3. Single-Circulator Distribution Systems

The simplest hydronic distribution system is a single series loop. A single loop of piping begins at the heat source, passes through several heat emitters, and ends back at the heat source. System operation is usually controlled by a single room thermostat in the heated space. An example of such a system with multiple fin-tube heat emitters is shown in Figure 3-1.

Because heat input to the entire building is regulated based on the temperature at one thermostat location, overheating or under-heating of rooms other than where the thermostat is located is possible. It is therefore critical that all heat emitters be sized for the load of their respective rooms, as well as the water temperature at their location within the piping circuit. The latter condition is often ignored in favor of sizing heat emitters based on the estimated average water temperature within the circuit.

When sized based on the average water temperature in the circuit, heat emitters near the beginning of the series circuit tend to be oversized, and thus release heat at a rate higher than necessary. Heat emitters near the end of the circuit tend to be undersized, and therefore release heat at a rate lower than required.

To prevent this, designers should keep track of the fluid temperature within the circuit as it cools while passing from one heat emitter to the next. There are various calculations that can be used for this. They range from relatively simple estimates to detailed mathematical models.

A relatively simple approach uses Formula 3-1 to calculate the temperature drop across each heat emitter.

$$\Delta T = {Q \over 500* f}$$

Where:

$\Delta T$ = temperature drop across the heat emitter (ºF)

$Q$ = rate of heat output from the heat emitter (Btu/hr)

$500$ = a constant based on properties of water. Change the 500 to 485 if a 30% glycol solution is used in the circuit, or to 450 if a 50% glycol solution is used.

Once the $\Delta T$ is determined for a given heat emitter, it is subtracted from the inlet temperature of that emitter to get the outlet temperature from that emitter. This outlet temperature becomes the inlet temperature for the next heat emitter in the series circuit.

Use of Formula 3-1 requires a value for the circuit flow rate. This can be determined through standard hydraulic analysis of the circuit, and the pump curve of the circulator used in the circuit.

idronics™ #16 describes methods and data for estimating flow rates in hydronic circuits.

idronics #16

In addition to circuit flow rate, designers must access data that gives the heat output of the heat emitters based on water temperature and flow rate. This data is typically published by heat emitter manufacturers. Figure 2-12 can be used to estimate the heat output of standard residential fin-tube baseboard at various average water temperatures.

There is also software that can simulate the combined hydraulic and thermal characteristics of single series circuits that use fin-tube baseboard heat emitters. One example is the Hydronics Design Studio. Figure 3-2 shows an example of this software.

This software analyzes a user-specified single serial loop containing up to 12 fin-tube baseboard heat emitters. The sizes of each baseboard, as well as the fluid temperature at locations within the circuit are reported.

The limiting factors in designing single series loop systems are temperature drop and flow resistance. In North America, series circuits have traditionally been designed around a temperature drop of approximately 20°F under design load conditions. However, there is nothing special about a 20ºF temperature drop. Systems can be designed to operate properly with circuit temperature drops greater than 20ºF, as well as less than 20ºF. In Europe, hydronic distribution systems are routinely operated with temperature drops of 30°F to 40°F. 

The advantage of designing around higher temperature drops is that the flow rate can be relatively low. This allows use of small-diameter piping and a smaller, less energy-consuming circulator. The disadvantage of a high temperature drop system is that heat emitters will need to be larger, especially those near the end of the series circuit. This is necessary to meet the load requirement with the lower temperature water. When the circuit is supplied from a conventional boiler, the water temperature returning from the last heat emitter needs to be high enough to prevent sustained flue gas condensation.

Systems that operate with low temperatures drops usually have relatively high flow rates. This may require increased pipe sizes to keep the flow velocity under a suggested limit of 4 feet per second. Systems designed around low temperature drop may also require larger circulators. However, because the heat emitters will operate at higher average temperatures, their size can usually be slightly reduced compared to sizes required in systems operating at higher temperature drops.

It is impossible to say what the optimum temperature drop is for a single series circuit. The answer depends on the cost of piping, circulators, heat emitters, electricity and several other factors. Designers should investigate different possibilities based on various combinations of supply water temperature and circuit flow rate.

