7.9

Thermostats and Humidistats B. G. LIPTÁK

1440 © 2006 by Béla Lipták

(1985, 1995, 2005)

Types:

A. B. C. D.

Conventional HVAC thermostats Advanced, digital, intelligent thermostats Conventional HVAC humidistats Advanced, digital humidistats

Standard Ranges:

A. 45 to 85°F (7 to 29°C) for heating and from 55 to 105°F (13 to 40°C) for cooling services B. 40 to 90°F (5 to 32°C) to 23 to 122°F (-10 to 50°C) or higher C. 15 to 95% RH D. 0 to 100% RH

Inaccuracies:

Sensor Error: A. ±1°F (0.5°C) B. ±0.5°F (0.28°C), even better with RTD-based digital units C. ±3 to 5% RH D. ±2% RH (some w/1% resolution) Offset Error: A. ±1 to 5°F (0.5 to 2.8°C) B. Zero with integral C. ±2 to 10 % RH D. Zero with integral

Air Capacities:

Relay types (high volume): 10 to 30 SCFH (0.3 to 0.84 m /hr) 3 Non-relay types (low-volume): 1 SCFH (0.028 m /hr)

Costs:

A. Room thermostat $50 to $100. Low-volume, proportional, pneumatic room thermostats cost $75 to $125; high-volume, proportional, dual room thermostats for heating/cooling and with automatic changeover cost $175 to $250; averaging, proportional thermostats for duct mounting cost about $150; on/off electric thermostats can be obtained for $100 or less B. Microprocessor-based on/off home thermostats cost $100 to $200, programmable microprocessor-based energy-saver thermostats cost $200 to $400; C. Proportional, pneumatic room humidistats cost $200 to $250; duct-mounted, proportional, pneumatic humidistats cost about $300; on/off electric humidistats can be obtained for about $100, HVAC transmitter $600 to $800. D. For the costs of higher-quality humidity instruments see Section 8.32 in the Process Measurement volume of this handbook (Note: Installed total costs can double the hardware costs listed above.)

Partial List of Suppliers:

(For Type D see Section 8.32 in the Measurement volume.) Ametek Thermox (C) (www.thermox.com) Barber-Colman Co. (A, B, C) (www.barber-colman.com) Dwyer Instrument, Mercoid Div. (A) Emerson Electric Co. (A, B) (www.emersonprocess.com) Eurotherm/Barber Coleman (A, B) (www.eurotherm.com) Foxboro-Invensys (C, D) (www.foxboro.com) General Eastern Instruments (C) (www.geinet.com) Honeywell Inc. (A, B, C) (www.honeywell.com/sensing)

3

7.9 Thermostats and Humidistats

1441

Johnson Controls Inc. (A, B, C) (www.johnsoncontrols.com) Jumo (A, B, C) (www.jumousa.com) Ohmic Instrument Co. (C) (www.ohmicinstruments.com) Panametrics Inc. (C, D) (www.panametrics.com) Powers Controls (A/B) (www.powerscontrols.com) Princo Instruments Inc. (A, C) (www.princoinstruments.com) Robertshaw/Invensys (A/B) (www.robertshaw.com/cli-fam-robTherm.htnl) Rotronic Instrument Corp. (C) (www.rotronic-usa.com) Staefa/Siemens Building Automation (A/B) (www.sbt.siemens.com/hvp/Components/ catalog/thermostats.asp) Vaisala Oyj. (C) (www.vaisala.com) Watlow Controls (C) (www.watlow.com) Westinghouse Electric Corp. (A, B) (www.westinghousepc.com) White Rogers (A, B) (www.whiterogers.com) Yokogawa Corp. of America (C) (www.yokogawa.com/us)

INTRODUCTION There is an overlap between the measurement and the control of room temperature and humidity. The first volume of this handbook on Process Measurement and Analysis covers these topics from the measurement perspective in Sections 4.11 and 8.32. In this section, the emphasis will be placed on the control capabilities of these devices. Thermostats and humidistats have been developed to serve the heating, ventilating, and air conditioning (HVAC) industry. While energy costs were low, the consideration of performance was secondary to the purchase price of these devices. When energy costs increased, new, better-quality units were developed, but the conventional HVAC devices still remain in use in existing systems and in new installations, where the designers are less sensitive to the quality of performance. In this section, the conventional quality units and the more advanced designs are both described. On/off and throttling devices are both referred to as thermostats or humidistats. Therefore, both versions will be discussed, but more emphasis will be placed on the throttling-type designs. First the narrow proportional band controllers, the stats, will be described. After that, the various thermostat and humidistat designs will be covered. ACCURACY OF THERMOSTATS AND HUMIDISTATS Conventional In conventional HVAC thermostats and humidistats, the total error is the sum of the sensor error and of the offset error. Even after individual calibration, the sensor error of conventional thermostats cannot be reduced to less than ±1°F (±0.55°C) for thermostats or to ±5% relative humidity (RH) for humidistats. Added to the sensor error is the offset error, which increases as the throttling range increases. This, for conventional thermostats, can be as high as ±5°F (±2.8°C) and ±20% RH for humidistats. If the two errors are additive, the total errors of conventional units can approach ±6°F (±3.3°C) and ±20–25% RH, respectively.

© 2006 by Béla Lipták

In conventional HVAC electrical thermostats or humidistats, the total error is the sum of three components. The first error component is the element error, which is the same as the sensor error in pneumatic designs. The second error component is the mechanical differential set in the thermostat switch, which does not modulate the final control element, but turns it on and off. The differential (or bandwidth) can vary for the thermostats from 0.5 to 10°F (0.3 to 5.6°C) and for humidistats from 2 to 50% RH. The third error component is due to the fact that turning on a heat or humidity source does not instantaneously return the space to within the control differential. Instead, conditions will continue to deviate further for a while. The total error resulting from these three contributing factors can be as high as for HVAC pneumatics, although if the final control elements are sized large enough and if the control differential is set small, electrical units are likely to outperform their pneumatic counterparts. Advanced If space conditions are to be accurately maintained, such as in clean rooms where the temperature must be controlled within ±1°F (±0.55°C) and the humidity must be kept within ±3% RH, neither of the above designs is acceptable. In such cases it is necessary to use a resistance temperature detector (RTD)type or a semiconductor transistor-type temperature sensor and a proportional-plus-integral thermostat controller, which will eliminate the offset error. This can be most economically accomplished through the use of microprocessor-based thermostats that communicate with sensors over data highways. The performance of advanced, microprocessor-based digital units is much better than that of conventional thermostats. The error in an RTD-based digital thermostat is around 1% of span. Similarly, the resolution of a digital hygrometer is around 1%. THERMOSTATS Conventional thermostats are usually uncalibrated devices, and their manufacturers usually do not guarantee their accuracy. This is a limitation, because it is possible to have some

