Automotive Air Conditioning Information Forum

Finally got around to finish building up the A/C system on this RX7, but I'm finding after evacuating the system, it won't hold a vacuum over time. So it's likely something is leaking. Quick rehash of the restoration project and summary of where I'm at now:

- When I got the car, the only A/C equipment it had was the OEM Denso evaporator core, and its HVAC controls/wiring was all fully intact and working.
- Removed the evaporator core, and with the expansion valve removed, flushed it out with an A/C solvent. I replaced the expansion valve with a OEM Denso part and new O-rings, and reinstalled the evaporator unit under the dash.
- Picked up a very good condition used Denso compressor for the car; hand turning it it appears healthy, you can feel the compression and suction on the appropriate ports as you turn it. Rotates smoothly with no odd binding or funky noises. FWIW, I got the compressor from another rotary guy who said it was fully functional when removed it from his car (that car was converted to track duty)
- Picked up a new aftermarket generic parallel flow condenser, 24"x12" size which is the largest one I can fit. It was delivered capped and filled with a pressurized nitrogen charge; upon cracking open the 1st cap, I can hear the gas escape, which means the condenser itself is free of leaks
- Picked up a generic aftermarket dryer & binary switch that fits the dryer. Dryer fits in the OEM location/bracket. It too was nitrogen filled, and I could hear the hiss when I cracked open the 1st port
- Due to the turbo I'm running, the OEM A/C hard lines & hoses would not work, so I had to fabricate my own lines and route them slightly differently from the OEM installation. I used all reduced barrier hose (no hard lines) and crimped on bead lock fittings for the hose ends. New proper size green/HBNR O-rings were used on all fittings

Which brings me to today, after closing up the system I pulled a vacuum on it to evacuate and see if the vacuum would hold. The vacuum would only go down to 28 inches Hg. Was expecting 29~30 inHg, but I thought that might be due to cheap Horror Freight vacuum pump i was using. FWIW, same pump on prior jobs would pull a vacuum down to about 30 in-Hg. Anyway, I shut off the pump when I saw it wasn't going to pull anymore vacuum, closed the manifold valves and came back to check the gauges about 10 minutes later. At that point, the vacuum dropped to about 15 in-Hg, and by about 30 minutes later, it was down to zero.

Obviously something is leaking, the question is what? I systematically went over every single connection in the system, opening it up, checking the O-rings, closing it back up & torquing the fitting to spec one by one and repeating the evacuate & vacuum hold drill. None of the connections had any impact on the results, it still lost vacuum within the same period of time.

One concern I have is my manifold gauge set isn't a high quality tool - it's another Horror Freight item, and a tiny leak in any of those fittings would have me chasing my tail to run a conclusive test. Plan A is to see if I can borrow/rent a better manifold gauge set and see if I get the same results, to confirm that the leak is in the car and NOT the manifold gauge set.

Assuming the 2nd gauge set confirms my car's A/C plumbing is indeed leaking, what would be the most efficient way to isolate the leak so it can be fixed?

Thanks!

I think your plan is a good one. Try another manifold gauge set, pull a vacuum, and see whether it holds. If not, I would suspect the front compressor shaft seal since you've double-checked just about everything else.

Place a shower cap around the front of the compressor, charge the system with a few ounces of refrigerant, and wait a little while. Then, probe under the shower cap with an electronic sniffer and see whether it detects anything. If the used compressor was sitting for some time without being run and the front seal internally lubricated, it could have developed a leak.

JohnHere wrote:I think your plan is a good one. Try another manifold gauge set, pull a vacuum, and see whether it holds. If not, I would suspect the front compressor shaft seal since you've double-checked just about everything else.

Place a shower cap around the front of the compressor, charge the system with a few ounces of refrigerant, and wait a little while. Then, probe under the shower cap with an electronic sniffer and see whether it detects anything. If the used compressor was sitting for some time without being run and the front seal internally lubricated, it could have developed a leak.

Agree the compressor is the most likely culprit, though I still haven't pulled out the evaporator core to see if the new O-rings installed with the new expansion valve are an issue. Don't think they are bad though, because it's a simple block type expansion valve with 4 o-rings that bolts up between the core manifold and a piping assembly - the design is pretty much idiot proof to assemble & install incorrectly on a bench. That's a great tip for testing the compressor, though I don't have an electronic gas sniffer yet. Hopefully one can be rented/borrowed from one of the local auto parts retailers (Advance, Autozone, etc.)

Al9 wrote:Orings have to be lubed (with a non-hygroscopic AC lube/AC o-ring lube, such as the good ol' mineral, and Ester, unfortunately, is hygroscopic).

I purchased the proper O-ring lube along with all the fittings & O-rings and used it, so that's not an issue here.

