the factory or matched to a common reference point. If they are not matched, then the offsets should be documented and applied to the readings.

2   Place the sensors to represent the entire averaging element location.

3   Create a uniform, non-stratified temperature profile by using 100% return air or a fixed setting on the preheat coil with non-varying load.

4   Let the temperature stabilize and mathematically average the data logger measurements. Compare the data logger and averaging element values.

5   Perform a two-point calibration, or a single point calibration to a temperature near the operating temperature. If the averaging sensors have a calibration problem, the element can be replaced if it is excessively inaccurate or the slope and intercept at the transmitter can be tweaked.

3.5.1.4. Outdoor air temperature

Accuracy

± 0.5ºF or better for temperature

Installation

For an accurate reading, the outside air sensor should be installed on the north side of the building, shielded from the sun. The sensor should be located in free air that is not heated by exhaust air or heat from the building or roof. Radiant effects can be significant and should be guarded against. On some sites, it may be desirable to experiment with data loggers placed at various locations to select the best location for this sensor.

Calibration

Calibrate per the manufacturers recommendations, or at least annually if no standards exist.

 

Outdoor air temperature measurement can significantly impact building energy consumption since many operational decisions are based on this temperature. Sensor error can affect the economizer cycle and reset schedules. The outdoor air conditions can also be used in degree-day and enthalpy calculations.

3.5.1.5. High temperature applications

Accuracy

± 2ºF

Installation

Use the thermocouple wire well within its temperature rating.

Calibration

Calibrate per the manufacturers recommendations, or at least annually if no standards exist.

 

For many high temperature measurements, such as boiler flue temperature, accuracies in terms of plus or minus several degrees are adequate. There are instances however where tight accuracy is important for reasons of product quality control or other factors, and accuracy specifications of a degree or less are required. In these cases, RTDs and RTDs coupled with transmitters generally offer the best solution. Thermocouples are economical and adequate when accuracy is not as important.

3.5.1.6. Critical control, monitoring, and billing applications

Accuracy

± 0.1ºF to ± 1ºF depending on the application and how the measurement is taken and used. Relative calibration may be more important than absolute accuracy, so matched sensors are effective. A sensor that can measure accurately to one-tenth of a degree needs to be coupled with a system capable of being controlled to that level.

Installation

Need to minimize other sources of error like installation errors, transmission errors, etc. An RTD that looks really good can be made less accurate by hooking it up to a transmitter. Using it directly with lead length compensation may be a better option and may drive the decision about what controller or control system you are going to use.

Calibration

Calibrate per the manufacturers recommendations, or at least annually if no standards exist.

 

Many consumption calculations rely on differential temperature measurement as a part of the energy consumption formula. In these instances, the accuracy of the sensor can have a significant impact on the calculated result. There is little tolerance for error if the data will be used for performance verification and diagnostics, or if billing is based on these measurements.

For instance, consider two sensors that have ± 1ºF degree accuracy and are to be used in a tons calculation. Assume that there is a known constant flow of through the chillers and one sensor reads one degree high and the other reads one degree low. For an actual chilled water temperature differential across the chillers of 12ºF, the inaccuracy of the temperature sensors represents 17% of total system capacity.

3.5.1.7. Water temperature differentials 20ºF or less

Accuracy

± 0.5ºF or better, or as driven by the needs of the application. This is another area where relative calibration may be more important than absolute calibration.

Installation

For any temperature differential, a matched pair of sensors will eliminate error based on the calibration of one sensor relative to the other.

Calibration

Install calibration wells. Calibrate per the manufacturers recommendations, or at least annually if no standards exist.

 

A 1ºF error on a chilled water system with a 10ºF differential represents 10% of system capacity. Less accuracy than this can make the measurement useless.

3.5.1.8. Water temperature differentials of over 20ºF

Accuracy

± 1ºF or better

Installation

For any temperature differential, a matched pair of sensors will eliminate error based on the calibration of one sensor relative to the other. As the temperature differential increases, the effect of unmatched sensors decreases.

Calibration

Install calibration wells. Calibrate per the manufacturers recommendations, or at least annually if no standards exist.

3.5.2. Humidity

The two most common humidity sensing technologies are bulk polymer resistive and thin film capacitance. Lithium chloride salts are an older technology, and chilled mirror hygrometers are often used for precise humidity control for labs and clean rooms.

Table 3.2 Humidity Measurement Technologies

Technology

Function

Advantages

Disadvantages

Bulk Polymer Resistive

Measures change in resistance as the polymer absorbs or emits molecules of water.

Surface contamination will not affect accuracy. Some sensors are interchangeable without calibration.

Variable accuracy with changes in temperature.

Thin Film Capacitance

Measures change in capacitance between a thin film polymer and electrode due to change in relative humidity.

High linearity, low hysteresis, long-term stability, wide temperature range. Some sensors are interchangeable without calibration.

Variable accuracy with changes in temperature

Lithium Chloride Salts

Saturated salt solutions produce a known RH. Often used for calibration of sensors.

One of the older technologies; was available before capacitance and thin film technologies were developed and was actually one of the more cost effective approaches

Sensitive to dirt and moisture contamination, expensive, requires frequent calibration.

Chilled Mirror Hygrometer

Mirror chilled through thermoelectric cooling until dew/frost forms. The dew point temperature is measured by an RTD.

Inherent accuracy since measure dew or frost point temperature directly. Long-term stability.

Expensive relative to some of the new technologies. Contaminants on mirror reduce accuracy, so cleaning is required. Some of these sensors have a self-cleaning cycle.

 

Accuracy

± 5% of reading represents a compromise between accuracy and first cost. For critical applications (clean rooms and operating rooms), ± 3% or better can be achieved. Obtaining a higher level of accuracy can increase the sensor cost from hundreds to thousands of dollars.

Installation

Outdoor air humidity should be picked up at the same location as the outdoor air temperature.

Calibration

The humidity sensor needs to be immersed in salts or a humidity reading must be taken using a sling psychrometer in the duct location. This is usually easiest with a sling psychrometer that has a built in fan so you can set it in the duct and let it run. Swinging a sling in a small duct can be tricky, and an access panel is necessary. Salts can be used for calibration by pulling the transmitter out of the duct, but it needs to be mounted so this is possible (flex connection, big enough probe opening to retract probe and turn it with the flex attached, etc.). Otherwise you need an access panel.

