There isn't one fix for an F22 fault on a Vaillant boiler. I know that's an unsatisfying way to open a troubleshooting article, but the alternative — a single five-step guide that supposedly works for everyone — is how people end up replacing a perfectly healthy pressure sensor, or topping up a system twice a week for two years and then wondering why the heat exchanger is full of sludge.

I work as a quality and compliance manager at a heating contractor. I review commissioning reports, service sheets and warranty claims before they go anywhere near a customer — roughly 180 to 200 jobs a year. In our 2024 review, we pushed back on about 9% of first-pass installs, and roughly half of those were pressure-related. Almost every one came down to the same thing: someone treated the code as the problem instead of the symptom.

When I first started reviewing commissioning sheets (this was back in 2019), I assumed F22 meant the pressure sensor had failed. Six years of warranty claims later, my rough sense is that the sensor itself is the culprit in maybe a third of cases. The rest are the system telling you something you'd rather not hear.

What F22 actually means

F22 is the low water pressure fault on Vaillant's gas condensing range — ecoTEC and similar. Vaillant's installation and maintenance instructions describe it as heating circuit pressure below the permitted minimum, with the boiler locking out until pressure is restored and the fault reset. Check your own model's manual for the exact threshold; it usually sits somewhere near the 0.5 bar mark, and that number changes how you read the gauge.

What follows is four scenarios. Find yours before you touch anything.

  • Scenario 1: Gauge confirms low pressure
  • Scenario 2: Gauge reads normal, F22 keeps coming back
  • Scenario 3: Pressure drops repeatedly
  • Scenario 4: It isn't that kind of system at all

Scenario 1: The gauge agrees — pressure is genuinely low

This is the straightforward one. Cold system, gauge sitting below about 1.0 bar, F22 on the display. The heating circuit has lost pressure, usually through a small weep, a recently bled radiator, or a system that's been sitting idle over summer.

Top up via the filling loop to somewhere between 1.0 and 1.5 bar cold. Not 2.5 bar because it felt safer — over-pressurising puts the expansion vessel and the pressure relief valve under unnecessary load, and the PRV may not reseat properly afterwards (which turns a five-minute job into a call-out).

Then bleed the radiators. Order matters here: if you top up first and bleed afterwards, you'll be topping up again. Cold system, heating off. Radiator key on the bleed valve at the top of the radiator, cloth underneath, quarter turn until you hear the hiss, then close it the moment you get a steady trickle of water rather than sputtering air.

Work through the system furthest from the boiler first and come back towards it. Honestly, the exact sequence matters less than people claim — but bleeding cold, in one pass, and re-checking pressure afterwards means you only do it once.

After 20 to 30 minutes, check the gauge again. If it's dropped, you've bled more air out than you put water in, and you need another top-up.

The trigger for Scenario 3: if you're doing this more than once or twice a year, that's not a maintenance routine. That's a leak.

Scenario 2: The gauge reads normal but F22 keeps returning

Gauge in range (1.0 to 1.5 bar cold), and the fault either won't clear or comes back within hours. Now you're looking at the signal side, not the water side.

The Vaillant water pressure sensor sits on the hydraulic block. In my experience the failure modes, roughly in order of frequency:

  • Debris or magnetite blocking the sensing port — the sensor reads correctly, but it's reading a blocked tapping
  • Connector or wiring fault at the sensor plug
  • Sensor failure itself
  • On older appliances, a blocked sensing line

Here's the bit people get backwards. The assumption is that a pressure fault means the pressure sensor has failed. More often, the sensor is the one component in the chain doing its job properly — it's reporting accurately, and the reading it's reporting is correct. The question to ask isn't why is the sensor lying, it's why is that reading what it is.