Single series loops that contain one or more heat emitters with high flow resistance must be carefully designed to avoid problems. A single heat emitter, or other component, that has high flow resistance, can restrict flow through the entire series circuit, limiting its total heat output. A high head circulator with a larger power requirement may be required. The installation and life cycle operating cost of such a circulator should be considered against other modifications that could eliminate the flow restriction in the system. In most cases, a distribution system that uses a parallel arrangement of heat emitters will prove more appropriate for use with heat emitters having high flow resistance characteristics.

The use of high head circulators in single series loop systems can also cause excessive flow velocity through the small tubes or valves in the restrictive heat emitter(s). This can create noise and erosion inside some components.

Another drawback of a single series loop is that heat output control is limited to the control features present on the heat emitters. In the case of fin-tube baseboard, heat output can be reduced up to about 50% by closing the dampers on the enclosure. When fan-coils are used, the blower speed can be reduced to slightly limit heat output. Both these methods require manual adjustments to the heat emitters in response to changing load conditions. Most occupants either do not know these adjustments can be made or quickly tire of making them.

There are ways to retain the general concept of a single loop system while also providing a means of individual flow control through each heat emitter. One approach is to create a diverter tee distribution system.

Diverter tees can be installed in either the “upstream” location, as shown by the upper image in Figure 3-4, or in the “downstream” location, as shown by the lower image in Figure 3-4. The pressure differential created in either location, under the same operating conditions, is approximately the same.

A single diverter tee is usually sufficient to create flow through low resistance heat emitters such as a few feet of fin-tube baseboard or a panel radiator, assuming the heat emitter is above the piping loop containing the diverter tee.

Two diverter tees may be needed on heat emitters with high flow resistance. When two diverter tees are used, they create a greater pressure differential across the branch. This creates a higher flow rate through that branch. Two diverter tees are also suggested to overcome buoyancy forces when the heat emitter is located several feet below the distribution circuit, as shown in Figure 3-5.

When two diverter tees are used, there should be at least one foot of pipe between the tees to allow turbulence created by the upstream tee to dissipate before the flow enters the downstream tee.

Figure 3-6 shows how multiple branches can be supplied from a single loop system powered by a single circulator. 

Each branch includes a thermostatic radiator valve that can modulate flow through that branch based on the set room temperature. Flow through a given branch can be completely stopped if necessary. There will be slight changes in a given branch flow rate as other branches open, close or modulate flow rate.

Any heat emitter on what would otherwise be a single series circuit can be selected to have independent temperature limiting control by piping it into the circuit using one or two diverter tee(s) and some type of flow-regulating valve.

Figure 3-7 shows a series circuit where one heat emitter is connected using a diverter tee arrangement, while the other heat emitters remain in series. This is a convenient way to limit the heat output of one heat emitter without significantly affecting the remainder of the heat emitters. This arrangement could be used, for example, in a guest bedroom that only needs to be heated to normal comfort temperatures a few days each year. Keep in mind, however, that the branch connected to the system using the diverter tee cannot independently call for heat. There must be flow in the main circuit in order to have flow through the branch. If independent on/off control of several heat emitters is needed, it is better to use a parallel distribution system.

When non-electric thermostatic radiator valves are used for individual room temperature limiting, the designer must provide controls to operate the heat source and distribution circulator whenever the building might need heat. One approach is to equip the system with an outdoor reset control that automatically turns on the heat source and distribution circulator when the outdoor temperature falls below a preset value. When such a control system is used, the thermostatic valves on the heat emitters serve as temperature-limiting devices for their respective rooms. Since this approach requires constant circulation in the main distribution circuit, all piping should be insulated to minimize heat loss.

Whenever diverter tees and thermostatic radiator valves are used to reduce or totally stop water flow to a heat emitter, it is crucial that the heat emitter and the piping leading to it are protected against freezing during cold weather. Some thermostatic radiator valves have a freeze-proof minimum setting that only allows a trickle of heated water flow through the branch to provide this protection.

There is another way to construct what appears to be a series piping circuit, but retain the ability to regulate flow through each heat emitter. It is called a series bypass system, and is commonly used with panel radiators that are equipped with thermostatic radiator valves.

The H-pattern diverter valve allows up to three panel radiators to be configured for individual flow control. However, as is true with a diverter tee system, as well as a single series circuit, there is a drop in water temperature as it passes each active radiator. Designers should calculate this temperature drop based on the flow in the circuit, and the rate of heat release from each panel radiator. Formula 3-1 can be used for this calculation.