1442

Regulators and Final Control Elements

High temperature switch (Detects temperature rise) Accuracy

Low temperature switch (Detects temperature decrease) Tolerance Reactivation point

Set point

Adjustable range

Differential

Differential

Reactivation point

Set point Tolerance

Accuracy

FIG. 7.9a A cooling thermostat operates as a high-temperature switch, and a heating thermostat functions as a low-temperature switch.

thermostats with as much as 5–10°F (3–6°C) error in their measurement. In the last decades, a new generation of thermostats have been introduced that guarantee to limit their error to 1°F (0.6°C) or less. ElectroMechanical Designs There is little difference between a two-position thermostat and a temperature switch. A thermostat that is controlling the heating of a space operates like a low-temperature switch, and a cooling thermostat operates as a high-temperature switch (Figure 7.9a). The movement produced by a bimetallic temperature sensor or by the expansion of a humidity-sensitive element can be used to build two-position controllers. The electric thermostats in Figure 7.9b illustrate the various differential gaps in two-position control. The thermostat on the left operates load contacts directly from the bimetallic element. The set point is adjusted by setting one contact or by repositioning the whole bimetal element. The gap in this arrangement is theoretically zero, although in actual practice there is a very small gap due to contact “stickiness.”

+G −G Zero gap

Positive gap

Negative gap

FIG. 7.9b Design variations of electrical two-position thermostats.

© 2006 by Béla Lipták

A positive differential gap is often necessary in twoposition control in order to save wear on the control apparatus. The mechanism shown in Figure 7.9b employs a toggle action to produce a positive gap. The gap is adjustable by the tension of the spring, because this determines the suppressive (hysteresis) force acting on the element. A negative differential gap is often used on domestic thermostats in order to reduce an abnormally long period of oscillation. An auxiliary heater operated by the load contacts turns on a small heater within the thermostat when the contact is made. Thus, the bimetal element opens the load contact before the actual room temperature attains the same point. The load contact make point is normal. This is sometimes termed “anticipation.” There are many forms of two-position thermostat- or humidistat-type controllers for both domestic and industrial applications. These offer many arrangements of load contacts, set point, and differential gap. Generally, a solenoid valve or an electrical heating element, motor, or other power device is operated by the controller. For floating controller action, a neutral zone is usually necessary. This is obtained by a second contact position in the controller and is often obtained with two independently set on/off control mechanisms. Single-speed floating control can be achieved by means of the same equipment as for two-position control with an electric-motor-operated valve. The difference lies only in the speed of operation of the motor or valve actuator. In twoposition control the valve stroke is 120 sec or less. In singlespeed floating control, the valve stroke is usually 120 sec or more. An electrical interrupter is sometimes employed in conjunction with the motor to decrease the speed of opening and closing the valve. Electrical/Electronic Design Features The most common thermostat is the low-voltage unit that uses external interposing relays to control heating or cooling equipment (typically 24 VAC). Other models include: Line voltage units that directly control AC circuits (120 or 240 VAC). Heating only or heating/cooling models. Digital or analog indication of room temperature. Bimetallic, filled system or electronic sensing elements. Direct or reverse acting. Local or remote set point. Local set points may be external or internal and may be key protected. Limited control range thermostats allow the occupant of an office to move the set point to any value desired, but will disregard any setting that exceeds the limit value. Programmable set-back models with 24-hour or 7-day programs.

7.9 Thermostats and Humidistats

Microprocessor-Based Units A smart thermostat is usually a microprocessor-based unit with an RTD-type or a transistorized solid-state sensor. It is usually provided with its own dedicated memory and intelligence, and it can also be provided with a communication link (over a shared data bus) to a central computer. Such units can minimize building operating costs by combining timeof-day controls with intelligent comfort gap selection and with maximized self-heating. Microprocessor-based units continue to incorporate new features. These are programmable devices with memory and communication capability. They can be monitored and reset by central computers using pairs of telephone wires as the communication link. Microprocessor-based units can be provided with continuously recharged backup batteries and with accurate electronic room temperature sensors. They can also operate without a host computer (in the “stand-alone” mode). In this case the user manually programs the thermostat to maintain various room temperatures as a function of the time of day and other considerations. Control by Phone In the year 2000, the first residential gateway for the HVAC industry’s new open communications standard for residential environmental control was introduced. This gateway enables secure, remote access to residential HVAC systems via touchtone telephone, standard phone line, and pass code. Some systems offer either “dial-in” or “dial-out” functions. Owners can call in to check the temperature, change the status of their HVAC system, and perform other thermostat-related functions. These telephone access modules enable monitoring of the building’s temperature and its heating/cooling equipment. It can report when a furnace filter needs replacing or an electronic air cleaner’s cells need cleaning. And it can call up to three phone numbers to immediately alert the owner, contractor, or others of problems, such as freezing temperatures or an extended power outage. The Proportional-Only Controller A thermostat or humidistat is a simplified controller, having only one control mode —proportional—that is narrow and fixed. The pressure of the output signal from a pneumatic “stat” is a near straight-line function of the measurement, described by the following relationship: O = Kc ( M − M 0 ) + O0

7.9(1)

where O = output signal Kc = proportional sensitivity (Kc can be fixed or adjustable depending on the design) M = measurement (temperature or relative humidity)