3.5.3. Pressure and Flow

Table 3.3 presents a variety of pressure measurement technologies for air and water. The capacitance, strain gauge, and piezoresistive technologies compete with each other since they offer accuracy at a reasonable cost in a variety of levels of quality. LVTDs compete with the lower end versions of these technologies. For all of these devices, pressure transients in the air or water may cause erroneous readings. For example, pressure pulses from the pumps can create noise in the pressure measurement. A pressure pulse can also be created from a door opening that is near a diffuser location and near the duct static pressure measurement point. In this case, the pressure pulse can cause the control loop to hunt.

Table 3.3 Pressure Measurement Technologies

Technology

Function

Advantages

Disadvantages

Velocity Probes

Example application:

isolation rooms

Measure pressure differential based on velocity through a tube across the pressure difference.

Can accurately and repeatable measure to thousandths and ten thousandths of an inch water column; good for low pressure applications like clean rooms, building pressure control and isolation room monitoring or control where pressure differentials in the range of 0.05 in.w.c.or less must be accurately measured and maintained.

More expensive that other technologies, sensing line length critical since there is active flow through the lines - long lines could impact the accuracy of the input because the pressure drop through them would affect the flow and flow is the indicator used to measure pressure differential.

Manometer

Example application:

filter pressure drop

Measure the change in height of a column of liquid between a reference pressure and the pressure being measured. Typically a manual instrument used to calibrate and verify other instruments.

 

 

Capacitance

Pressure changes cause a change in capacitance between a metal diaphragm and an electrode. The capacitance is measured and used to generate an output signal.

Low hysteresis, high repeatability, high resolution, fast response, and the ability to measure low pressures.

Must be zeroed routinely since temperature or change in physical position can affect performance. May not be as rugged as some other technologies.

Strain Gauge

The deflection of a diaphragm due to pressure change is measured by strain gauges.

High accuracy, long-term stability; very tolerant of extreme over-pressurization in some packages.

Strain gauge bond with diaphragm may degrade

Piezoresistive

A pressure change causes the resistance of a semiconductor (solid-state chip) to vary.

Detects larger pressure differences than capacitive transmitters (greater than 5”). Withstands vibrations.

 

Linear Variable Differential Transformer (LVDT)

An electric output is produced in proportion to the displacement of a movable transformer core. Usually coupled to a bourdon tube to measure pressure.

High reliability since no mechanical wear/friction between the transformer core and coil. High resolution. Lower cost for a given accuracy spec as compared some other technologies.

Inherent nonlinearity of standard LVTDs is about 0.5% of full scale. Not as rugged or accurate as some other technologies.

 

Differential pressure-based flow readings are a function of the square of the flow (at 50% of full flow, the signal is 25% of full flow), so the signal becomes less accurate at low flows (high turndown). If differential pressure based measurements are being employed to measure flow, it is important to consider the magnitude of the available output signal that will be generated at the lowest flow point that the system can be expected see.

Smart differential pressure transmitters can get up to a 10:1 turndown, but another way to get better turndown is to hook up multiple transmitters to a flow sensor and then have the software pick the one that is best for the current flow range. For example, one transmitter is full scale at 25% of design flow, one is full scale at 50% of design flow, and one is full scale at 110% of design flow. The lower range transmitters must be rated to withstand the higher differential pressures (beyond their range) without failure.

The maximum output of the flow meter should be chosen at 5-10% above design conditions to guard against over-range errors at start-up and overload conditions. This sizing is important in flow measurement applications because the sensor ranges tend to be customized to the application rather than being defined by physics (like 0-100% RH) or being defined by standard ranges based on the HVAC process.

Table 3.4 Differential Pressure-based Flow Measurement Technologies

Technology

Function

Advantages

Disadvantages

Pitot Tube

Measures velocity pressure at a point. Flow is then calculated based on the velocity pressure. Typically used for manual measurements with an inclined or electric manometer or as a sensing probe for one of the pressure transmitter technologies listed in the Pressure Measurement section.

Fairly inexpensive but must be combined with some sort of indicator or transmitter to provide useful data and/or interface with a control system. Little need for calibration. Low pressure drop. Small access holes.

Small tube misalignment results in poor accuracy. Reliable measurement requires multiple pitot tubes. Flow output is a function of the square of the velocity pressure, so Turn down capabilities are not good. For example, 50% flow produces only a 25% signal.

Orifice

Pressure is measured before and at restriction in liquid flow (orifice plate). Velocity is calculated from the pressure drop using conservation of energy and mass equations. Combine with an indicator or transmitter to provide useful data and/or interface with a control system.

Lower first cost relative to the more exotic technologies.

Energy penalty from high pressure drop. Flow output is a function of the square of the velocity pressure. 50% flow produces only a 25% signal, so turn down capabilities are not good.

Nozzle

As above, but restriction is a nozzle internal to the pipe, with no expanding outlet area for pressure recovery. Combine with an indicator or transmitter to provide data and/or interface with a control system.

Lower first cost relative to the more exotic technologies. Used for high velocity flow.

Pressure drop lower than an orifice plate, but higher than a venturi. (and same problems as far as turn down and signal go)

Venturi

As above, but restriction is a gradually narrowing diameter of pipe, followed by an expanding section. Combine with an indicator or transmitter to provide data and/or interface with a control system.

Lower first cost relative to the more exotic technologies. Nearly returns the flow to its original pressure. High accuracy, used for larger pipes.

(and same problems as far as turn down and signal go)

Insertion Tube

Measures velocity pressure

Combine with an indicator or transmitter to provide useful data and/or interface with a control system.

Low pressure loss. Often, this probe can be installed in an operating system via a hot tap, thereby eliminating a shut down.

Flow output is a function of the square of the velocity pressure. 50% flow produces only a 25% signal, so turn down capabilities are not good.