Another update. My local Advance Auto parts didn't have a manifold gauge set available yesterday, so I couldn't use it verify the readings I got from my questionable Harbor Freight gauges. But I had a thought - can I self-test my Harbor Freight gauge set to see if it holds vacuum on its own, without being connected to the car?

The manifold gauge set in question is this one: https://www.harborfreight.com/ac-r134a- ... 60806.html

What I did to "self test" it was connect the yellow hose to my vacuum pump, and connect the red/blue hoses to their respective quick connect thumbscrew valves as you normally would, but not connect them to the car's service fittings. Keep both thumbscrew service valves on the gauge set in their CLOSED positions, and OPEN both of the manifold gauge valves. Then start the vacuum pump, and draw a vacuum as far as it would go - about 29~30 in-Hg per the low side gauge; high side lacks #'s for vacuum, but the needle goes below the zero mark, and comes close to resting on the needle stop. Turn off the vacuum pump, and let it sit that way overnight. I checked the gauges again in the morning, and both were showing the same vacuum they were at when I left them the day before.

So is that a good test to verify that the leak I'm chasing lies in my car's AC system, giving me confidence that the gauge set is OK?

Sounds like your gauge set holds vacuum.

Well I managed to get an electronic leak detector and found my leak yesterday - it was the brazed joint I made between the male O-ring fitting and the pictured compressor fitting pipe, discharge side. Heck, once I added some refrigerant, I didn't need the sniffer to find the leak - the compressor lube with UV dye made the leak pretty obvious.
I salvaged that pipe and the suction side fitting from a used OEM hose set. After removing the OEM beadlock ferrules, my original plan was to just stick a #8 reduced barrier hose over the barb end of that discharge pipe and crimp on a new ferrule, but I couldn't get the hoses & ferrules to fit over the barbs on either the suction or discharge pipes - apparently Mazda used a slightly larger ID hose than the current #8 & #10 reduced barrier hose standards? Same hose I purchased fit easily with no struggle on the new #8 & #10 fittings I got. Gave up on trying to force the hoses to fit, and went to plan B - brazing aluminum male O-ring fittings on the end of each aluminum pipe. While my brazing work resulted in mechanically solid connections on both parts, only the suction fitting tested gas-tight.

I'm a newbie to brazing AL, are there any suggestions on how to get a gas tight seal when I do-over the discharge side? So far I un-brazed the discharge pipe from the male O-ring fitting. Pipe end cleaned up well, and should be re-usable. Worst case, I still have enough length on it where I can cut off 1/4" or so of the brazed end if need be. The male O-ring fitting got too trashed upon removal to reuse, so I ordered a new one.

pete_89t2 wrote:Well I managed to get an electronic leak detector and found my leak yesterday - it was the brazed joint I made between the male O-ring fitting and the pictured compressor fitting pipe, discharge side. Heck, once I added some refrigerant, I didn't need the sniffer to find the leak - the compressor lube with UV dye made the leak pretty obvious.

Glad you found the source of the leak.

pete_89t2 wrote:I'm a newbie to brazing AL, are there any suggestions on how to get a gas tight seal when I do-over the discharge side?

Techniques for a gas-tight seal is to ensure that you heat the parts and not the brazing rod so that the rod material flows smoothly into the heated joint. Typically, the parts need to be heated to around 700 degrees F, which depends on the melting point of the brazing rod you're using. Check the packaging for that info. Be sure to clean the parts well with a small wire brush with brass bristles. And keep the torch moving. Otherwise, you risk melting the parts themselves rather than just the rod. MAP gas and air/acetylene torches will heat the parts quicker than propane, so extra care is needed when using the former.

JohnHere wrote:Techniques for a gas-tight seal is to ensure that you heat the parts and not the brazing rod so that the rod material flows smoothly into the heated joint. Typically, the parts need to be heated to around 700 degrees F, which depends on the melting point of the brazing rod you're using. Check the packaging for that info. Be sure to clean the parts well with a small wire brush with brass bristles. And keep the torch moving. Otherwise, you risk melting the parts themselves rather than just the rod. MAP gas and air/acetylene torches will heat the parts quicker than propane, so extra care is needed when using the former.

Sounds like I just need a bit more practice then, as I pretty much did all of the above. The AL brazing rods I used were these guys, available locally: https://www.homedepot.com/p/Bernzomatic ... /203710179

I was using a MAP gas torch, and applying the heat to the parts until they were hot enough to melt the brazing rod so it would flow into the joint by capillary action. I only applied heat to the parts, periodically touching the part with the brazing rod to test when if it was up to melting temp yet. The packaging for these rods claim a melt point of about 725~750*F, IIRC, and that you don't need to use flux with them. Some of the reviews of these brazing rods say they work much better when used with a proper AL brazing flux, so I may try using flux next time. Any recommended fluxes for AL brazing? Seems that none of the hardware stores local to me carry fluxes for AL brazing.

Take the ass'y to a local welding shop and have them TIG weld it and be done with this.