© 2006 by Béla Lipták

1443

M0 = “normal” value of measurement corresponding to the center of the throttling range O0 = “normal” value of the output signal, corresponding to the center of the throttling range of the control valve (or damper) Throttling Thermostats are distinguished from other controllers in that they provide proportional action only and that their proportional band (gain) is narrow and fixed. In this section, the operation of thermostats will be described on the basis of their pneumatic designs, but all concepts and features that are discussed are equally applicable to analog or digital electronic thermostats, also. A conventional pneumatic room thermostat might have a set point range of 55–85°F (13–29°C) and a fixed sensitivity (gain) of 2.5 psi/°F (0.3 bar/°C). Assuming that this thermostat operates a control valve having a 9–13 PSIG (0.6–0.9 bar) spring range, we can convert the fixed sensitivity of 2.5 psi/°F into a percent-proportional band value that is better understood by instrument engineers. Sensitivity (or gain) is stated as the ratio between a change in measurement and a change in the corresponding controller output. A 1°F (0.56°C) change in measurement is 3.3% of the measurement range of 30°F (55–85°F, or 13–29°C), while the 2.5 psi change in controller output is 62% of the controller output range of 4 psi (9–13 PSIG). If a 3.3% change in input results in a 62% change in output, that is a gain of 62/3.3 = 19. Converting this typical thermostat gain to proportional band, we get: P=

100 100 = = 5.3% G 19

7.9(2)

If a thermostat is controlling a valve or a damper, one needs to determine Kc, M0, and O0 in Equation 7.9(1) in order to draw the straight-line relationship between measurement and output signal. Figure 7.9c illustrates the behavior of a control loop consisting of a direct acting thermostat, having a fixed Kc of 2.5 PSIG/°F (0.3 bar/°C), a “normal” measurement value of M0 = 72°F (22ºC), and a damper spring range of 8–13 PSIG (0.55–0.9 bar). This combination results in a normal output signal: O0 = 10.5 PSIG (0.73 bar). Does it Have a Set Point? It is a common error to call M0 in Figure 7.9c the set point of that thermostat. Stats do not have setpoints in the sense of having a predetermined temperature value to which they would seek to return the condition of the controlled space. (One must add integral action in order for a controller to be able to return the measured variable to a set point after a load change.) M0 does not represent a set point; it only identifies the space temperature that will cause the cooling damper in Figure 7.9c to be 50% open. This can be called a “normal” condition, because relative to this point the thermostat can both increase and decrease the cooling air flow rate as space temperature

1444

Regulators and Final Control Elements

Cold air FC 8−13 PSIG

Range: 40−90°F TC Normal: 72°F PB = 5% D/A Flow sheet representation

Output signal “O” (PSIG)

15 14 13 12 Spring 11 range 10 (5 PSIG) 9 8 7 6 5 4 3

Offset error, which is also called “Throttling range” 2°F

Maximum cooling (Damper open)

“Normal” condition

O0 = 10.5 PSIG

Operating line

Minimum cooling (Damper closed)

40 Note: °F − 32 °C = 1.8 1 PSIG = 6.9 kPa

Measurement “M” (°F) 50

60

70

80

90

M0 = 72°F Scale range = 40 − 90°F

Slope of operating line = Proportional sensitivity = 2.5 PSIG/°F = Proportional band = 5% = Gain = 20

FIG. 7.9c Operating characteristics of a narrow and fixed proportional band, pneumatic thermostat.

changes. If the cooling load doubles, requiring the damper to be fully open, this cannot take place until the controlled space temperature has first risen to 73°F (23°C). As long as the cooling load remains that high, the space temperature must also stay up at the 73°F (23°C) value. Similarly, the only way this thermostat can reduce the opening of the cooling damper below 50% is to first allow the space temperature to drift down below 72°F (22°C). Therefore, stats do not have set points, but they have throttling ranges, and if a throttling range is narrow enough, this gives the appearance that the controller is keeping the variable near set point, when in fact the narrow range allows the variable to drift within limits. Offset Error This controller has a fixed proportional band of 5%; it is is very sensitive, has a small “offset,” and, as its gain is fixed, it cannot be tuned. Let us examine the consequences of these characteristics separately. A sensitive controller is suited for the control of very slow, large-capacity processors. Temperature control in the HVAC industry usually fits that description, and therefore thermostats with fixed, high gains can give acceptable results, if space temperature can change only very slowly. On the other hand, such thermostats cannot control spaces with fast dynamics (short time constants) and will cycle or lose control when applied to such service. The “offset” inherent in all proportional controllers is not a serious drawback because the resulting error is small, due

© 2006 by Béla Lipták

to the narrowness of the proportional band. Assuming that the thermostat output is set at 11 PSIG (0.76 bar) when the error is zero (thermostat is on set point), we can calculate the maximum error due to offset. When there is no error, the control valve is 50% open (output is 11 PSIG, or 0.76 bar). As the load changes, an error must be allowed to develop in order for the valve opening to change. Assuming that the valve has been correctly sized and is large enough to handle all expected loads, the maximum error due to offset will be that deviation that is required to move the valve from half to full opening. If the thermostat’s sensitivity is 2.5 PSIG/°F, this 2 PSIG (0.14 bar) change in thermostat output will occur when the deviation from set point is 0.8°F (0.4°C). Therefore, this is the size of the permanent offset error for a gain of 19:1. There is an inverse linear relationship between gain and offset error, such that if the gain is cut in half, the offset error will be doubled. From the above discussion it might be concluded that the conventional room thermostat is a good selection for the HVAC-type applications and that more expensive instruments, such as PID controllers, would not necessarily improve the overall performance. Throttling Range, Gain, or Sensitivity The control response of thermostats is described by the slope of the operating line in Figure 7.10c. This slope is described in different ways depending on the industry and manufacturer involved. People in the process control fields tend to be more familiar with terms such as “gain” and “proportional band,” while people working in the HVAC field are more used to the terms “proportional sensitivity” or “throttling range.” All of these terms describe the same slope shown in Figure 7.9c and they are also defined below: The throttling range is the gap within which the space conditions are allowed to drift as the final control element is modulated from fully closed to fully open (2°F or about 1°C in Figure 7.9c). Figure 7.9d describes the throttling range of a 3–13 PSIG (0.2–0.9 bar) thermostat. Thermostat Output (PSIG)

13

3 Throttling Range

Temperature (°F)

FIG. 7.9d The throttling range of a 3–13 PSIG (0.2–0.9 bar) thermostat.