I had a batch of warranty claims in 2023 where three of eleven boilers on the same estate had the pressure sensor replaced within eighteen months. All three were the same brand of aftermarket sensor, fitted because the OEM part was roughly three times the price. Two failed again inside a year. The savings were about £90 a unit; the repeat visits, drained systems and customer goodwill cost us closer to £600 each. That's not an argument against cheap parts in the abstract. It's an argument against treating unit price as the whole cost.

Any work inside the boiler casing — sensor replacement included — is a Gas Safe registered engineer's job, not a DIY one. The pressure gauge and filling loop are fair game; the hydraulics behind the case aren't.

Scenario 3: Pressure keeps dropping — and topping up is the wrong answer

This is the scenario where the instinctive fix does real damage.

Repeated pressure loss usually means one of three things: a leak somewhere on the system, an expansion vessel that's lost its charge, or a pressure relief valve that's discharging and not reseating.

The trap is that topping up works. It works for a week, or a month, and it's free, so it becomes the routine. What it's actually doing is introducing fresh, oxygenated water into a sealed system. Oxygen plus steel equals corrosion, corrosion produces magnetite, magnetite settles in the lowest points and the slowest flow paths — which is to say, in the heat exchanger and around the pressure sensor port.

Two years of £15 top-ups and a bit of inconvenience tends to become one £1,200 to £1,800 heat exchanger or a system flush, which is a job nobody enjoys explaining to a facilities manager. I'd argue the arithmetic here is lopsided enough that it stops being a judgment call.

What to do instead: find the loss. Check the pressure relief discharge pipe for signs it's been wet, look for weeping at radiator valves and compression fittings, check for damp patches or staining under floors. If there's no visible leak and the system still loses pressure, get the expansion vessel charge checked — a recharge is a fraction of a replacement, but only if the vessel itself is still sound. If the PRV has discharged and isn't reseating, it needs replacing; they rarely come back from that.

One caveat on cost: those figures are indicative for a mid-sized domestic system as of January 2025. Verify current rates with your contractor — region and system size move them a lot.

Scenario 4: It isn't that kind of system at all

Not every heating appliance has a water pressure circuit, and not every pressure problem is a boiler problem.

F22 is specific to Vaillant's gas condensing appliances. If you're standing in front of an oil-fired appliance — a kerosene heater, for example, or a portable kerosene space heater in a workshop — there's no water pressure sensor, no F22, and none of the logic above applies. A kerosene heater has its own fault modes entirely: fuel supply, wick condition, airflow, flue spillage. Transferring sealed-system diagnostics to it will send you looking for a gauge that doesn't exist.

The same reasoning shows up somewhere much less expected. A tyre pressure sensor light comes on, and the instinct is to blame the sensor. Almost always, the sensor is telling you the truth about a tyre that's slowly losing air — a nail, a bead leak, a valve. The monitor is the messenger. I've lost count of how many times I've seen someone replace a component because they didn't like the message it delivered.

The question everyone asks is how do I clear the code. The question worth asking is what changed to make that reading appear.

Working out which scenario you're in

Four checks, in this order, before you call anyone or buy anything:

  1. Read the gauge cold. Boiler off, system at ambient temperature. A hot reading is meaningless for comparison.
  2. Top up if needed, then watch it for 24 hours. Holds steady? Scenario 1. Falls again? Scenario 3.
  3. Look for evidence of loss. PRV discharge pipe, valve weeps, staining, damp patches, corrosion crusting around fittings.
  4. If the gauge is fine and the fault persists, you're in Scenario 2 — sensor or signal, and that's an engineer's job.

One more thing, mostly for anyone managing a portfolio rather than a single home: log the readings. Not because it's clever, but because pressure history is the single most useful piece of information you can hand an engineer. A system that's dropped 0.2 bar over six months is a different conversation from one that's dropped 0.8 bar in a week.

My sample here is roughly 200 domestic and light-commercial jobs a year in one region. If you're working on large commercial plant or district heating, the ratios probably shift. But the underlying point holds: the cheapest response is usually the one that costs the most six months later. Read the gauge before you reach for the parts catalogue.