A/C charging question?

There is no air in the AC system. You're seeing vaporized refrigerant, which is what should be on the low side. If there was any liquid there, the system would be severely over-charged. You don't want liquid anywhere near the compressor as that would cause it to lock up at a minimum, and could cause damage to the valves or pistons. Home refrigerators can only tolerate a couple tenths of an ounce of over-charge, but those are standing still. Automotive systems can handle a lot of over-charge before there's a chance of liquid sloshing into the compressor, but there's another problem. Refrigerant gets very cold in the spot where it changes from a liquid to a vapor. You want that to be right in the middle of the evaporator. If some leaks out, but there's still some liquid in the evaporator, the system will still cool the incoming air. Same with a little over-charge. It's when so much has leaked out that the liquid in the high side changes to a vapor before the evaporator, that the cooling takes place under the hood, not in the incoming air stream, and you feel the warm air. Similarly, with a severe over-charge condition, the evaporator is full of liquid, plus some of it gets into the hose under the hood going to the compressor. Here again, the change of state to a vapor takes place in that hose, hopefully before it gets to the compressor, and the cooling takes place there, under the hood.

The problem now is there is no way to know how much refrigerant is in the system. Very experienced HVAC specialists can get an idea by watching the high and low-side gauges, but even those don't tell the whole story. Thinking back to older vehicles that used R-12, it was just a dandy coincidence that with the system at rest, the equalized pressure in the system would almost perfectly match the outside temperature in Fahrenheit degrees. If the outside air temperature was 60 degrees, for example, the pressures in the system were close to 60 psi. That assumes there was enough refrigerant in the system that some was liquid and some was vapor.

Here's the interesting part. If you bleed some air out of a tire, the pressure goes down. You'd expect that. But in an AC system, when you bleed off some of the vapor, the pressure goes down at first, but then some of the liquid vaporizes and expands, so the pressure goes right back up to where it was. As long as there's some liquid still in the system, no matter how much vapor you bleed out, that liquid continues to vaporize, expand, and the pressures will stay right where they are. It's the fact that pressure keeps some of the refrigerant in a liquid state, and less pressure lets more of it turn to vapor, that makes it work.

The point of this sad story is there are only two ways to know when the system is properly charged. For most cars, that is to recover all that is currently in the system, pump the system into a vacuum, then pump in the exact measured amount called for. The only other way is on older Chrysler products that used a sight glass on top of the receiver / drier. You add refrigerant just until the bubbles flowing through the glass disappear, and no more.

For the benefit of Ford owners researching this topic, some models as far back as the 1980s had a similar sight glass, but they must never be used. Due to the design of the system, there will still be a lot of bubbles in the sight glass when the system is fully charged and when it is over charged. As with all other brands, you have to start from empty and add the measured amount.

To address your observation of intermittent cooling / warm air, the first thought is the system is leaking and there's no longer any liquid in the evaporator, but not all the clues you observed agree with that. Another thing to consider is if the system was open recently for other service, such as to replace the compressor or expansion valve, it must be pumped into a vacuum for a good 30 minutes. Humidity in the air will cause problems. That moisture, or any water for that matter, will boil at 77 degrees under vacuum and allow it to be drawn out. If that moisture remains in the system, over a long time it promotes corrosion of metal parts, but in the short term, it will freeze and block refrigerant flow when it hits the expansion valve. That's where it turns from a high-pressure liquid to a low-pressure liquid, and starts to get cold. That blockage can stop the flow for 20 minutes to an hour, then the ice melts, the flow resumes, until that water makes its way around again.

Expansion valves have also caused this problem. They can block flow, they can let too much flow, and they can work fine for weeks or months before causing a problem again. High-side pressures will be too high, and low-side pressures will be too low when the valve is stuck closed.

When this occurs again, keep the system running, then look if there's frost buildup around either hose under the hood. That will show where the liquid is changing to a vapor and where the cooling is taking place.

Remember, the refrigerant gets cold and does its thing right at the spot where it turns from a liquid to a vapor. We want that to be right where the temperature sensor is. That's why it's critical to pump in the exact measured amount. By adding a little or bleeding a little off, you know at some point the wrong amount was in there to place the liquid level right by the sensor. You may have gone from one wrong level to the other, while missing the exact perfect level.

One more thing to consider. While I did say the system will still cool with a small over or under charge, there is usually a temperature sensor right in the middle of the evaporator. There is going to be a gas-charged sensing bulb connected to the expansion valve, or an electronic temperature sensor used to cycle the compressor on and off. Whatever system is used, it is designed to maintain pressure, and therefore, temperature, in the evaporator at no less than 40 psi. AC systems do cool the air, but the real comfort comes from removing the humidity from that air. It does that by condensing it on the evaporator, then it drips down and drains onto the ground. If the evaporator temperature drops below 32 degrees, that water will freeze into a block of ice that restricts air flow. With no warm air passing through the evaporator, it gets colder faster. That ice won't melt until well after the system has not been running for at least a few hours.

The newer refrigerants probably do not track temperature with pressure like the old R-12 did. That was just a coincidence. With the newer refrigerants, the systems are still designed to prevent the evaporator from going below roughly 40 degrees so the condensate doesn't freeze. If that's done by sensing pressure, it likely won't work properly if the system has been retrofitted to use a different refrigerant. If it's regulated with a temperature sensor, the water won't freeze, but system performance could be compromised. For example, years ago when it was common to retrofit from R-12 to R-134, the newer refrigerant needed to be at roughly twice the pressure to get the same cooling power. This retrofit started out with changing the compressor, changing to hoses that could handle the higher pressures, and a lot more stuff. Over time the procedure got simplified to changing the oil in the system, adding hose adapters, and putting in the R-134. Air duct temperatures usually weren't as low as before, but who wants to sit in a 50 degree car? Even the normal 60 degrees is too cold, but the humidity was removed. That's why the retrofits seemed to work so well.
The point is, in regard to controlling the flow of refrigerant, the two of them behaved very similarly. That may not be the case with some of the aftermarket refrigerants sold as "green" products.