7.9 Thermostats and Humidistats

Proportional sensitivity is the amount of change in the output signal pressure that results from a change of one unit in the measurement (2.5 PSIG/°F in Figure 7.9c). Gain is the ratio between the sizes of the changes in control output and measurement input. In case of Figure 7.9c, a 100% change in output (5 PSIG, moving the damper from fully open to fully closed) will result from a 2°F (1°C) change in measurement, which represents only 4% of the thermostat span of 40–90°F (6.6–32°C). Therefore, the gain of this thermostat is 25. Proportional band (PB) is related to gain as follows: PB = 100/G = 100/25 = 4%. Therefore, the proportional band of the thermostat in Figure 7.9c is 4%. Having defined the “normal” conditions (M0 and O0) and also the slope of the operating line, the behavior of the thermostat is now fully defined.

1445

Thermostat Action and Spring Range The thermostat in Figure 7.9c is said to be direct acting (D/A) because its output signal is increased as the measurement rises. If, instead of the Fail Closed damper, a Fail Open damper were used, a reversal in the thermostat action would be required, because now maximum cooling would take place when the thermostat output signal is low. Hence, in that case a reverse-acting (R/A) thermostat would be used, and the left-to-right slope of the operating line in Figure 7.9c would change to right-to-left. Spring Ranges The operating line in Figure 7.9c varies with the spring range of the final control element. The typical spring ranges that are available are listed below: Fail Closed HVAC Quality Dampers

Gain or Sensitivity Adjustment The simplest thermostats are manufactured with fixed sensitivities. Such a unit was illustrated in Figure 7.9c, having a fixed proportional sensitivity of 2.5 PSIG/°F (0.3 bar/°C). Control flexibility is improved in those designs where the proportional sensitivity is adjustable. Figure 7.9e illustrates a thermostat with a 5:1 range of gain adjustment, which is typical to standard thermostats. Lowering the proportional sensitivity tends to improve stability, but it also increases the offset error due to the widening of the throttling range.

A—3 to 7 PSIG (0.2 to 0.5 bar) B—5 to 10 PSIG (0.35 to 0.7 bar) C—8 to 13 PSIG (0.55 to 0.9 bar) HVAC Quality Valves D—Fail Open, 4 to 8 PSIG (0.9 t 0.55 bar) E—Fail Closed, 9 to 13 PSIG (0.6 t 0.9 bar) F—Three-way, 7 to 11 PSIG (0.5 to 0.76 bar) Speed or Blade Pitch Positioners (Fail Closed) G—3 to 15 PSIG (0.2 to 1.0 bar)

Cold air FC 8−13 PSIG

Range: 40−90°F TC Normal: 72°F PB = 5%−25% D/A Flow sheet representation

Output signal “O” (PSIG)

15 14 13 Spring 12 range 11 10 (5 PSIG) 9 8 7 6 5 4 3 Note:

Offset error, which is also called “Throttling range” Range of 10°F adjust(2°F) ability Maximum cooling (Damper open) O0 = 10.5 PSIG

“Normal” condition

Minimum cooling (Damper closed)

Operating lines

Measurement “M” (°F)

40 °F − 32 1.8 1 PSIG = 6.9 kPa

°C =

50

60

70

80

M0 = 72°F

Range of adjustability: Offset error or throttling range: 2 to 10°F (1.1 to 5.6°C) Proportional sensitivity: 2.5 to 0.5 PSIG/°F (31 to 6.3 kPa/°C) Proportional band: 5% to 25% Gain: 20 to 4

FIG. 7.9e The features and characteristics of a thermostat with adjustable sensitivity.

© 2006 by Béla Lipták

Figure 7.9c illustrated the operating line of a thermostat with a 2.5 PSIG/°F (0.3 bar/°C) proportional sensitivity in combination with a final control element having a type C spring. If the range spring of the final control element is changed, this would also change the operating lines, as illustrated in Figure 7.9f. From Table 7.9g, it is obvious that the range spring selection can substantially affect the throttling range and the apparent proportional band. The wider the throttling range, the larger the control error (offset or drift away from the “normal” measurement value). On the other hand, the wider the throttling range, the more stable (less cycling) the final control element is likely to be. Therefore, the trade-off is between lower stability or higher offset error. By narrowing the spring ranges to 4 or 5 PSIG (0.3 or 0.35 bar), the conventional HVAC control manufacturers have selected lower stability as the more acceptable of the two undesirable options. The process control industry has made just the opposite choice, as shown in Figure 7.9f by the 12 PSIG (0.8 bar) spring line (type G). This selection stems from the desire to maximize stability while being unconcerned with offset error because it was eliminated by the addition of the integral mode. Thermostat Design Variations Figure 7.9h illustrates the design of the simplest thermostat that is manufactured. This fixed-proportional controller uses

1446

Regulators and Final Control Elements

Output signal “O” PSIG (kPa)

Due to the small restriction and small nozzle openings in Figure 7.9h, the resulting output airflow is rather small, around 1.0 SCFH (0.028 m3/hr). This type of design is also referred to as “nonrelay” or “low volume” design. The 1.0 SCFH air capacity can be insufficient to fill connecting tubing and operate final control elements. In order to increase the air capacity of a thermostat, a booster or repeater relay can be added. Such “relay type” or “high-volume” thermostats will usually provide an output airflow of 10–30 SCFH (0.3– 3 0.84 m /hr), which is sufficient for most applications. In the case of the relay-type design (Figure 7.9q), the nozzle backpressure, instead of operating a final control element directly, is sent to the bellows chamber and acts against the bellows. Because the bellows has some stiffness (spring gradient), the pilot valve plug is positioned between the inlet and outlet ports in accordance with the nozzle back-pressure. A low nozzle back-pressure positions the valve plug to the left, throttles the exhaust port, and causes a high output pressure. The advantages of adding the pilot are (1) the actuating signal vs. output pressure relationship can be made linear, and (2) the capacity for airflow can be considerably increased.

