Controlling heat output by adjusting the flow rate isn't as intuitive as it might seem.

If you install hydronic radiant floor heating systems, chances are you work with valved manifolds. Each circuit has its own valve that can be used to adjust the flow rate through the circuit — what we call “balancing.”

But what’s the purpose of balancing tubing circuits? Is it:

a. To assure a certain flow rate in each circuit.

b. To assure that every circuit on the manifold has the same flow rate.

c. To regulate the heat output of each circuit.

The answer is c. Heat output, not flow rate, is the desired “end product” of balancing. Heat output from a floor circuit can be adjusted by changing the water temperature supplied to the circuit, or by adjusting the circuit’s flow rate. The latter method, although theoretically plausible, tends to be a bit “touchy” at times. This month we’ll look at why controlling heat output by adjusting flow rate is not as intuitive as it might seem.

Anything BUT Proportional

Before attempting to balance a radiant heating circuit, it helps to know how its heat output will change as flow changes. It’s reasonable to assume that reducing a circuit’s flow rate by 50 percent would also reduce its heat output by 50 percent. Unfortunately nature doesn’t agree.

The graph in Figure 1 shows the heat output of a typical 300-ft. floor heating circuit being supplied with 105 degrees F water over a wide range of flow rates. At 2 gpm, heat output is about 6,800 Btu/hr. Now imagine closing the balancing valve on the circuit until its flow rate dropped to 1 gpm. After allowing the heat output to stabilize at the reduced flow a Btu meter would tell you that heat output has only dropped to 6,200 Btu/hr. That’s still 91 percent of the heat output the circuit delivered at twice the current flow rate. Seems that reducing the flow rate at the upper end of its range doesn’t have much of an effect on heat output.

Now imagine you cut the flow rate in half again. At 0.5 gpm the circuit delivers 5,000 Btu/hr. Still about 73 percent of the original heat output at only 25 percent of the original flow rate.

As the flow rate keeps getting reduced, the heat output drops faster and faster. Interestingly, at 0.2 gpm (only 10 percent of the original flow rate) the circuit still puts out about 3,400 Btu/hr. (about half of its full flow heat output). That’s right. Half the heat output at 10 percent of the flow! Since the circuit’s heat output must be zero when the flow rate is zero, the remaining 50 percent of the circuit’s heat output must be regulated between 0.2 and 0.0 gpm. Doing so with any degree of precision requires a specialized balancing valve and a steady hand. More on this later.

Warm Over Here … Cool Over There

As the flow rate is lowered, the temperature drop along the floor heating circuit increases. This effect is shown in Figure 2. Notice that at full flow (2 gpm) the ³T across the circuit is about 7 degrees F. At 1 gpm, the ³T is still acceptable at about 12 degrees F. However, at 0.5 gpm, where the circuit is still delivering 73 percent of its full heat output, the ³T has increased to about 21 degrees F. Such a wide ³T is acceptable across the coil of an air-handler unit, or when using a panel radiator, but it could easily lead to complaints about uneven surface temperature in a floor heating system. Taking it one step further, at 0.2 gpm, where the circuit is still delivering about half its full flow heat output, the ³T has increased to over 30 degrees F.

I’ve plotted the water temperature at various positions along the 300-ft. circuit in Figure 3. Notice how the water temperature drops fast at first and then slower farther along the circuit. The first half of the circuit is delivering about 73 percent of its total heat output. There’s little doubt where the family dog will curl up on that floor during a cold winter night.

All Plugs Are Not Equal

In an ideal hydronics world, the heat output from a distribution circuit would be proportional to the flow rate through it. For example, closing a balancing valve 50 percent would reduce the heat output of the circuit by 50 percent, and so forth. It is possible to make this happen. Doing so requires an “equal percentage” balancing valve with a specially shaped plug rather than the flat disk used in some globe valves and manifold valves. Valves with equal percentage plugs are designed to open the gap between the plug and its seat very slowly over the first portion of the stem lift. When the stem has lifted through 50 percent of its travel, the valve only allows about 10 percent of its fully-open flow to pass through. This compensates for the heat output versus flow rate characteristic of the hydronic circuit we looked at earlier. The combined effect yields heat output that varies close to proportionally with the lift of the valve stem.

Valves with flat disks are characterized as “quick-opening.” They allow flow through the valve to increase rapidly as the stem starts to lift. The quick-opening characteristic of flat disk valves, combined with the way heat output varies with flow rate, makes adjusting the heat output of a floor circuit “touchy” at lower flow rates. Very small changes in valve stem position (even 1/100 of an inch of stem lift) can result in significant changes in heat output. It’s like trying to precisely control a racecar with an 800-horsepower engine in slow, stop and go traffic. Theoretically it’s possible — but you better have a nimble foot on the gas pedal!

Some of you have probably learned this on the job. Perhaps you remember getting a callback where one room of a building in which you installed floor heating was consistently overheating. No problem, you thought, I’ll just turn down the manifold valve on that circuit a bit. You went back to the job and closed the manifold valve from fully open to half open. Recognizing it would take several hours for the thermal mass of the slab to adjust itself to the reduced heat input, you told the owner, “Give it a couple of days to settle in, then call me if it’s not OK.”

Sure enough, two days later the owner called with the same complaint. So back you went to close the valve a little more, this time leaving it about one-fourth open. If you got lucky, that last adjustment did the trick. If not, you probably made a couple more trips trying to tweak the circuit into submission. After each adjustment, you had to wait several hours for the system to settle out at the new flow rate before knowing whether you closed the manifold valve too much, too little or finally found the “sweet spot” where the circuit’s heat output matched the room’s load.

Holding The Future In Balance

Setting each of several balancing valves in a multiloop hydronic system such that each circuit yields the proper heat output can be a real challenge. Every time one valve setting is changed, the flow rates in all the remaining parallel circuits also change. Field experience certainly helps you close in on the solution faster, but getting acceptable results may still require a couple of trips back to the site to tweak those valves in search of the elusive settings that deliver the right amount of heat to each room.

Believe it or not, the entire process of balancing parallel hydronic circuits can be simulated on a computer. Imagine if you could tell your computer what circuits you have, what the heat output of each needs to be and what balancing valves are installed. It digitally digests all this information in a couple of seconds and tells you the setting of each balancing valve. I’m currently working with my colleage Mario Restive to set up the models for this. Thus far the math seems like something only a direct descendent of Albert Einstein could appreciate. It’s going to take some time. In fact don’t expect any results until the next millennium.

In the meantime, small changes in the stem position of a balancing valve as it starts to open can result in big changes in heat output. Keep a steady hand on those valves, and remember a little twist goes a long way!