 

Table 3.5 Non DP-based Flow Measurement Technologies

Technology

Function

Advantages

Disadvantages

Hot Wire Anemometer

(Airflow)

The heat lost due to airflow across a heated sensor is correlated to air velocity. Some technologies measure the heat transferred from one sensor relative to another to detect flow rate and direction

High frequency response. Good for low flow measurements. Good for high turn-down applications.

Fragile, only used in clean air flows. Calibration is required frequently. High cost.

Rotating Anemometer

Vanes spin and velocity is measured. Typically used for manual measurements.

Portable, low cost, can detect and measure low flow rates.

Small tube misalignment results in poor accuracy. Reliable measurement requires multiple readings

Vortex Shedding Sensor

Example application:

Steam flow

Measures the frequency of pressure spikes due to vortices made by inserting the vortex probe in the air or liquid stream. The frequency of the pressure spikes is translated into fluid velocity.

Highly accurate. High turn down ratio.

For steam, mass flow varies with line pressure. High first cost relative to some of the other lower accuracy, lower turn down technologies. difficult to field calibrate. Needs to be sized and may be smaller than line size to provide the necessary turn down capability.

Critical Flow Nozzle

The gas (air, steam, oxygen, etc.) reaches the velocity of sound through the nozzle, which cannot be exceeded. The mass flow rate is proportional to the pressure upstream of the nozzle since the velocity through the nozzle is fixed. Must be combined with an indicator or transmitter to provide useful data and/or interface with a control system.

Requires only two measurements (inlet static pressure and temperature).

Nozzle and diffuser result in 10% reduction in upstream pressure, so used in steam transitions to lower pressure applications.

Positive Displacement

Applications include water metering such as for potable water service, cooling tower and boiler make-up, and hydronic system make-up. Positive displacement meters are also used for fuel metering for both liquid and gaseous fuels. Common types of positive displacement flow meters include lobed and gear type meters, mutating disk meters, and oscillating piston type meters. 

high accuracy at high turndown

high permanent pressure loss

close tolerance required between moving parts of positive displacement flow meters, they are sometimes subject to mechanical problems resulting from debris or suspended solids in the measured flow stream

High cost. Loss of flow if the meter fails or seizes up.

Turbine

Flowing water rotates a turbine in the pipe, speed of rotation is detected and related to the velocity of the fluid.

Full bore turbine for critical flow measurements. Good turndown. Often, this probe can be installed in an operating system via a hot tap, thereby eliminating a shut down.

Pressure loss, but less with insertion type. Reduced accuracy and susceptible to damage with debris in water. Bearing life problems.

Transient Time Ultrasonic Flow Meter

Ultrasonic waves are sent with and against the direction of fluid flow, measuring the time difference for the wave to travel.

Non-invasive (strap-on or weld on) or in-line meter. Very good turn-down capability. Often, this probe can be installed in an operating system via a hot tap, thereby eliminating a shut down.

High first cost relative to some of the other lower accuracy, lower turn down technologies. Used to measure clean water flow. Errors in readings when air trapped in pipes or with dirty water. Difficult to field calibrate.

Doppler Ultrasonic Flow meter

Sound waves are reflected back to the sensor from solids or bubbles in the fluid. The echoes return at an altered frequency proportionate to flow velocity, which is measured to calculate flow.

Non-invasive (strap-on) or in-line meter. Liquids measured must contain solids. Often, this probe can be installed in an operating system via a hot tap, thereby eliminating a shut down.

High first cost relative to some of the other lower accuracy, lower turn down technologies. Difficult to field calibrate.

Magnetic Flow Meter

Magnetic induction: water (a conductor) moves through a constantly applied magnetic field and a voltage is induced proportional to the speed of the water.

In-line meter has high accuracy. High turndown ratio (30:1). Low maintenance.

High first cost relative to some of the other lower accuracy, lower turn down technologies. Difficult to field calibrate.

Target Flow Sensor

An arm with a disc at right angle to flow measures the force of fluid flow.

Can be used with dirty fluids. 20:1 turndown

 

 

3.5.3.1. High Pressure Applications

Accuracy

For flow metering applications, the sensor should be selected for ± 2% of full scale with full scale selected to match the meter output at 110% of the maximum flow.

Pressure transmitters for piping: ± 1% of full scale or better. The standard gauge ranges for pressure transmitters for water and steam systems and other relatively high pressure fluids are typically 30, 50, 100, 150 psi, so 1% of full scale as a rule implies accuracies of 3-15 psi depending on the range. Ultrasonic, magnetic flow, and vortex shedding meters can achieve ± 5% of the flow reading or better with turndown ratios of at least 15: 1 to 30:1. Vortex meters can also be used on steam and compressed air systems for superior turn-down capabilities.

Installation

The upstream and downstream lengths of straight piping are important to consider for accurate flow measurement. During design, the piping layout should incorporate the manufacturer requirements for straight run in and out of the flow sensor. Installing the sensor just after and just before tees or elbows will result in an inaccurate reading.

For field-welded meters with precise sensor alignment requirements (ex., ultrasonic flow meters), adjustment jigs that are specific to the meter should be used for installation. These jigs usually need to be provided by the supplier. If the ultrasonic meter alignment is correct, then any sensor error should be related to the electronics and transmitters, which can be removed and returned to the factory for re-certification. A venturi meter or orifice meter usually does not have a special jig.

Calibration

Calibrating water flow meters is very difficult, since it is not always practical to install the flow meter in a location that provides enough room to add an additional calibration flow meter at the location. Using a calibration technology more accurate than the installed flow meter can also be difficult. In many cases, factory calibration is the only practical option. Differential pressure measurement across a chiller, pump curves, or calibrated balancing valves can be used as a cross-check, rather than actually calibrating to the device. For a precise flow meter, like some of the electronic technologies, using differential pressure for a calibration reference would probably decalibrate the flow meter.

One possibility is to make provisions for the installation of a portable, strap-on ultrasonic meter in series with the permanent meter. A section of pipe must be provided with insulation that can be removed and replaced easily. There are vapor sealing issues that need to be addressed with lines that condense the insulation is removed.

A strap-on flow meter is also a viable option for a system where good readings are necessary for troubleshooting or set-up, but the facility cannot justify or afford the cost of a permanent meter. Design place places for a strap-on meter, then furnish one portable meter for the owner after the initial balancing.