G

15 (104)

E C

14 (97) 13 (90) 12 (83) 11 (76)

B F

10 (69) 9 (62) 8 (55)

A

7 (48) 6 (41) 5 (36) 4 (28) 3 (21)69 (20.5)

D Measurement “M” , °F (°C) 70 (21.1)

71 (21.7)

72 (22.2)

73 (22.8)

74 (23.3)

75 (23.9)

FIG. 7.9f Illustrating the effect of damper or valve spring range on the gain or throttling range of the thermostat.

a flapper arrangement, which moves in front of a nozzle as the temperature detected by a bimetallic element changes relative to the manually adjusted “normal” value. Air at about 20 PSIG (140 kPa) is supplied through a restriction to the open nozzle, the back-pressure of which is inversely proportional to the distance between nozzle and flapper. Because the A/B ratio is fixed in Figure 7.9h, the proportional sensitivity (Kc = change in output pressure per °F measured) is also fixed. Figure 7.9q describes a humidistat design in which the proportional sensitivity is adjustable.

Special-Purpose Thermostats Application engineers can choose from a fairly large variety of design features when specifying thermostats, because they can not only be electromechanical, pneumatic, or electronic, but they can also • • • • • •

be indicating or blind be direct or reverse acting automatically switch their actions in response to a pneumatic or electronic signal have bimetallic, filled, or electronic sensing elements have local or remote set points have their set points under key, concealed, or externally adjustable.

Advanced Features If the set point of a pneumatic thermostat is adjusted remotely by an air signal, each 1 PSIG (0.07 bar)

TABLE 7.9g Thermostat Gain (Proportional Sensitivity) and Throttling Range as a Function of Its Spring Range Range Spring Type

PB (%)

GG

Proportional Sensitivity PSIG/°F (bar/°C)

Throttling Range °F (°C)

“Normal” Output PSIG (bar)

Spring Range PSIG (bar)

A

3.2

31

2.5 (0.3)

1.6 (0.9)

5.0 (0.35)

3–7 (0.2–0.5)

B

5.0

20

2.5 (0.3)

2.0 (1.1)

7.5 (0.52)

5–10 (0.35–0.7)

C

5.0

20

2.5 (0.3)

2.0 (1.1)

10.5 (0.72)

8–13 (0.55–0.9)

© 2006 by Béla Lipták

D

3.2

31

2.5 (0.3)

1.6 (0.9)

6.0 (0.4)

E

3.2

31

2.5 (0.3)

1.6 (0.9)

11.0 (0.76)

4–8 (0.3–0.55) 9–13 (0.6–0.9)

F

3.2

31

2.5 (0.3)

1.6 (0.9)

9.0 (0.6)

7–11 (0.5–0.76)

G

9.6

10

2.5 (0.3)

4.8 (2.7)

9.0 (0.6)

3–15 (0.2–1.0)

7.9 Thermostats and Humidistats

Sets the “normal” value (M0 in Figure 2.10 a) Manual setpoint adjustment

This thermostat is called to be a “fixed” proportional band design because the distances “A” and “B” are fixed. A standard response is to give a 2.5 PSIG (17 kPa) change in output signal for each °F change in measurement.

Set-point scale

°F °C

1447

50 60 70 80 °F 10 15 20 25 °C

0.01"(0.25 mm) diameter restriction A INCR.

B

20 PSIG (140 kPa) air supply

As the temperature rises the bimetallic element moves to the right, increasing the backpressure on the nozzle and thereby raising the thermostat output signal. (This is called direct action)

Bimetallic element

Flapper 0.015" (0.375 mm) Diameter nozzle

Low volume (1 SCFH or 0.028 m3/hr) output signal to dampers or to control valves

FIG. 7.9h Direct-acting bimetallic thermostat with manual set point and fixed proportional band. The output airflow is limited because no booster relay is included. For this reason, it called a “low-volume” thermostat.

change in set point pressure will move the set point by an adjustable preset amount. The range of this adjustment is usually from 0.15–1.4°F (0.1–0.8°C) per 1 PSIG (0.07 bar). If the set point is to change as a function of the time of day, a timer can automatically operate a solenoid and thereby switch the set-point signal. Some of the more recently developed and more advanced thermostat features include: Dual Set Points Dual set point thermostats will switch their settings in response to a change in the air supply pressure. Both set points can be manually adjustable, with the day setting made by external thumbwheel and the night setting concealed internally. One design variation is a dual thermostat that has two set points, one corresponding to the minimum, the other to

D.A.

the maximum allowable space temperature. These thermostats can use their dual outputs to operate a variable air volume (VAV) damper to cool the space and a reheat coil to heat it (Figure 7.9i). In other control systems the dual outputs of the thermostat can be used to throttle the hot or the cold water to condition the air being supplied to the room (Figure 7.9j). Night–Day or Set-Back These thermostats will operate at different “normal” temperature settings for day and night. They are provided with both a “day” and “night” setting dial, and the change from day to night operation can be made automatic for a group of thermostats. The pneumatic day–night thermostat uses a two-pressure air supply system, the two pressures often being 13 and 17 PSIG (89.6 and 117 kPa) or 15 and 20 PSIG (103.35 and 137.8 kPa).

D.A.

Hot water control valve (FC)

V.A.V. damper control (FC) Dual control room thermostat Primary

Room

Duct

Supply air

air D.A. = Direct action FC = Fail closed V.A.V. = Variable air volume

Volume regulator (Damper FC)

Reheat coil

FIG. 7.9i Dual thermostat for controlling a variable air volume (VAV) box with reheat. (Courtesy of Powers Controls.)

© 2006 by Béla Lipták

1448

Regulators and Final Control Elements

Conditioned air to room

Hot FO Heating coil Cold FC ing

l Coo

D.A.

D.A.

water D.A. = Direct action FC = Fail closed FO = Fail open

Dual control room thermostat

water

coil

Return air

FIG. 7.9j Dual thermostat installed to control a heating and a cooling coil. (Courtesy of Powers Controls.)

