Cooling That Fades After Twenty Minutes
A 2017 diesel estate with an R134a system labelled at 590 g comes in with a complaint that the air conditioning blows cold for the first 15–20 minutes of a motorway run and then fades to ambient, and is cold again the next morning. Ambient is 28 °C with high humidity. With the engine off, static pressure reads 5.1 bar. With the engine at 1500 rpm, blower on maximum, recirculation selected: low side 2.2 bar and high side 12.0 bar at the start, vent temperature 7 °C. After 12 minutes the low side has fallen to 0.7 bar, the high side to 9.5 bar, the vent temperature is 16 °C, and the liquid line just downstream of the expansion valve is frosting and then clearing. The compressor clutch stays engaged, the condenser fan runs continuously, the condenser is clean, and there are no fault codes. The system has not been opened in the four years the owner has had the car and there is no record of an earlier repair.
1. Interpret the readings and explain which faults the evidence supports and which it rules out.
Read the worked answer
What the numbers say at the start: a low side of 2.2 bar corresponds to an evaporator at about −1 °C and the vent at 7 °C, so the system is initially working reasonably; a high side of 12 bar at 28 °C ambient implies a condensing temperature near 50 °C, a little high but not alarming in still air.
What changes: both pressures fall, low side to 0.7 bar, while the vent temperature climbs to 16 °C. Falling low side with reducing capacity is the signature of a restriction upstream of the evaporator - the compressor is pulling against a blockage, so it moves less mass and the high side drops too. Frosting at the expansion valve outlet that clears is the classic sign of moisture freezing at the valve seat: ice forms, flow stops, the ice melts when the flow stops and the local temperature rises, and the cycle repeats.
Ruled out:
- Undercharge - would show permanently low pressures and low subcooling, not a repeatable cycle that resets overnight.
- Compressor wear - would show low high side with a high low side, not a collapsing low side.
- Condenser airflow - fan runs, condenser is clean, and the high side falls rather than rises.
- Expansion valve stuck closed - possible, but it would not restart by itself after standing.
- Electrical inhibition - the clutch stays engaged and there are no codes.
The scenario repeats after a cool-down period, which supports moisture or ice rather than a mechanical failure.
2. Why is adding refrigerant the wrong response, and where did the moisture come from?
Read the worked answer
Why not top up Adding gas does not remove water. It raises the charge above specification, which on a fixed-orifice or TXV system reduces subcooling headroom, raises high-side pressure and can cause liquid flood-back to the compressor and oil dilution. It also masks the real fault long enough for the compressor to be damaged, and the customer then faces a much larger bill. A system that has held its charge for four years with no leak record does not need refrigerant; it needs the moisture removed.
Where the water comes from
- Air left in the system because it was never evacuated deeply enough, or evacuated without a micron gauge on a humid day.
- A receiver-drier that has been left open to atmosphere, or one that has exceeded its service life and is saturated. Desiccant has a finite capacity; once it is loaded, water passes through to the expansion valve.
- Moisture introduced during charging from a contaminated drum or a cylinder left uncapped.
- A leak that has been pulling in moist air under vacuum, even though the customer has no repair history.
The judgement point: the low side reaching 0.7 bar with an ice signature at the valve is a moisture diagnosis. The correct response is to open the system, replace the drier (and inspect or replace the expansion valve), and re-evacuate properly. Recovering the charge first is mandatory - you cannot diagnose moisture by adding to it, and you cannot recover cleanly once you have overcharged the system.
3. Set out the repair and the verification you would perform before returning the car.
Read the worked answer
Repair sequence
- Recover the charge, weigh it and compare with the 590 g label - record the difference, because a genuine shortfall changes the diagnosis to include a leak search.
- Remove the receiver-drier, cap the open ports immediately, and replace the expansion valve if there is any sign of corrosion or debris in the screen.
- Renew all disturbed O-rings, lubricate with the specified oil, torque to specification.
- Blow through with dry nitrogen to confirm no restriction and to carry out a pressure-decay leak test at a sensible pressure well below component limits.
- Triple evacuation with dry nitrogen breaks, using a micron gauge at the system, not at the pump. Hold below 500 microns for at least 10 minutes with the pump isolated; a rise that plateaus in the low thousands indicates moisture, not a leak.
- Charge by weight to 590 g using a calibrated scale, with the correct oil quantity topped up for the parts replaced.
Verification
- Run at 1500 rpm, blower high, recirculation on, and log vent temperature, low and high side pressures, and ambient at two-minute intervals for 20 minutes. The low side should stay stable, not drift down, and vent temperature should hold; that repeated-run stability is the real proof the ice is gone.
- Check superheat at the evaporator outlet and subcooling at the condenser outlet against specification.
- Check the compressor clutch cycling and the condenser fan operation.
- Road test and then let the car idle for 15 minutes to reproduce the customer's conditions as closely as possible.
- Record pressures, vent temperature, charge weight, parts fitted and vacuum figures on the job card, and explain to the customer that the cause was moisture in the system, not a shortage of gas.