Five-Valve Manifold

Most differential pressure sensor locations serving high-pressure fluids require a field-fabricated or factory-built five-valve manifold. Differential pressure sensors and transmitters can be damaged, decalibrated, or destroyed by imposing the static pressure of the measured medium upon the differential pressure sensor and transmitter. There are many styles of five valve manifolds, so it is important to understand the function of each valve and install one that performs the exact function you need.

The equalizing valves, opened before the process flow valves, serve to bring the process pressure onto both the high and low sides of the differential pressure sensor at the same time. The equalizing valves are then closed and the vent between them is opened to verify that there is no cross flow bypassing the transmitter. Figure 3.6 illustrates a typical manifold and how it is used.

Figure 3.6 Typical 5-Valve Manifold and Schematic

Opening the equalizing valves (A) prior to opening the service valves (B) guarantees that the sensing element of the transmitter will see little if any differential pressure even if one service valve is opened before the other. Closing the equalizing valve after opening the service valves subjects the sensing element to only the differential pressure signal. Opening the vent valve (C) proves that there is no cross-flow in the equalizing line, which could throw off the measurement. Some manifolds include another pair of valves in tees next to the transmitter connection to allow air to be cleared from the lines and allow a calibration meter to be connected in parallel with the process sensor.
(Images courtesy of the Hex Valve web site)

Snubbers

Snubbers should be installed at all pressure sensor and pressure transmitter locations to dampen pressure pulsations in order to provide stable readings and extend the life of the pressure-sensing element. Snubbers should be installed after the isolation valve and ahead of the calibration and sensing connections. Several different styles are available, as shown in Figure A. The style selected depends on the application.

Figure 3.7 Pressure Gauge Snubber Styles

 

 


  

Some styles dampen pulses via a fine meshed porous filter (left). Others use interchangeable pistons that can be selected to tune the snubber to the process (center). Others have a field adjustable screw that allows the dampening effect to be adjusted (right). (Images courtesy of the Ashcroft and Weiss Instruments web sites)

Indicating transmitters

Indicating transmitters eliminate the need for an independent indicator, saving first costs and ensuring that the reading the operators get at the equipment is the same as the control system is getting. You still need to install second well or gauge for calibration purposes.

3.5.3.2. Low Pressure Applications

Accuracy

Air flow can be measured with transmitters designed for much lower ranges, like 1/10th of an inch or even better if some of the velocity-based flow measurements are used. The accuracy necessary for airflow measurement is dependent on the need for turndown and/or absolute accuracy, specific to each application. For example, supply and return fan tracking in a VAV system with high turndown requires sensors that can measure a wide range of flow rates. To determine the absolute accuracy required, consider what would happen to system control if the flow was measured 10% low on the return and 10% high on the supply. This could be unacceptable for an operating room or laboratory with strict pressurization requirements, but satisfactory for office space.

VAV box flow is another common application for flow measurement. Most controllers use a velocity pressure signal as an indication of flow. Thus, due to the square law effect (pressure drop varies with the square of flow), the low end of the flow measurement rating (i.e., 1000 cfm for a 1000-2000 cfm VAV box) is generally set by the ability of the flow measuring element to generate a useable signal when applied at the connection size associated with the VAV unit. There are several important design and commissioning implications to this, which are discussed in Chapter 19: Terminal Equipment.

Consider the magnitude of the available output signal that will be generated at the lowest flow point that the system can be expected see. This could be equivalent of the sum of the minimum flow rates of all of the terminal units served by the system if all zones on the system operate on the same schedule. But, for a system where the schedules were applied to the zones so that an unoccupied zone was driven to 0 cfm during the unoccupied cycle while other zones remained active, it also could be the sum of the minimum flow rates of the smallest possible collection of occupied zones.

Installation

There are several ways to improve the signal with high turndown ratios:

·       Install the flow sensors in a section of duct that is sized to provide the required velocity pressure at the minimum anticipated flow. This section could have significantly higher velocities and friction rates than the bulk of the duct system. Careful design and attention to regain considerations can minimize the impact of this approach on overall system static, but there will be some impact.

·       Install flow sensors designed for the eye of the fan wheel. Velocities in this area tend to be higher than the trunk duct velocities and several manufacturers have equipment configured specifically for this type of application. Figure 3.8 is one such example.

·       Use “smart” sensor technology. This approach takes advantage of digital technology to more finely resolve the velocity pressure input and provide better resolution of the signal, often in the range of 8:1 or 10:1 vs. the 4:1 or 5:1 possible with conventional differential pressure cells.

Figure 3.8 Flow Sensor Designed for the Fan Wheel Inlet

(Image courtesy of Tek-Air)

·       Install multiple sensors on the flow-measuring element ranged for different percentages of the output signal and select the best sensor for the current flow condition. This allows one differential pressure sensor to be optimized for the higher flow for instance from 50% to 100% while another is optimized for the flow range from 25% to 50% and a third is optimized for 10% to 25%. (This approach can also be used for differential pressure based water or steam flow measurements.) In this application, all sensors must be capable of withstanding the differential associated with the full flow indication (which will be over-range for the sensors optimized for the lower flow rates) without damage or calibration problems. This is usually more of a problem with water and steam systems than air systems.

·       Use an alternative technology that does not required differential pressure measurement for systems with large turn-down ratios. Hot wire anemometer based sensors are an example of this approach for air systems. Ultrasonic and vortex shedding flow meter technologies are a similar example for water and steam systems.

Calibration

Field calibration of an air flow meter is easier than a water flow meter because you can traverse the duct at a good spot with good technique and instruments and probably get a reasonable cross-check.

Filter Pressure Drop

Filter pressure drop can be automatically detected and compared to a set point to determine when the filter should be changed. This point is called filter status. This information can make filter maintenance easier on a constant volume system, but is not critical as long as some form of filter pressure drop indication is provided (most codes require filter pressure drop measurement either directly or by reference to ASHRAE which cites it as good practice). Installing a Photohelic™ or equivalent device that combines pressure drop indication, set point indication, and a contact closure at set point meets code requirements. Since the flow in a constant volume system is constant, the pressure drop of the filters will not vary with load on any given day, and the pressure drop observed by the operators on rounds will be a good indication of the condition of the filters.

