Micron gauge: turning an aircon evacuation into evidence

A micron gauge measures what is left inside a circuit, not how hard a pump is working. It reads absolute pressure far below atmospheric, in the region where almost everything has already been drawn out of the pipework. That region is where an evacuation either succeeds or does not, and an ordinary manifold set cannot resolve it.

The needle on a manifold gauge is built for the working range of a running system. Its scale covers the pressures refrigerant reaches in normal operation, and the whole of a deep vacuum sits inside the last short arc before the stop. Once the needle reaches that stop it has nothing further to report. A circuit still holding a great deal of water and a properly dried one look identical on it.

The gap in resolution is not a small one. Everything a manifold treats as a single mark at the end of its travel is the entire span a micron gauge exists to divide up. A technician watching the needle sees the pull finish long before it has finished. A technician watching microns sees the slowest and most decisive part of the job.

So the instrument, and not the pump, is what distinguishes one evacuation from another. Pumps are ordinary equipment and most vans carry one. A gauge that resolves the deep range is a deliberate purchase, and it commits whoever bought it to reading a figure they may not like. That second part is the harder half.

The gauge stays on because the pull is the easy half. A pump left running on a small circuit will reach a low figure sooner or later, and that figure describes the pump at least as much as the pipework. What nobody knows at that moment is whether the circuit can keep it.

Shutting the pump off turns the instrument into a different test. Until then the reading reports a balance between what the pump removes and what the pipework gives up. Close the valve and the pump leaves the equation entirely. Whatever the number does next belongs to the circuit alone.

A dry, sealed circuit sits still. Nothing inside is left to evaporate and nothing outside is getting in, so the figure stays where it was left. That stillness is the actual result of the whole exercise. The low number before it was only the setup.

This is why an answer built on the figure alone is incomplete. Two systems can arrive at the same reading in completely different conditions. One of them keeps it. The other starts climbing the moment the pump is out of the way, and it was climbing underneath the pump's work the whole time.

A rise after the pump is shut off has two possible sources, and each points somewhere different. Either something inside is still turning into vapour, or something outside is finding its way in. The first is moisture in the refrigerant circuit that the pull did not finish removing. The second is a leak.

The shape of the rise separates them. Water left clinging inside the copper has a limit. It evaporates into the remaining space, lifts the pressure, then stops, because at any given temperature that space will hold only so much vapour. A wet circuit therefore climbs, slows, and levels off at a plateau it will not pass.

A leak has no such ceiling. The atmosphere outside is effectively unlimited, so air keeps arriving at a steady rate and the reading keeps going. It does not flatten and it does not settle. Left long enough it returns all the way to where it started. Watching the movement over a standing period, instead of glancing at the display once, is what makes the two cases separable.

Manufacturers publish the interpretation as bands rather than as a single pass mark. Daikin's installation and service reference for its R-32 condensing units gives three of them. A rise that steadies below one thousand microns reads as leak free. A rise that holds between one and two thousand points at leftover air or moisture, or at a small leak. A rise past two thousand is called a leak outright, to be found and repaired. Other makers set their bands elsewhere, so the numbers on any given job come from that system's own manual.

A leak found here costs a joint. The pipework is still open to the technician, no refrigerant has been bought or released, and no ceiling has been closed over anything. Remaking the joint and pulling down again is an inconvenience inside the same visit. Nothing already finished has to be undone.

The same leak found long after handover costs a diagnosis first. By then it presents as a room that takes longer to get cold, and that description fits several unrelated faults. Someone has to rule those out before anybody starts hunting for escaping gas. Leak detection methods all have limits, and a slow loss from concealed pipework is the case they handle worst.

What comes after the diagnosis costs more again. The lost gas has to be replaced, the joint has to be reached, and reaching it can mean opening finished carpentry. If the system ran short of charge for any stretch of time, the compressor has been working in conditions it was not built for, and that wear does not reverse.

The gauge prevents none of that by itself. Its contribution is to move the discovery forward to the one moment when a leak is still cheap to put right. That is the entire economic argument for it, and it is why the standing reading counts for more than any figure the pump reached.

Ask for two photographs of the gauge face. The first belongs to the instant isolation begins, the second to the close of the watching period. Each needs a clock in frame or a timestamp on the file. Those two images and their times are the entire record, and the phone in a technician's pocket already produces them.

A picture beats a figure recited afterwards, and not because anybody is suspected of inventing numbers. A number recalled after the fact has travelled through somebody's memory on the way. An image of the display is the reading itself. It also shows the instrument, which quietly settles whether one was present.

Put the request in before a quote is accepted. A firm that already reads a gauge this way hears it as routine, since nothing new is being added to the job. A firm that does not has to change what its technicians do on site, and the difference between those two answers is easy to hear.

Nothing technical is required to make the request. It is two images and the times printed on them. Those photographs are one entry on a longer commissioning record, and what else belongs on that sheet is set out separately.

• •The gauge photographed as isolation begins, with a clock or timestamp in frame
• •The same gauge photographed again once the watching period closes
• •Whether the reading moved between the two, and by how much
• •Which manual's bands that movement was judged against
• •Whether the circuit was pulled down afresh after any joint was redone