Changing the pressure at a central point from one value to the other actuates switching devices in the thermostat and indexes them from day to night or vice versa. Supply air mains are often divided into two or more circuits so that switching can be accomplished in various areas of the building at different times. For example, a school building may have separate circuits for classrooms, offices and administrative areas, the auditorium, and the gymnasium and locker rooms. In some of the electric designs, dedicated clocks and switches are built into each thermostat. Heating–Cooling or Summer–Winter This thermostat can have its action reversed and, if desired, can have its set point changed by means of indexing. It is used to actuate controlled devices, such as valves or dampers, that regulate a heating source at one time and a cooling source at another. It is often manually indexed in groups by a switch, or automatically by a thermostat that senses the temperature of the water supply, the outdoor temperature, or another suitable variable. In the heating–cooling design there frequently are two bimetallic elements, one being direct acting for the heating mode, the other being reverse acting for the cooling mode. The mode switching is done automatically in response to a change in the air supply pressure, similarly to the operation described for the day–night thermostats. Limited Control Range These thermostats allow the occupant of an office to move the set point to any value desired, but will disregard any setting that exceeds the limit value. For example, in heating applications, the limit could be 74°F (23°C). In this case the space temperature will be limited to a maximum of 74°F, regardless of the setting by the occupant. In the cooling season the minimum limit value of 75°F (24°C) can be set for cooling. This is done internally without placing a physical stop on the setting knob. Zero Energy Band Control The idea behind zero energy band (ZEB) control is to conserve energy by not using any when the room is comfortable. Zero energy band thermostats will provide heating when the zone temperature is below

© 2006 by Béla Lipták

72°F (22°C) and will provide cooling when it is above, say, 78°F (25.6°C) (these are adjustable settings), and in between, they will neither heat nor cool. One way to achieve this goal is to use the dual thermostat illustrated in Figures 7.9i and 7.9j. If in these installations a single set-point thermostat was used instead of a dual one, energy would be wasted, because as soon as heating was stopped, cooling would be initiated (Figure 7.9k). This is wasteful, as no energy needs to be added or removed from a space that is comfortable as it is (within the ZEB band). This approach can reduce the yearly operating cost by up to about 33%. ZEB control can be accomplished in one of two ways. The dual set-point approach is shown on the right of Figure 7.9l. The single set-point and single output approach is illustrated on the left side of Figure 7.9l. In the single set-point, spilt-range configuration, the cooling valve fails closed and is shown to have an 8–11 PSIG (0.55–0.76 bar) spring range, while the heating valve is selected to fail open and has a 2–5 PSIG (0.14–0.34 bar) range. Therefore, between 5 and 8 PSIG (0.34 and 0.55 bar) both valves are closed, and no pay energy is expended while the thermostat output is within this range. The throttling range is usually adjustable from 5–25°F (3–13°C). Thus, if the ZEB is 30% of the throttling range, it can be varied from a gap size of 1.5°F (.85°C) to 7.5°F (4.2°C) by changing the throttling range (or gain). ZEB control can also allow buildings to become selfheating by transferring interior heat to the perimeter. (For a detailed discussion of self-heating, see Section 8.2.) Slave or Submaster This thermostat has its set point raised or lowered over a predetermined range, in accordance with variations in the output from a master controller. The master controller can be a thermostat, manual switch, pressure controller, or similar device. For example, a master thermostat measuring outdoor air temperature can be used to adjust a submaster thermostat controlling the water temperature in a heating system. Master–submaster combinations are sometimes designated as single-cascade action.

7.9 Thermostats and Humidistats

1449

No-energy-band, dual control, thermostat 100% Heating load

100% No energy band

50%

50%

Cooling load

Adjustable 0% 65°F 18°C 100%

Heating load

0% 85°F 29°C

72°F 75°F 78°F 22°C 24°C 26°C Room temperature

100%

Conventional thermostat

50%

50%

Wasted energy

0% 65°F 18°C

Cooling load

0% 85°F 29°C

72°F 75°F 78°F 24°C 26°C 22°C Room temperature

FIG. 7.9k Energy can be saved by neither cooling nor heating a space if its temperature is within a zone that is comfortable.

When such action is accomplished by a single thermostat having more than one measuring element, it is known as compensated control. Split-Range Control Multistage thermostats are designed to operate two or more final control elements in sequence. While the split-range approach is a little less expensive than the dual set-point scheme (shown on the right of Figure 7.9l), it

Full heating 12 11 10 9 8 7

Cooling valve open

9 Cooling valve closed 8 7 5 4 3

Full cooling

13

Gain = 0.75 PSI/°F

10

6

1. The gap width can only be adjusted by also changing the thermostat gain; maximum gap width is limited by the minimum gain setting of the unit. 2. In this design, the heating valve must fail open, which is undesirable from an energy conservation point of view.

Dual output pressure from thermostat (PSIG)

Output pressure of single thermostat (PSIG) 11

is also less flexible and more restrictive. The two basic limitations of the split-range approach are:

Zero energy band

6 Heating valve closed

5 4

Heating valve open

3 2

2 1 0

67 Heat

71

75 Zeb

79 Cool

Single set point split-range system

Space temperature °F

1

Heat off

Cooling off

0 64 66 68 70 72 74 76 78 80 82

Space temperature °F

Dual set point system

FIG. 7.9l The zero energy band (ZEB) can be obtained by single set-point split-range design (left) or by dual set-point system (right).

© 2006 by Béla Lipták

Regulators and Final Control Elements

HUMIDISTATS Basically the same design variations are available for humidity control as have already been discussed in connection with thermostats. By replacing the temperature sensor element, such as a bimetallic spring, with a humidity-sensitive sensing element, such as a polymer, a humidistat can be obtained. A wet-bulb thermostat is also often used for humidity control with proper control of the dry-bulb temperature. A wick, or other means for keeping the bulb wet, and rapid air motion to ensure a true wet-bulb measurement are essential. A dew-point thermostat is a device designed to control from dew-point temperatures. Just like thermostats, humidistats can also be on/off, pneumatic, electric, electronic, or microprocessor-based digital. For the design features of these and for a discussion of the nature of fixed and narrow proportional control and the resulting offset, the reader is referred to the previous paragraphs.

