Commissioning a Glycol Loop That Will Not Reach Design Flow
You are commissioning a new closed-loop glycol system: three 4 × 8 ft flat-plate collectors (96 ft² aperture), 30% propylene glycol, a 1/12 hp bronze pump rated 10 gpm at 8 ft of head, 3/4-in. copper supply and return, and a 30-gal diaphragm expansion tank precharged to 12 psi. Cold fill pressure is 18 psi at the mechanical room. At startup the flow meter reads 2.6 gpm, the collector outlet sensor reads 112 °C, the tank bottom is 38 °C, the pump is noisy, and the pressure gauge swings between 9 and 21 psi. The sight glass shows bubbles, and the controller display shows "pump on" continuously. The installer says the system "just needs to run a few days to settle." Design flow for the array is 0.03–0.05 gpm/ft².
1. Calculate expected flow, interpret the readings, and give a ranked diagnosis.
Read the worked answer
Expected flow = 96 ft² × 0.03–0.05 gpm/ft² = 2.9–4.8 gpm. Measured 2.6 gpm is below even the low end.
The ΔT is the loudest signal: 112 °C outlet minus 38 °C tank bottom = 74 K. Design ΔT at full flow is typically 10–20 K. A 74 K ΔT with a noisy pump, swinging pressure, and bubbles in the sight glass points to air and/or cavitation, not a settling-in issue.
Ranked diagnosis:
- Air in the loop - incomplete purge, a high-point air trap, or a vacuum breaker/air vent stuck open. Bubbles plus fluctuating pressure support this.
- Pump cavitation - low suction pressure or entrained air collapsing at the impeller; the 9 psi low-side swing is near the point where a 1/12 hp pump loses prime.
- Expansion tank issue - precharge 12 psi vs. 18 psi fill means the bladder is partly compressed; if the tank is undersized or waterlogged, pressure spikes to 21 psi and flow falls.
- Mechanical restriction - throttled valve, reversed flow meter, clogged strainer, or a solder blob in a fitting.
"Let it run a few days" is not acceptable: sustained 112 °C with air will degrade glycol and can trip the high-limit.
2. Give the order of checks and the corrective actions, including what values you would accept at sign-off.
Read the worked answer
Order matters - verify the cheap, reversible things before touching the pump.
Valves and flow path: confirm every isolation valve is fully open, the flow meter is oriented correctly and not installed backward, and the strainer is clean.
Pressure and purge: raise cold static pressure to the correct value for the system height (about 0.43 psi/ft of rise above the fill point, plus 5–10 psi margin). Then purge with a purge pump at high volume - typically 1–2 gpm per collector - with the expansion tank isolated, until no bubbles appear at the return. Isolating the tank during purge is essential; otherwise the purge flow just compresses the bladder.
Air elimination: verify the high-point air vent and vacuum breaker operate, and that no local high spot traps air. Recheck after the loop reaches temperature, because dissolved air comes out of solution when hot.
Pump: confirm rotation, that the pump is not air-locked, and that suction piping is not undersized or throttled. Recheck the pump curve against measured head; if it cannot make design flow at the actual head, replace it.
Expansion tank: set precharge to the cold fill pressure at the tank, confirm tank volume against loop volume and maximum temperature, and support/isolate it properly.
Sign-off values: flow 3–5 gpm; stable pressure with no more than a few psi swing; ΔT roughly 10–20 K at steady state; no bubbles; no cavitation noise; controller switching pump on at 8–12 K and off at 3–4 K, not running continuously.
3. The controller is set to run the pump continuously and the outlet sensor is strapped to the tank inlet pipe 3 ft from the collectors. What is wrong, and how do you correct the control strategy?
Read the worked answer
Continuous pump operation is wrong for a differential-controlled solar loop: at night and in low-light conditions it circulates cooler tank water through the collectors and creates standby losses, and it masks the real problem (air/low flow) by keeping the loop hot.
Sensor placement: a tank-inlet strap 3 ft from the collectors reads pipe temperature, not collector temperature. It is influenced by the tank, by insulation gaps, and by ambient. Move the collector sensor into a dry well at the collector outlet header, insulated from ambient, within about 6 in. of the collector. The tank sensor belongs low in the tank, in a well, measuring the coldest usable water.
Control settings to establish: • Pump on at ΔT 8–12 K above the tank sensor. • Pump off at ΔT 3–4 K. • Collector high-limit at the manufacturer's stagnation rating (often 120–130 °C) to protect glycol. • Freeze protection set per the glycol concentration (typically around −25 °C for 30% PG), not a generic setpoint. • Anti-short-cycle delay of a few minutes.
After moving the sensor and resetting the controls, re-trend for at least one full day: the pump should cycle with sunlight, ΔT should settle in the teens, and the tank should stratify with a hot top and cool bottom. If the pump still runs all night, the sensor or wiring is faulty, not the strategy.