The circuit should be designed assuming that all radiators are active simultaneously. Each panel radiator should be sized based on its entering water temperature, which decreases along the circuit as the flow passes each panel.

All of the distribution systems discussed thus far have heat emitters connected in a series or “quasi-series” arrangement.

The disadvantage of these arrangements is that the water temperature supplied to each heat emitter will be lower than that supplied to the upstream heat emitter — assuming all heat emitters are operating. Each time a heat emitter in a diverter tee or series bypass system goes from inactive to active, the water temperature to the downstream heat emitters changes. 

This is “tolerable” when accompanied by accurate design calculations, but it is not ideal.

Parallel distribution systems do not experience this temperature drop from one heat emitter to the next. In most cases, it’s safe to assume that the water temperature supplied to each heat emitter in a parallel distribution system is the same. The exception being where the piping lengths to one heat emitter are significantly longer, or subject to significantly higher heat loss, compared to the piping supplying the other heat emitters.

There are several types of parallel distribution systems that can be supplied by a single circulator. They will be discussed in this section. There are also parallel distribution systems that use multiple circulators, which will be discussed in the next section.

One of the most common parallel distribution systems uses a separate circuit to heat (or cool) each zone of a building. Flow through each zone circuit is allowed or prevented by an electrically operated zone valve. A typical piping layout is shown in Figure 3-11.

This system uses a single fixed-speed circulator to create flow in any zone circuit that has an open zone valve. Since all zones are supplied from a common header, they have the same supply water temperature. The heat emitters used in each zone circuit can be different, but they should be sized based on the supply water temperature produced by the heat source at design load conditions.

This system also includes flow-balancing valves on each zone circuit. This allows flow rates within each parallel circuit to be adjusted in proportion to the percentage of the total load supplied by that circuit. The flow-balancing valves can also be completely closed to isolate the zone valves below if servicing is needed.

The zone valves are located on the supply side of each zone circuit. This placement is preferred because it prevents heat migration into the zone circuit when the zone valve is closed.

Each parallel circuit also includes a purging valve where it connects to the return header. This placement allows bulk air to be quickly purged from each circuit when the system is commissioned. It also allows each circuit to be fully isolated if servicing is needed.

The piping within the dashed lines in Figure 3-11 is called the “common piping.” It is the piping through which all system flow passes. To provide optimal performance, the common piping should be selected to provide a very low head loss. Doing so will minimize change in flow rate through a given zone circuit when another zone valve turns on or off.

A differential pressure bypass valve is connected between the right ends of the supply and return headers. Its function is to prevent excessively high differential pressure from developing between these headers when only one or two of the zones are on. It should be set to a pressure that is 0.5 to 1 psi above the differential pressure across the headers when all zone valves are open. A differential pressure bypass valve should always be used when valve-based zoning is used in combination with a fixed-speed circulator.

idronics™ #5 provides a more complete description of how to size and select differential pressure bypass valves.

idronics #5

In systems where a fixed-speed circulator is used in combination with zone valves, the circulator used should have a relatively “flat” pump curve. Figure 3-12 illustrates the difference between such a pump curve compared to a “steep” pump curve.

Notice that the differential pressure bypass valve shown in Figure 3-11 is no longer present. It is not needed when a pressure-regulated circulator is used. If a variable-speed pressure-regulated circulator is retrofitted to a system that has a differential pressure valve, that valve should be completely closed or removed.

Systems of this configuration, and using variable-speed pressure-regulated circulators, should also have low flow resistance common piping. The variable-speed circulator should be set for constant differential pressure mode. The differential pressure “setpoint” should be adjusted to that required when all zone valves are open (e.g., under design load conditions). Once set, this circulator will automatically adjust speed in an attempt to maintain the differential pressure between the supply and return headers approximately constant. When a zone valve closes, the circulator will reduce speed, which also reduces electrical power input. When a zone valve opens, the circulator will increase speed.

It is also advisable to include a magnetic dirt separator in any system using an ECM-type variable-speed circulator. These circulators have powerful permanent magnets inside their rotors. Iron oxide particles are attracted to these magnets. Use of a magnetic dirt separator will reduce the presence of such particles in the system, and thus reduce any potential for them to lodge within the circulator.

idronics™ #5 provides a more complete description of how variable-speed pressure-regulated circulators can be used in hydronic systems.

idronics #5

It is also possible to create a parallel distribution system using manifolds, as shown in Figure 3-14.