%

%

70

ir

ry a

of d

und

60

%

r po

. pe

40

%

50 %

−B TU

eat

al h Tot

Weight of water vapor in one pound of dry air − grains

Pressure of water vapor − PSI

ir

0

5

0

.5 12

0.00

13.

0.10

13.

0.20

14.

0.30

240 230 90 220 210 200 44 190 85 42 180 40 170 s 80 38 160 re u at 150 36 r pe 140 34 em 75 t 130 nt i 32 po 120 0 w 7 30 de 110 d % 28 an 65 30 100 b l W et b bu 90 t u lb li e 60 nes 80 W % 0 2 70 55 60 50 50 10% dity 40 umi 30 ve h i t a l Re 20 10 0 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120 Dry bulb temperature − degrees F Copyright, 1942, General Electric Company 5 14.

0.40

250

58 56 54 52 50 48 46

a ry

0.50

Humidity refers to the water vapor contained in the air at a particular temperature. Warm air has a greater capacity for water vapor than does cold air. Relative humidity is the mole fraction of moisture in air to the mole fraction of moisture in a saturated mixture at the same temperature and pressure. In other words, RH is the ratio of how much water vapor is in the air vs. the maximum it could contain at the particular pressure and temperature. Relative humidity is the ratio of the actual partial pressure of the water vapor to the saturation vapor pressure at a particular temperature. Therefore, in a room that has 50% RH at 20°C, the RH will drop to 37%, if the temperature rises to 25°C. (Note that temperature has no effect on the dew point temperature, which in the above example stays constant at 9.3°C.) Relative humidity can be calculated using the wet and dry bulb temperature readings, or relative humidity can be read from a psychrometric chart such as the one shown in Figure 7.9m. The sensor in a humidistat can be moisture-sensitive hair, cellulose, synthetic polymer fiber, surface conductivity, or resistance or capacitance elements. Other, more accurate

d . of r lb .pe 15.0

0.60

Relative Humidity Sensors

.ft Cu

0.70

250 240 230 220 Barometric pressure 14.696 PSIA 210 200 190 180 170 160 150 140 130 26 120 24 110 22 100 20 90 18 80 16 70 14 60 50 12 45 40 40 10 35 30 30 5 2 20 10 0 20 25 30 35 40 45

Thermometers and hygrometers can also be combined into single instruments, and their microprocessor-operated versions can also be provided with memory, so they can recall past temperature and humidity records and can allow remote access to such data.

1 90% 00%

These limitations are removed when a dual set point, dual output thermostat is used. Here both valves can fail closed and the bandwidth is independently adjustable from the thermostat gain. The gains of the heating and cooling thermostats are also independently adjustable. In Figure 7.9l, the heating thermostat is reverse acting and the cooling thermostat is direct acting. For a detailed discussion of controlling and optimizing HVAC systems, refer to Section 8.2.

80

1450

FIG. 7.9m One can read the relative humidity of air from the above psychrometric chart if one knows any two of the air properties shown in the chart.

© 2006 by Béla Lipták

Vio let

Gr ay

e Blu

Gr een

Ye llo w

Or an ge

W Blahite Br ck ow Re n d

7.9 Thermostats and Humidistats

70

80

Lower (20 K) upper (909 K)

Resistance in k/m ohms

10 M 1M 100 K 10 K 1K

1451

0

10

20

30

40

50

60

90

100

Relative humidity in percent

FIG. 7.9n The chart describes the RH-to-resistance relationship of ten color-coded Dunmore sensor elements. (Courtesy of Ohmic Instruments Co.)

humidity and moisture sensors are also available and are discussed in detail in Section 8.32 in the first volume of this handbook. Cellulose Hygrometers Cellulose strips or other shapes are used to measure humidity. Like human hair and synthetic materials, cellulose changes its dimensions as the water vapor concentration varies. This elongation has been used to build dial and digital indicators and also relatively inexpensive HVACtype thermostats, transmitters, and recorders. The operation of the cellulose element instruments involves the detection of dimensional change of the element. Solution Resistance Elements (Dunmore Cells) The Dunmore sensor consists of a wire grid on a insulating substrate that is coated with lithium chloride solution. Lithium chloride is hygroscopic and therefore takes up moisture from the air. The resulting resistance of the sensor is an indication of the relative humidity in the air. One disadvantage of this otherwise widely used, inexpensive, and good sensor is its narrow range.

Figure 7.9n illustrates the ranges of ten color-coded elements that are used by one manufacturer to cover the full range of 0–100% RH. Wide-range Dunmore sensors can be configured by placing several narrow-range elements in a single housing and connecting them to a microprocessorbased readout. Polystyrene Surface Resistivity (Pope Cells) The Pope cell is similar to the Dunmore element, but instead of lithium chloride solution it uses polystyrene as a substrate for the conductive wire grid. In this design the polystyrene is treated with sulfuric acid, which produces a thin hygroscopic layer on its surface. Changes in humidity cause large changes in the impedance of this layer, and because the operating portion of the sensor is only its surface, its speed of response is fast — only a few seconds. An AC-excited Wheatstone bridge circuit can be used to measure the impedance. The output signal is nonlinear and can cover a range of 15–99% RH. This type of humidity probe can be mounted on the wall or in ducts (Figure 7.9o).

Relative humidity transducer 2.3" (58 mm)

1.0" (25 mm)

Temperature transducer

Threaded probe body (13 − 20 thread) 8

Duct-mount

Shoulder O-ring Wall-mount

FIG. 7.9o Polystyrene surface conductivity element. (Courtesy of General Eastern, which used to be Phys–Chem Scientific Corp.)

© 2006 by Béla Lipták

1452

Regulators and Final Control Elements

Permeable gold layer-vacuum coated 5" UNF 8 (16 mm) Aluminum oxide dielectric Ultra high purity aluminium substrate

3 inch (75 mm)

FIG. 7.9p Thin-film aluminum oxide capacitance element packaged as a probe. (Courtesy of Michell Instruments Ltd.)

Thin-Film Capacitance The largest number of manufacturers use some form of a capacitance measurement to detect humidity or dew point in air.

These units are available as transmitters with intelligent electronics and are capable of detecting low dew points while requiring reduced maintenance due to their stability.

