Sometimes one less circulator makes sense.

Figure 1

If one circulator is good, two must be better, right? If boosting flow rates is a legitimate way of correcting underperforming radiant systems, just slap in another circulator and call it a day, right? If a standard “header-type” distribution system works OK, just think how good it will work with another “helper” pump.

Imagine a conscientious plumbing and heating contractor who has installed zoned baseboard systems for years, and has always had those systems perform as expected. Along comes a great opportunity for his company to install a system in an 10,000-square-foot custom home where the owner wants radiant heating in some areas, baseboard in others, plenty of hot water for the eight showers, and no snow on the steps or sidewalks.

Our man sees this as the break he’s been waiting for. A chance to jump into the high-end housing market with systems that have all kinds of bells and whistles.

Down at the supply house, he’s told this type of job needs primary/secondary piping. “You know - just throw in a primary pump and connect all the zone circuits to the loop using closely spaced tees,” he’s told. After one of the counter guys sketches out a primary/secondary system, it all seems to make sense. Besides, in the back of his mind he’s sure he heard or read something about the “magic” of closely spaced tees.

So he loads the truck full of piping, circulators and all the other hardware needed to install this system, stuffs the sketch of primary/secondary piping in his pocket, and heads to the jobsite. On the drive over, he’s even thinking that if the piping is all neatly installed, one of the trade magazines will publish a photo of his installation.

What he initially creates is shown in Figure 1, and it is indeed a pure series primary/secondary system.

When the system is turned on, flow in all the circuits seems to be fine. However, the heat output of the 3rd and 4th secondary circuits along the primary loop are well below expectations. A few temperature readings along the primary loop gets followed by a hard swallow. He realizes that he didn’t account for the temperature drop along the primary circuit in the downstream direction. The amount of heat being extracted by secondary circuits No. 1 and No. 2 doesn’t leave much left for No. 3 and No. 4.

Visions of the photo op quickly fade. He starts wondering if the advice he got at the supply store was any good, and thinks he now knows how to solve the problem on his own. So out come the tools, and a few hours later the system has morphed into the schematic shown in Figure 2.

This piping design is based on his previous familiarity with header-type piping. He remembers how this approach provided the same water temperature to a zone circuit, and reasons that it should work just as well, if not better, with one extra circulator (e.g., the “primary circulator”). He adds piping and a valve to the end of the headers to form a loop. He feels the valve is necessary to “balance” flow in the primary loop.

He’s now created a mixture of what he knew worked in the past with what the guys at the supply house told him was the latest and greatest piping technique for sophisticated hydronic systems.

I’ve seen this piping configuration more times than I’ve cared to. It usually comes from someone who’s just gone through a “conversion” to the enlightened use of primary/secondary systems, but who also clings to the familiarity of header-type piping. Sounds like a nice compromise, doesn’t it? Too bad the laws of physics don’t like compromises.

When the reworked system is turned on, and a single zone calls for heat, that zone promptly gets heat. Unfortunately, so do all the other zones with inactive circulators!

Figure 2

Take This Piping And Shove It!

To figure out what went wrong with this well-intentioned but hydraulically challenged system, we need to start with the inescapable truth that pressure always decreases in the downstream flow direction. Start at Point A on the schematic in Figure 3.

Let’s say the pressure at Point A is 15 psi. Assume the primary circulator is wired to operate when any of the zone circuits calls for heat. As flow moves along the upper piping, pressure is dropping. The pressure at Point B is lower than at Point A. How much lower? It depends on piping diameter, length and flow rate, but rest assured it can be calculated. For now let’s just estimate there is a 0.5 psi pressure drop from A to B.

Flow then passes through the partially closed valve connecting the two headers. The pressure drop along this segment of the path obviously depends on the setting of the valve. The more restrictive its setting, the greater the ∆P. In any case, the pressure at Point C is lower than at Point B. Let’s say the pressure drop across the valve with the “primary circulator” operating is 2 psi.

Finally, flow passes back across the lower piping. The pressure at Point D is again lower than at Point C. Again, let’s assume the pressure drop along the header from C to D is 0.5 psi.

This makes the total pressure drop from Point A to Point D 3 psi (0.5 psi + 0.5 psi + 2 psi). Thus, the pressure at Point D is 15 - 3 = 12 psi.

The 3 psi pressure differential from A to D is significantly higher than the forward-opening pressure differential of a typical spring-loaded check valve or flowcheck in a zone circuit (typically about 0.3 psi). The result: Flow develops in ALL zone circuits, leading to overheating in many of them when only a single zone really needs heat. Been there, seen it, and trust me, it’s very real.

Figure 3

Lose The Primary Pusher

The easiest way to prevent this situation and save money in the process is to just stick with a header-type system. Don’t morph it into a piping aberration by adding another “helper” circulator and calling it a primary circulator.

Besides eliminating the zone heat migration problem, this approach eliminates the installation cost and, more importantly, the operating cost of a primary loop circulator. The latter can easily exceed $1,000 over the life of the system, even in a small residential system equipped with an 80-100 watt primary loop circulator.

If the primary loop is legitimately in the system to serve other secondary loads, connect the header assembly to the primary loop with a pair of closely spaced tees as shown in Figure 4. This hydraulically isolates the header assembly from the primary loop as well as from any other circuits in the system.

Be generous when it comes to header sizing. Short, fat headers are good - long, skinny headers are bad. The shorter the header and the greater its diameter, the lower the pressure drop along its length. In an ideal case, there would be zero pressure drop along a header regardless of the flow through it. This would allow perfect hydraulic separation of all circuits attached to the header. Realistically we can’t get the pressure drop to zero, but we can make it very low, thus the term “low-loss header.”

My suggestion is to size headers for a maximum flow velocity of 2 feet per second under full design flow rate. This provides insignificant pressure drop and allows all the hydraulic separation effect needed. Any minute pressure differentials are easily contained by the forward-opening resistance of spring-check or flowcheck valves in the zone circuits.

Figure 4

Stick To Standards

If you want to install primary/secondary piping systems, learn exactly how they are supposed to be assembled and don’t “morph” your design by thinking back to header systems. Remember, if you don’t have closely spaced tees, you don’t have primary/secondary piping.

Likewise, if you are relying on a low-loss header for hydraulic separation, don’t throw an extra circulator into the mix under the misguided assumption that it acts as a “helper” for the zone circulators.

Keep in mind that primary/secondary piping always involves one more circulator than an equivalent system using a low-loss header. Both piping concepts work when properly configured, but one obviously spins the electrical meter a bit slower.

Circulators are designed to be “pushy,” but don’t let them shove hot water were you don’t want it.