Left alone, a 2-pipe direct-return loop self-unbalances — the nearest load gets too much flow and the farthest load too little.

True. In direct return, each load goes "back the way it came," so the near load has a short total trip and the far load crosses the whole building and back. Water takes the path of least resistance, so flow favours the near loads. Every load needs a balancing valve and someone to set them — cheap to pipe, fussy to commission.

In a direct-return loop, why does the nearest load tend to hog flow?

Its total pipe path is shorter, so it has less resistance — and water takes the path of least resistance.. Shorter loop = less resistance, and like current, water splits toward the path of least resistance. The near load over-draws and the far load is left short until balancing valves add restriction to the near branches to even it out.

How does a 2-pipe reverse-return loop come out roughly self-balancing?

The return main runs the same direction as the supply, past every load, then loops back — so the first load supplied is the last to return and each load's total path length is about equal.. The trick is in the return: instead of heading straight back, it continues past every load and only then doubles home. First-supplied is last-to-return, so everyone's total pipe path — and therefore resistance — comes out roughly equal, and flow splits close to evenly on its own.

Reverse return's self-balancing comes for free — it uses the same amount of pipe as direct return.

False. It costs an extra full run of pipe: the return doubles back the whole length of the building, so you're essentially doubling the return run. The trade is real — you spend pipe and labour to save commissioning effort. (You'll still fit balancing valves, but they're trimming, not heavy lifting.)

In a primary-secondary "twin-T" system, how much of the building's water actually flows through the boiler?

Only the slice the injection pump diverts from the primary loop; most of the building's water never touches the boiler.. The boiler sits on its own short primary loop with a dedicated pump holding constant flow through it. A separate injection pump taps hot water off the primary into the system loop. Only that injected slice ever touches the boiler; the rest of the building's water circulates the loads and mixes with the injection downstream. That decoupling is why a 100-GPM boiler can serve a 200-GPM building.

Walking a plant, you find two tees piped right next to each other with a short stub between them — it looks like a short-circuit straight across the boiler. The pipefitter swears it's correct. Who's right? primary loop ──┬──┐ (closely-spaced tees) │ │ short common pipe system loop ───┴──┘ injection pump on the bridge

The pipefitter — those are the "closely-spaced tees," a common pipe with near-zero Δp that decouples the loops so the three pumps don't fight each other.. The closely-spaced tees are the entire mechanism, not a mistake. With essentially no pressure drop across that little stub, the boiler, injection, and system pumps each do their own thing instead of fighting one another's discharge pressure. That common pipe is the whole reason a building with swinging demand can be served by a boiler that wants a steady diet. Recognise the twin-T and the plant snaps into focus.

In a twin-T system, what sets the supply temperature delivered to the loads?

The injection pump's speed — it controls how much hot primary water mixes into the system loop.. The injection pump is the control point. Colder outside, building wants more heat → the injection pump speeds up, pulling more hot water from the primary into the system, and supply temperature to the loads climbs. The boiler's own flow rate never changes — that constant flow is what protects it.

In a twin-T arrangement the boiler pump holds constant flow through the boiler no matter what the building demand does, and that constant flow is what protects the boiler.

True. Right. The dedicated boiler pump keeps a steady flow through the boiler always — shielding it from the low-flow and thermal-shock conditions a variable building would otherwise impose. The building's demand swings are absorbed by the injection pump and the system pump, not by the boiler's flow.

A boiler pump holds 100 GPM (6.3 L/s) through the boiler. The injection pump is pulling 40 GPM (2.5 L/s) from the primary into the system loop. How many GPM stay in the primary and go straight back to the boiler to be reheated?

60 GPM. 100 GPM circulates the boiler; the injection pump diverts 40 GPM into the system, so 100 − 40 = 60 GPM stays in the primary (in SI: 6.3 − 2.5 = 3.8 L/s) and returns directly to the boiler. The system loop, meanwhile, moves its own (larger) flow through the loads — most of which never touches the boiler.

On a reverse-return loop, do you still install balancing valves?

Yes, but only for fine trimming — the geometry already does the heavy lifting.. Reverse return gets you close to balanced for free, but real loops still want balancing valves for the last bit of trim. The difference from direct return is that here they're polishing an already-even split, not fighting a strong near/far imbalance.

Practice · Hydronic Loops

Hydronic Loops Quiz Hydronics

Read the Hydronic Loops lesson first, then test yourself — ten questions on hydronic piping topology: why direct return self-unbalances and reverse return self-balances, and the primary-secondary "twin-T" with its closely-spaced tees, constant boiler flow, and injection-pump control point. Each reveal links back to the section it came from.

← All practice