The manifold replaces the need for site-built headers. It consolidates multiple zone connections into a common assembly. This type of distribution system is often referred to as a “homerun” system. The name stems from each piping path making a complete round trip from the supply manifold, to the heat emitter, and back to the return manifold.

As is true in systems using zone valves, the heat emitters used in each zone circuit can be different, but they should all be selected and sized based on the supply water temperature produced by the heat source under design load conditions.

Figure 3-17 shows how multiple panel radiators, each equipped with a thermostatic radiator valve, can be supplied by a manifold distribution system.

The flow rate through each panel radiator is regulated by the thermostatic radiator valve, which responds to changes in room air temperature. Thus, each panel radiator creates a separate zone within the building.

Some radiators, such as those depicted in Figure 3-17, are supplied with the body of the radiator valve integrated into the radiator. The installer just screws on a thermostatic actuator to allow automatic actuation of the valve’s stem.

A variable-speed pressure-regulated circulator is ideal for this type of system. It would be set for constant differential pressure mode, and would automatically speed up or slow down as thermostatic valve actuators allowed more flow or reduced flow. The circulator could be turned on using a master thermostat within the building, or whenever the outdoor temperature dropped below some threshold at which heating was deemed necessary.

The manifolds may be valveless, or they may be equipped with isolation valves. The latter has the advantage that each circuit can be isolated at the manifold station if service is necessary.

If zoning control will be accomplished using electric thermostats, the manifold station can be equipped with manifold valve actuators, as shown in Figure 3-18.

Some manifold valve actuators are equipped with an isolated end switch. This switch closes when the actuator reaches its fully open position. A low power circuit through the end switch is used to signal that the circulator and heat source need to operate.

idronics™ #5 provides more details on how to wire manifold valve actuators with other electrical components in the system.

idronics #5

Another common parallel piping method, one this is especially common in commercial buildings, is called a 2-pipe direct return distribution system. Figure 3-19 shows an example.

Each heat emitter is located within a “crossover” connected to the supply and return “mains.” If the supply main is properly insulated, heat loss will be minimal, and each heat emitter will be supplied with water at approximately the same temperature.

Notice that the pipe size of the mains decreases as one moves from the heat source toward the farther end of the distribution system. This is possible because the flow in the mains decreases as the mains pass outward beyond each crossover.

The heat emitter closest to the circulator on the supply main is also closest to the circulator on the return main. The next heat emitter connected to the supply main has a greater length of piping between it and the circulator. The heat emitter farthest away from the circulator has the longest overall piping path length.

If each crossover in the system has the same flow resistance, the highest flow rate will be through the shortest piping path. If uncorrected, this situation reduces the flow rate through heat emitters located farther away from the circulator.

This situation can be remedied through use of properly set balancing valves on each crossover. These valves can be adjusted so that the flow rate through each crossover is in proportion to the rate of heat release from the heat emitters in that crossover as a percentage of the total rate of heat release from the system.

idronics™ #8 provides details on how to balance a 2-pipe direct return distribution system.

idronics #8

Flow through each crossover can be allowed or prevented by installing zone valves, as shown in Figure 3-19. The flow rate through each crossover can also be modulated when a modulating valve is used in lieu of a zone valve. Either configuration allows each heat emitter to be independently controlled, and thus room-by-room comfort control is possible when the system is properly configured.

As with other parallel distribution systems, different types of heat emitters can be used on the crossovers provided they are all sized for the same supply water temperature.

The system shown in Figure 3-19 uses a variable-speed pressure-regulated circulator. This is the same type of variable-speed circulator described in previous parallel distribution systems. However, when used in a 2-pipe direct return system, this circulator should be set for proportional differential pressure mode. 

In this mode, the differential pressure created by the circulator decreases in a pre-determined manner as the flow rate decreases.

Proportional differential pressure control accounts for the head loss in the supply and return mains. In a 2-pipe direct return system, this head loss is a significantly higher percentage of total head loss compared to the head loss in the common piping of a manifold-type distribution system. Proportional differential pressure control approximates the ideal scenario of maintaining constant differential pressure across each crossover, regardless of which crossovers are active or inactive. Most circulators that offer constant differential pressure control can also be set for proportional differential pressure control.