Detection of filter status is required for VAV systems because the filter pressure drop will vary with load and flow. The peak load and flow condition, and therefore the peak pressure drop, may occur at a time when the operators are not observing the filter pressure drop indicator. Failure to realize that the filters are exceeding their rated pressure drop can lead to structural failure of the filter or blow-through of the filter media. Additionally, an unacceptable loss of system capacity can lead to performance problems, IAQ issues, and loss of efficiency due to soiling of the heat transfer surfaces downstream of the filters.

3.5.4. Electrical

Electrical measurements are useful to indicate proof of operation, measure efficiency, and measure power demand and energy consumption.

Table 3.6 Electrical Measurement Technologies

Technology

Function

Notes

Current Switch

Binary signal based on the current flow relative to a limit value.

Inexpensive way to measure proof of operation based on a fairly direct indicator (motor current) that can detect drive failures if properly applied and adjusted. Needs to be matched and adjusted to the system operating characteristics to reliably detect belt and drive failures on variable flow systems. Some switches will not work with VFDs due to line harmonics.

Current Transformer (CT)

Measures a small voltage that is proportional to the current flowing through the device. Used with current transmitter to measure current.

Some designs require a shut down to install the sensing element since it is a continuous loop that must go around conductors and bus bars. Need to be protected from an open circuit condition.

Potential Transformer (PT)

Steps down the voltage into a range suitable for sensing. Used with voltage transmitter to measure voltage.

 

kW Transmitter

Used in combination with PT and a kW transmitter to measure power.

Three phase power measurements require multiple CT and PT. Can get three phase power transmitters or implement a separate transmitter for each phase and calculate the total power in the controller. Any kW reading integrated to get kWh.

kW Pulse Train

Pulse rate proportional to kW typically furnished by the utility

Often averaged over a 15 or 30-minute interval. Can be integrated to determine consumption (kWh).

kWh Pulse Train

Pulse represents the consumption of a fixed number of kWh. Keep track of kWh pulses.

 

 

3.6. Point Structure and Interface at the BAS

Chapter 2, Control System Design Process, presents the components of the design of control systems. This section builds upon Chapter 2 by focusing on aspects of interfacing the control and monitoring points to the BAS that often show up as commissioning and operational issues.

The following recommendations for spending time upfront planning the point structure and user interface to the building automation system (BAS) ultimately saves time by reducing the changes that need to be made during the final phases of construction. If the control contractor needs to make changes to point names and other fundamental database parameters after the initial programming is complete, it is often necessary to modify a significant portion of the software associated with the system’s operating logic because the point names show up in the software sequences as variables. In some systems, the controls contractor must literally go through all of the programming in the system and modify the name of the point at every occurrence. This can be time consuming and introduces opportunities for errors, which will require that the software be re-commissioned to verify that no new bugs were introduced by the changes and that all changes were picked up. The following BAS point structure and interface issues help make the BAS more useful in controlling and monitoring system performance:

·       Point Naming Conventions

·       Virtual Points

·       BAS Settings

·       Programmable Alarms

·       BAS Graphical User Interface

·       Point Trending

3.6.1. Point Naming Conventions

Most current technology systems have both a name and a descriptor associated with each point. Generally, the point name is short 8- to 16-character text string that is stored in the memory of the controller. In most systems, you can trace its origin to early generations of the controller architecture where the ability to store and handle text and other data strings was limited by the addressing capability of the microprocessor. The point name provides a way to link the raw data associated with a point’s physical location on the input/output terminal strip with the real world data represented by the point. As a machine, the microprocessor would have no problem working with the raw numbers that represent the wire termination location on its input/output termination board. But since humans needed to work with the machine the point name provided a way to represent this data in more human terms. To be truly useful, the name had to be coded to represent both the system and information associated with the point. This representation can be somewhat cryptic since 8 or 16 characters can be used up very quickly when generating a name that is decipherable, but gives a feel for what the point represents, especially on larger systems where the name might need to reflect multiple air handling units, each with multiple points of a certain point class such as temperature.

As microprocessing technology evolved to the point where larger text and data strings could be handled, most manufacturers added point descriptors to the point’s database. In most instances, the descriptor consists of a 32- to 64-character data string that enhances the point name to provide additional clarity. Since memory is still a fairly precious commodity at the controller level, the relatively memory intensive point descriptors are often stored in memory at the host computer, operator work station, or at a supervisory controller on the network, upstream of the controller they are associated with. The descriptors are then put together with the point name by the system’s operator interface software to allow them to be viewed with the points lists, graphics and other information presented at the operators terminal (host computer). If a technician is working directly with the microprocessor at the controller via a laptop computer or other programming tool connected directly to the controller’s programming port, they will probably be working only with the point names unless the programming too/ has the descriptors in its database and has software running that can make the association. Thus, it is advisable to code in the point names in some manner that allows them to convey the basic information associated with a point, such as the system it serves and the data it represents.

As you may surmise from the preceding discussions, point names and descriptors represent a powerful tool in making the control system useful to the operators and clarifying the information it presents. Information that is clearly presented is more likely to be responded to in a timely fashion and interpreted correctly, especially during the faced paced events associated with some sort of operating emergency. The ability to correctly interpret the data presented is also critical to the efficient operation of the facility and its equipment.

Many systems have default parameters that are used as placeholders for the memory locations associated with the point name and descriptor values. If the owner and designer to do not become proactive in directing the control contractor in the point naming conventions that should be used for a project, the programming technicians may simply use these defaults since it can save them time and labor cost during system’s installation and programming process. When this occurs, a great deal of the flexibility utility that can be provided creatively using these data fields will be lost, probably for the life of the system since changing this information (especially the point names) subsequent to program development can result in the need to go through all of the software in the system as mentioned previously.

