Why does my solar thermal system keep losing pressure?

What pressure should a solar loop actually hold?

Work it out before deciding anything is wrong. The standard rule sets the cold fill pressure from the height of the system:

Fill pressure = 1.0 bar at the highest point + 0.1 bar per metre of static height + about 0.1 bar of venting reserve.

For a house with 8 metres between the solar pump station and the top of the collectors, that is roughly 1.9–2.0 bar cold. The expansion vessel pre-charge is then set about 0.3 bar below the operating pressure — around 1.5 bar in that example — and the 6 bar safety valve stays untouched, with the design end pressure kept at or below 5.5 bar.

Two normal behaviours are often mistaken for faults. The gauge moves with temperature: a swing of 0.3–0.5 bar between a cold morning and a hot afternoon is expected. And a new system loses 0.1–0.3 bar over its first weeks as dissolved gas is removed by the air separator. That is commissioning, not a leak.

What actually causes the pressure to fall, in order of likelihood?

1. Expansion vessel failure. The diaphragm perishes — solar vessels live with fluid that reaches 100 °C-plus — or the nitrogen charge slowly leaks past a tired Schrader valve. Either way the vessel can no longer absorb the volume produced when collectors stagnate, so pressure spikes to 6 bar, the safety valve lifts, and fluid is lost. Every stagnation event repeats the cycle.
2. A weeping safety valve. Once a valve has lifted on debris or a pressure spike it often fails to reseat, dripping continuously afterwards. Look for dried glycol crystals under the discharge.
3. An automatic air vent left open on the roof. During stagnation it vents steam and, with it, fluid and frost protection.
4. Micro-leaks at joints and seals. Compression fittings that loosen through 100 K of daily thermal cycling, collector union gaskets, and pump station fittings. Glycol leaves a sticky, dust-collecting film — follow the dust.
5. A failed collector or coil. Rare, but a pinholed absorber or a corroded tank coil moves glycol into the cylinder. Test the tank water with a refractometer: any glycol reading confirms it.

How do I test the expansion vessel?

Five minutes with a tyre pressure gauge:

If the pre-charge was correct and the diaphragm intact, the vessel is not your problem — move to the safety valve and the joints.

Is the vessel big enough in the first place?

A vessel that is healthy but undersized produces exactly the same symptom, because it cannot swallow the collector content when the array boils. A solar expansion vessel has to accept the whole collector volume as vapour plus the fluid's thermal expansion plus a reserve, which is why a 5 m² domestic array typically needs 25 litres where a boiler circuit of the same volume would use 12–18 litres.

Two protective details matter as much as capacity. A pre-vessel — or simply two to three metres of uninsulated copper between the tee and the vessel — keeps steam away from the diaphragm; without it the rubber cooks. And the vessel should hang so that hot fluid does not convect into it during stagnation.

When does losing pressure become urgent?

When frost protection goes with it. Every litre of propylene glycol blown out of a safety valve is a litre that gets replaced with plain water if the system is topped up carelessly, and a diluted mixture will freeze at a temperature the system was never designed for. Repressurise with premixed fluid at the original concentration and check it with a refractometer, as set out in the glycol replacement guide.

Below roughly 0.5 bar cold, stop running the system. The pump will cavitate, the fluid boils at a lower temperature so stagnation starts earlier and harder, and air is drawn in through fittings that never leaked outward. Refill, find the cause, then run.

Good service practice is to treat any repeat pressure loss as an expansion-vessel investigation until proven otherwise — chasing joints first wastes a day on most systems.