Another variation of the “2-pipe” approach is known as a 2-pipe reverse return system. An example is shown in Figure 3-20.

Like a 2-pipe direct return system, this system connects multiple crossovers, each with a heat emitter and valve trim, between common supply and return mains. Thus, each crossover is supplied with water at approximately the same temperature.

Notice that the crossover closest to the circulator along the supply main is now farthest from the circulator on the return main. This is what distinguishes a 2-pipe reverse return system from a 2-pipe direct return system.

Reverse return piping attempts to make the flow resistance of each flow path approximately equal, beginning at the circulator discharge, passing out along the supply main, through a crossover, back along the return main, and finally back to the circulator inlet. Emphasis is on the word approximately.

An “ideal” reverse return system, in which each crossover had the same piping, heat emitter, and valve trim, and with the ability to maintain the same head loss per unit length along the mains piping, could yield exactly the same flow resistance along each flow path. Such a system would be “self-balancing.” There would be no need of balancing valves in each crossover.

Unfortunately, this is unrealistic from a number of standpoints. First, not all systems will have identical crossovers. Some heat emitters in the system may be different from others in both heating output and flow resistance. The amount of piping within a given crossover may also be different from that in other crossovers. 

Second, there are only a finite number of pipe sizes that can be used for the mains piping. Although changing pipe size based on the flow rate at different locations along the mains is possible, it is impossible (or highly impractical) to create exactly the same head loss per unit length of piping.

Still, even though any real reverse return system can only approximate ideal conditions, it will be closer to “self-balancing” compared to a direct return system. The amount of balancing required is usually less, and this implies that less head energy will be throttled away in balancing valves. Less throttled head energy suggests that a smaller, lower power circulator could likely be used.

Reverse return systems are well-suited to applications where the supply and return mains run side by side and make a loop from the mechanical room around the building and back to the mechanical room. They are not well suited to “dead end” layouts such as shown in Figure 3-21.

“Dead end” layout requires a long length of the largest size piping to return all flow from the location of the farthest crossover. While this will work, it can add significant cost to the system relative to reverse return systems that “loop” around the building, as depicted in Figure 3-20.

A variable-speed pressure-regulated circulator, set for proportional differential pressure control, is also appropriate for 2-pipe reverse return systems.

The target flow rates within each branch of a parallel distribution system should be proportional to the design load heat output of that branch as a percentage of the total design heat output of the distribution system.

Thus, if the heat emitter in a given branch is sized to release 15% of the total heat output of the system, that branch should operate at approximately 15% of the total system flow rate at design load conditions. This criterion keeps the temperature drop across each branch approximately equal. 

The circulator used in a parallel distribution system is typically sized by identifying the branch that would have the highest head loss, assuming all branches are active. This branch might have to be determined by trial and error, in which several possible branches are identified, and the head loss through each branch is calculated and compared with that of the other branches.

When calculating the head loss of each branch, the design flow rate should be assumed in each piping segment between the circulator discharge and the circulator inlet. Different pipe sizes will likely be involved along with different flow rates. The head loss of each piping segment at its associated flow rate must be calculated. Likewise, the head loss through all common piping components, (e.g., heat source, air separator, etc.) should also be calculated at whatever flow rate would be present with all branches on. Once the head loss of the individual branch is calculated, it can be added to the head loss of the common piping to determine the total head loss of the complete flow path. This is illustrated in Figure 3-22.

idronics™ #12 and #16 discuss methods for calculating head loss in systems using smooth piping such as copper or PEX tubing. When steel piping is used, head loss can be determined using the Darcy-Weisbach formula, which is also discussed in idronics #16. 

idronics #12    idronics #16

Once the path with the greatest head loss is determined, that head loss, combined with the total system flow rate (e.g., the sum of the flow rates through all crossovers operating simultaneously at design load conditions), will set the required “duty point” for a circulator. Pump curves from candidate circulators can then be examined to find a suitable model.

Parallel direct return piping systems with up to 12 crossovers can also be simulated using the Hydronics Design Studio software. Figure 3-23 shows an example of a system with 10 user-defined crossovers, including balancing valves, and use of a differential pressure bypass valve.