Ideally designer should set the general criteria that will be used for naming points in the project’s specifications. This may include a list of point naming conventions already in use on a site where the project is an expansion of an existing facility. For new projects, it might include a standard list that the designer has developed from another project and past experience. In any case, the general criteria provided in the contract documents can alert the contractor to the need to coordinate these requirements into the database and programming development cycle for the project. As a part of that process, the owner and, ideally, the future facility operators, should become involved in finalizing the naming conventions that will be used for the project. Typically, this can be accomplished as a part of the shop drawing review process. The following paragraphs describe considerations for point naming that should be addressed in the specifications and finalized prior to programming the system.

·       Label each point with a unique name Ultimately, each point must have some unique name that distinguishes it from all other points on the system. Some systems accomplish this by arranging their database so that quite literally, each point has a unique name. However, this can become cumbersome on systems with large point counts, and many systems now organize their data in some sort of nested file structure similar to that used by with Windows operating system. Under this approach, each controller or system device has a unique name and the points residing with in that device have a unique name relative to each other. But the structure of the database allows different controllers to have identical point names. For instance, controller RTU1 may have a point named RTU_LAT with a default descriptor ANALOG INPUT and controller RTU2 may have a point with the same name and descriptor. From a programming standpoint, this approach has some significant advantages because, for projects with identical HVAC systems, programming can be written and debugged for one controller and copied to all the others serving an identical HVAC system. The system knows the points are different because they are in controllers with different names. But, an all-points log on such a system often looks like this:

      RTU_LAT       ANALOG INPUT        57.4

      RTU_LAT       ANALOG INPUT        56.2

Obviously, data presented in this manner is not very useful to the operators. This can be especially frustrating when alarms come in. Having an alarm printer spit out:

      HI ALARM      SPACETMP     ANALOG INPUT        78.0 F

does not give the operating staff much to go on if they are dealing with a system that has hundreds of terminal units. Contrast this with:

HI ALARM      Rm_1201          Space Temperature       78.0 F

·       Use consistent naming conventions The temperature at the supply fan discharge could be called SF1_LAT (leaving air temperature) or AHU1_DAT (discharge air temperature). To avoid confusion, always label like points with similar names.

·       Number points for easy sorting If there are ten or more of something that is numbered, then number them 01, 02, 03 … 10. This will give a better sort because most systems will sort 1 through 10 as 10, 1, 2, ..., but would sort 01 through 10 as 01, 02, 03,...

·       Take advantage of upper and lower case letters Most systems can accept upper and lower case letters for text strings. This can be used to advantage to make things easier to decipher, especially point names where a lot of information is being crammed into a short character string. Consider the following possibilities for the point name for the discharge air temperature associated with air handling unit 1 on a system that allows 8 characters for a point name.

AHU1_DAT - A good starting point, but may not sort well if the project has (now or in the future) more than 10 air handling systems.

AHU01DAT - This solves the sorting problem, but many would say it was less easily read, especially if scanning it quickly in a long list of similar, nearly identical names.

AHU01_DT - This may improve readability, but now 5 characters are tied up in describing the system leaving only two characters to describe the data.

Ah01_Dat - An option that uses upper and lower case letters and a shorter system description to convey the same data in a potentially more readable text string.

Obviously, this is somewhat subjective, but what matters is that the project specifications pave the way for involving the operating staff in the development of the way that the system data presentation is handled and that they be informed about what the possibilities are prior to software and database development. This will allow the operating staff to work with the controls technicians to tailor the system data to their needs.

·       Use engineering units to describe the function of the point For example, units can be used to differentiate between a command point (on/off) as compared to a point that allows the local control system to start or stop the equipment (enabled/disabled). For instance, the control signal commanding the fan to 50% speed should be labeled to distinguish it from the feedback signal indicating that the fan is actually running at 50% speed.

·       Consider the future The current project may only have 9 terminal units now, but a future remodel or expansion could add several terminal units. A system set up initially numbering the units 1 through 9 vs. 01 through 09 would encounter sorting problems when units 10 and above were added at a later date. While not a major issue, this sort of thing can be an annoying problem for operating staff and can result in frustration and misinterpretation of data in an operating environment. A little forethought during the early phases of programming can eliminate a lot of frustrations in the future.

3.6.2. Virtual Points

Virtual points do not physically exist, but are points that exist in the control system software (see Section 2.4.2 Points List for more details). Virtual points include calculated points, setpoints, and other parameters. The purpose of virtual points is to provide more information to the operator than the raw data can provide alone. For example, chiller load can be continuously calculated by the control system using the chilled water supply and return temperatures and flow rate raw data.

(This is also in BAS Display section). Setpoints and tuning parameters should be virtual points to allow them to be manipulated by the operating staff without having to open up and modify programming code. But often, without specific instruction, the technicians programming the system will simply enter the parameters into the program code as hard values instead of variables that reference virtual points outside of the program for their value. This practice saves some programming time and effort, and, since the programmers are intimately familiar with the programming language and the programs they are writing, there may be little if any utility for them in having the ability to modify parameters without modifying the actual program code. However, the operating staff needs to be able to quickly modify setpoints and tuning parameters when the need arises. With virtual points provided for setpoints and tuning parameters, the operator can make changes without changing the hard coded control algorithms, which is much less risky than making a change to the programming code itself.

The following points are commonly included as virtual points:

·       Calculated points such as chiller load, kW/ton

·       Temperature and humidity setpoints such as supply air temperature, mixed air temperature, or supply humidity

·       Static pressure setpoints like duct static pressure or building static pressure

·       Utility system setpoints like chilled and hot water temperature setpoints

·       Zone setpoints such as temperature and flow.

·       Changeover setpoints like the settings used to enable and disable the economizer cycle or allow humidifier operation.

·       PID tuning parameters for all control loops.

3.6.3. BAS Settings

Change of value (COV) limit parameter In most cases, the controller is working with real time values. These values are filtered for transmission to the operator terminal to keep minor variations from clogging system communications. Usually the COV limit parameter (or a similar function) controls this filtering. The value of the point will not be updated in the operator’s terminal until the change in measurement exceeds the limit. If the COV limit is set high, then the reading at the operator terminal may not be a reasonable representation of the actual system operation. For example, if the COV limit for a space temperature was set to 5ºF, the space temperature would not change at the operator terminal unless the temperature changed by more than 5ºF or the system was manually forced to read the temperature. COV limit parameters depend on the particular measurement.