Metal Oxide Sensor One variation is to form a capacitor by depositing a layer of porous aluminum oxide on a conductive aluminum substrate and coating the oxide with a thin film of condensation from evaporated gold. The aluminum substrate and the gold film serve as the electrodes of the capacitor (Figure 7.9p). When exposed to air, the water vapors penetrate through the gold layer into the aluminum oxide dielectric and are absorbed by it. The amount of water absorbed determines the capacitance registered by the sensor. Some suppliers “dope” the aluminum oxide dielectric with lithium chloride to extend its range down to dew points of −94°F (−70°C). The disadvantages of lithium chloride doping include the need for individual calibration, the slowing of the response, and the potential for shorting due to condensation. The accuracy of the aluminum oxide moisture sensor is low; each sensor requires a separate calibration curve, which is nonlinear; and the unit must be periodically recalibrated to compensate for aging and contamination. These sensors are designed for low dew-point measurements, but can be disabled if exposed to high humidity or if wetted. The advantages of these sensors include their small size, their probetype packaging, and the fact that their measurement range is wide and is well suited for the detection of low dew points. As a consequence of these characteristics, they are widely used in such applications as HVAC, where cost is the primary concern and accuracy of measurement is not critical.

Humidistat Design Features

Polymer Sensor Design variations of the capacitance-type humidity sensor include the replacement of aluminum oxide dielectric with hygroscopic polymer dielectrics. These units consist of a capacitive polymer sensor bonded to a resistive temperature sensor. The polymer sensor detects humidity directly, while dew point is calculated by a microprocessor, which also reads the temperature. These sensors are claimed to provide better linearity and allow the use of longer lead wires, because their low-frequency operation reduces the effects of stray capacitance. Other advantages include resistance to contamination by various chemical vapors. Polymer sensors are immune to condensed water and therefore can have a wider range than metallic oxide ones.

© 2006 by Béla Lipták

As shown in Figure 7.9q, the pneumatic humidistats are similar in design to thermostats, except that the bimetallic measuring element is replaced by a humidity-sensitive element. A common humidity-sensitive element is cellulose acetate butyrate. This substance expands and contracts with changes in relative humidity, and the resulting movement can be used to operate the flapper of the pneumatic humidistat shown in Figure7.9q or it can open and close the contacts of electric humidistats.

This humidistat is a reverseacting one, because an increase in ambient humidity causes the element to expand, which in turn increases the nozzle backpressure on the reverseacting relay, resulting in a lowering of the humidistat output signal

This humidistat is the “adjustable” proportional band type, because the distance “B” and therefore the ratio A/B are adjustable

Set-point scale Remotely adjustable pneumatic set point sets the “Normal” value “Mo” in Figure 7.9c

% RH 30 40 50 60 70 80

Humidity element

INCR. A

B

High volume (10−30 SCFH or 0.3−0.85 m3/hr) output signal to dampers and/or valves

Vent

Feedback bellows operates to reposition the flapper

Reverse-acting repeater relay 20 PSIG (140 kPa) air supply

0.01" (0.25 mm) diameter restriction

FIG. 7.9q Reverse-acting humidistat with pneumatic set point and adjustable proportional band. The use of the repeater relay increases the humidistat output flow capacity, and therefore this design is also referred to as “high volume.”

7.9 Thermostats and Humidistats

1453

Bibliography Linkage Element

Cover Control port

FIG. 7.9r Duct-mounted humidistat. (Courtesy of Johnson Controls Inc.)

The choice of humidistat features and characteristics is similar to those of thermostats. 1. They can be direct or reverse acting. 2. They can be high or low volume (relay type or nonrelay type). 3. The “normal” value can be set manually or by remotely adjustable pneumatic signal. 4. The proportional sensitivity determined by the A/B ratio can be fixed or adjustable. 5. Feedback bellows can be added to minimize the effects of variations in supply pressure or temperature, to reduce output leakage, and to increase the range of adjustability of the proportional sensitivity setting. Humidistats are also installed in air ducts to control upper and lower limits of the relative humidity of the air. One such duct-mounted device is illustrated in Figure 7.9r. This unit is usually set for either 85% RH or for 35% RH.

© 2006 by Béla Lipták

Air-Conditioning and Refrigeration Institute (ARI) Standard 750-76 (1976), “Thermostatic Refrigerant Expansion Valves,” ARI, 1976. American Society of Testing and Materials (ASTM) Standard B106-68 (1972), “Standard Method of Test for Flexibility of Thermostat Metals,” ASTM, 1972. ANSI: Z21.23-1975, “Gas Appliance Thermostats,” American Gas Association (AGA) standard, 1975. ASHRAE Handbook and Product Directory, 1980 Systems, American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc., 1980. ASHRAE Standard 17-75 (ANSI: B60.1-1975), “Method of Testing for Capacity Rating of Thermostatic Refrigerant Expansion Valves,”ASHRAE, 1975. ASTM Standard B389-65 (1975), “Standard Method of Test for Thermal Deflection Rate of Spiral and Helical Coils of Thermostat Metal,” ASTM, 1975. Carlson, D. R., “Temperature Measurement in Process Control,” InTech, October 1990. Eckman, D. P., Automatic Process Controls, New York: John Wiley & Sons, 1958. “Filled Systems Thermometers,” Measurements and Control, September 1992. Kerin, T. W., and Katz, E. M., “Temperature Measurement in the 1990s,” InTech, August 1990. Lipták, B. G., “Reducing the Operating Costs of Buildings by the Use of Computers,” ASHRAE Transactions, Vol. 83, Part I, 1977. National Electrical Manufacturers’ Association (NEMA) Standard DC131973, “Residential Controls — Integrally Mounted Thermostats for Electric Heaters,” NEMA, 1973. NEMA Standard DC3-1972, “Residential Controls — Low Voltage Room Thermostats,” NEMA, 1972. Spethmann, D. H., “Importance of Control in Energy Conservation,” ASHRAE Journal, February 1975. Underwriters’ Laboratories (UL) Standard UL-521 (ANSI: Z220.1-1971), “Standard for Fire-Detection Thermostats,” UL, 1971. Warmoth, D., “Temperature Measurement & Control,” Measurements and Control, February 1986.