User access Most systems allow you to control what level different operators can access based on a security level associated with their password. Less skilled operators can be kept from having access at a system level where a parameter could accidentally be changed. Senior staff should be able to access all levels as needed.

Usually, it is desirable to use the system security access features to control who can make changes to these parameters. For instance, junior grade operators may only be allowed to make changes to zone temperatures via the access level associated with their system password. Mid level operating staff may be allowed to change everything that the junior level staff can change and also change central system settings and change over setpoints. Access to loop tuning parameters and critical changeover settings may only be provided to the senior operating staff or the lead facilities engineer.

3.6.4. Programmable Alarms

High and low alarms can typically be programmed for any analog point, which can be useful if used properly and annoying if used improperly. Some systems allow a high and low anticipatory alarm in addition to the actual alarm settings. This feature can be very useful on critical parameters to give operators a “heads up” warning based on a trend towards an undesirable operating condition before things get too out of hand. For example, an alarm can be used to anticipate a freezestat trip.

Floating alarms indicate when a parameter is out of normal operating rang. For example, a floating alarm will indicate when the point goes a certain amount above or below a setpoint, even when the setpoint changes due to a reset schedule.

Another useful situation to alarm is when a piece of equipment is in manual override for a period of time. It is easy to forget that an override is in place.

Be sure that the alarms give the information you need, when you need it, for the manner in which you intend to operate your building. Consider these questions as you program alarms:

·       Should the alarm have a time delay? For example, a 30 second delay in the fan proof of operation alarm will avoid nuisance alarms with the time delay between when the fan is commanded on and when the status point proves it.

·       How will the alarms be prioritized by their importance?

·       What alarm message should be displayed? This can be useful for directing a response to the alarm and/or clarifying what is wrong.

·       Which alarms should call a pager? Does this requirement change with the time of day and day of week?

·       Should the alarm trigger a graphics screen to highlight the alarm condition? With an event with multiple alarms, multiple triggered graphic screens may clog the system communications and not allow the operators to see what is going on. Often, this problem can be handled with some programming that disables the alarm graphics during certain events. On larger systems, providing a separate alarm printer and alarm monitor dedicated to alarm graphics can alleviate this problem while providing enhanced operating and management capabilities.

Smart Alarms Programmable alarms that use information from multiple points and the system’s logic capabilities to alert operators to conditions of degraded system performance are called smart alarms. The alarms consist of short algorithms that detect problems through point comparisons. For example, if the preheat coil is active (there is a temperature rise across the preheat coil or the valve has been commanded open) and the economizer is not on minimum outdoor air, then an alarm could notify the operator that energy is being wasted. A smart alarm can alert operators to energy waste by indicating when the outdoor air temperature is below the current discharge air temperature setpoint but the cooling coil is active.

Smart alarms can have great benefit as a diagnostic tool, saving operators’ time by automatically detecting problems. As a management tool, the system can be programmed to segregate these more complex alarms to a separate report log directed to the lead operator or facility engineering staff. This practice allows the operating staff up to deal with the day-to-day operating issues and allows the management level staff to prioritize and direct the response to these more complex system performance related issues.

3.6.5. BAS Graphical User Interface 

A number of issues related to the way in which the BAS graphical user interface is set up can influence the usefulness of the system to the operators. Additionally, providing remote access to this interface is very useful to commissioning providers, operating staff, and the control system contractor, especially during the first operating year. The following issues should be considered for the user interface:

·       Readout accuracy The display indication should reflect the known accuracy of each measurement device. For example, if a temperature sensor has an accuracy of ± 1ºF, then the temperature should be displayed in increments of ± 1ºF or ± 0.1ºF at the most.

·       Setpoints and Schedules Setpoints, occupancy schedules, reset schedules, and high/low limits should be variables that can be adjusted without altering the code. To do this, the controls programmer must create system variables rather than hard coding the information. These points should be easily accessed from the graphical interface for editing and comparison to measured values. There are many ways to accomplish this including a table for all systems on the site where the values are displayed and can be modified, or sliders or setpoint adjustment knobs on the system graphic to name a few. This is another area where tailoring the presentation to the tastes of the operating staff early in the programming process can result in a more useful and used system.

At a high security access level, the loop tuning parameters should also be adjustable at the user interface.

·       Operator workstation graphics In general, the design of the operator workstation graphics should consider two key elements:

1    Allow for rapid, easy to understand and implement, penetration of the building’s database.

2   Present the information retrieved in a manner that is easy to understand and interpret by the operating staff.

The graphical user interface for the project can be structured to accomplish this in a variety of ways. The best way to accomplish the intended functions will vary from project to project with the tastes of the operating staff. If the development can be tailored to reflect their tastes, the system will seem user friendly and be more useful to them. If the system is more useful to them, it is much more likely that the building will be operated at the peak of its performance and efficiency capabilities and that the design intent of the systems will persist. The following items should be considered when specifying and developing the system graphics:

·       Use Building Floor Plans to Guide Database Penetration Using a graphic of the floor plan is a good starting point for system navigation functions. This screen might include important building parameters like space temperatures and occupancy status. It is often useful to be able to access system graphics from the floor plan view. Most current technology systems can load the architectural and mechanical floor plans directly from the project’s AutoCAD files, reducing development time and creating a synergy and consistency between the contract documents and the operating system.

·       Provide System or subsystem graphics Include a graphic for each system, that displays dynamic operating variables and calculated values related to that system’s operation and performance. For large, complex systems, it may be necessary to have a key system graphic that can be used to navigate to subsystems where more operating details are displayed. Ideally, the operators should be able to manipulate setpoints from the system graphic in addition to viewing performance. To facilitate navigation through the system, include links from the system graphics back to the master system graphic, related system graphics and building floor plan. If the project design documents used a systems based approach and include system diagrams or schematics, it is often desirable to develop the control system graphics directly from the contract document information by loading the AutoCAD files of the system diagram to serve as a background for the operating graphics. This will enhance the operator’s ability to run the facility by providing consistency and synergy between the building’s construction documents and the control system graphics, both of which are important tools in the day-to-day operation of the facility.

·       Develop a Master system parameters Table as a Graphic Screen Many operators and facilities engineers find that having critical system parameters displayed as a dynamic graphic in a spreadsheet format provides a useful way to get an overall picture of system operation, especially during emergencies and other critical operating situations where calling up multiple system graphics to understand what was going on in multiple systems would take time and focus away from managing the big picture. Including links in the spreadsheet to each system graphic makes it simple to get more detail as necessary. Including the ability to adjust critical system parameters from the table while still keeping the table in view is also useful in managing a critical operating situation.

3.6.6. Point Trending

Trending can give facilities staff and commissioning providers valuable insight into how the system is operating. The trending requirements during the commissioning process often are different from the trending requirements associated with the day to day operations of the facility, and may actually set some of the control system performance parameters and dictate the structure of the network. In general, trending for commissioning will require frequent data samples over long periods of time. This may require that the system be set up to archive data from the controllers on a regular basis to maintain a seamless, detailed record of system performance. This is because most systems store trend data at the controllers and then archive it to the host computer at set intervals. If there is not much memory available at the controller, then the trend data needs to be downloaded frequently to the host, or information will be overwritten. Large file sizes or frequent downloads can tie up system communications. Most systems stagger the download times or schedule them for times when system activity is low. Having this detailed data record available during the commissioning process can be critical in assessing complex system interactions and picking up system performance problems like hunting, which can be masked by less frequent data samples.

Once the system has been commissioned, it is possible and may even be desirable to reduce the trend frequency to minimize the system’s data management and data handling burden. This is another area where operator taste and preference come into play. Some operators might prefer to not have to deal with the management of the large quantities of data that are associated with ongoing, frequent trend sample times, electing to keep some limited trending running on critical points and then resetting the system for a more frequent sample rate only when they see a problem. Other operators prefer to take advantage of the relatively inexpensive data storage capabilities of current technology systems that make saving large quantities of data a non-issue. They find that being able to go back into their data archive and take a detailed look at how a system was performing months or years ago can often provide insights into current operating problems. As a result, they implement an archiving procedure that involves saving data to a large hard drive or writing data to CDs or zip disks periodically that allows them to maintain a detailed operating history of their systems. For some buildings that are participating in the US-GBC LEED program, having this ability is a key component of the Measurement and Verification Plan, which can be worth a credit point in the LEED rating system.

3.6.7. System Back-ups

Providing a means for backing up the control system database and operating logic is an important but often neglected area in the implementation of DDC control technology. An amazing number of sophisticated systems are specified, furnished and installed without this important capability. In the current technology environment, the cost of this added hardware is inconsequential when compared to the problems that will arise if all or a portion of the system’s database is lost or becomes corrupt and must be regenerated. Having a good back up available makes this a relatively short lived, easily recovered from problem. Without a good back up to work from, a facility could be crippled for weeks or even months while the lost database and operating logic is regenerated and re-commissioned.

The bottom line is that the database and operating software of a well-commissioned and well-tuned DDC system can represent an investment of tens or even hundreds of thousands of dollars in a large building. To protect that investment, it is important to:

·       Include the Necessary Hardware and Software Required to Back-up the System in the Project Construction Documents The first step in a good back-up procedure is to have the hardware and software necessary to do the procedure. Usually, this involves a tape drive, independent hard drive or a CD burner and the software necessary to support the back-up operation.

·       Include Developing a Back-up Procedure and its Associated Requirements as a Part of the Services Provided by the Control Contractor: Before they leave the site and receive final payment, the control contractor should be required to have developed and implemented a regular back-up procedure for the DDC system. In most cases, a once a week process will be sufficient, especially if it is supplemented by a manually initiated back-up any time significant changes are made to a controller or other system component. Usually it is a good idea to maintain two copies of the system back up, one at the operator workstation location, and one off site. This protects the system from an irrecoverable data loss should a catastrophic event occur at the operator workstation location (like a flood, fire or someone setting a stack of refrigerator magnets on the back-up disks).

·       Train the Operating Staff to Back-up the System as a Part of the Training Program Associated with the Commissioning Process For the benefits of the back-up procedure and hardware to be realized and persist, the operating staff need to understand the importance of the procedure and be familiar and comfortable with its implementation. For most systems, the process can be automated to the point where it requires very little operator interaction other than to remove and replace the media when necessary and manage the process.

 



[1]   Belts will tend to stretch slightly when they are first installed and tightened. So, ideally, they should be retensioned after 8 or so hours of operating time has been accumulated. Coordinating the commissioning of the current switch settings with this effort can save time because the belts can be loosened to the point of no load and then retensioned.

[2]   If there are detectable differences in the graphic update times, they are probably due to differences in the network communications burden at the times the graphic was called up to the console. If the delay times run into the range 30-60 seconds or more, then network communications problems may be significant enough to make the system difficult to use, especially in an emergency.

 

[3]   Based on Ohms law, a 250-ohm resistor will generate a 1-5 vdc signal with a 4-20 ma current flow. A 500-ohm resistor will generate a 2-10 vdc signal. Basically, this means that all systems ultimately work with voltage as the analog input measurement. Any system that can accept a voltage input can also accept a current loop input if care is taken with regard to the grounding of the input boards and the current loop power supplies relative to each other.

[4] From a commissioning and operations perspective, these terminals also eliminate that annoying little clink sound that you hear right after you loosened up a connection on a controller, not realizing that there was a precision resistor clamped under it. The sound was the resistor falling into something in the panel. Locating a supplier for one precision resistor to replace the one that was lost can be difficult, especially on projects in remote locations.

[5]   Auxiliary contacts are electrical contacts that change state any time the starter coil changes state. For instance, a normally closed auxiliary contact will be closed with the starter coil is de-energized, and open when the starter coil is energized.

[6]   Many current technology drives have programmable inputs and outputs that can be used to provide hard wired information to a non-networked system. The information that can be provided is generally any of the information available at the drive microprocessor. Common HVAC selections include drive operating status, drive output frequency, motor kW, drive safety status, and motor amps.

[7] Assume $0.08/kWh. This value includes both electricity consumption and demand charges.

[8] Assume $0.50/therm of natural gas.