944 A/C Condenser Leak

[dr bob]

Sorry to hear about the failure.

In larger exchanger duty, we hydro-test for leaks, but then get to deal with the moisture using heat, vacuum (for days...), and dry-nitrogen sweeps. Vacuum testing is a good screener of bigger leaks, but for effective evaluation especially on smaller systems you need to leave it for a few days and monitor an electronic gauge that reads in microns. Vacuum testing can be deceiving especially when the vacuum can draw an o-ringed connection tighter and artificially seal it.

Serious stuff gets helium and a sonic detector under pressure. For some condensing systems, a helium sensor goes at the vacuum pump exhaust, and you puff a little helium around the vessel, mostly connections and fittings, listen for the sensor to sing. Tape a wireless mike or a radio to the sensor so I can listen as we search even in high noise areas like a running power plant.

[ROB III]

Well, I said time will tell -- and time has now told. The eBay condenser leaks. Within a week, the refrigerant was too low for the a/c to kick on. My Harbor Freight sniff tester went crazy near one corner of the condenser, so that seemed to be the most likely leak (that, and the 'last thing you touched' theorem). I took it to a shop to evacuate the remaining r134a. While on the shop's machine, they pulled a heavy vacuum on the system and put it in test mode. His machine said the same as my HF gauges -- it passed with flying colors -- no loss of vacuum detected. I guess it needs higher pressures to leak (aka don't rule out leaks based on a vacuum test). My 30-day return window is closed, of course, but I may still see if the seller will exchange it, since it seems to be a production flaw. Otherwise, I may just by a new factory condenser and give up on the parallel flow.

AARRRRGGGGHHHHH!

[Tom]

Arggg indeed, but there was actually a sense of relief when I finally found the leak. I spent a lot of time (!) looking for leaks within the fins, since that's where the sniff tester seemed most active. I was starting to worry the leak was somewhere else, so was pretty happy when I finally found it.

On the plus side, the eBay seller was very gracious and immediately offered to replace it after I sent him that video. I figure I'll try one more and if it doesn't work out, I'll get a new factory (non parallel flow) version.

dr bob -- I did try to (ahem) 'hydro-test' at first by filling it up with a/c liquid flush and then putting a little pressure on it to see if any came squirting out. It didn't go well. But on the plus side, all areas within a 3 foot radius of my garage sink are much cleaner now.

[Pauerman]

Tom,
Did you end up using the factory expansion valve with the parallel flow condenser?

[Tom]

Tom,
Did you end up using the factory expansion valve with the parallel flow condenser?

I did. Never thought much about that. Am I missing something?

[dr bob]

Tom,
Did you end up using the factory expansion valve with the parallel flow condenser?

I have the same question as Tom, especially since the expansion valve is remote from the condenser. The drier hard pipe fit-up might be my concern. The thermal expansion valve operation will be the same, as the change in sensible heat of the liquid with a more efficient condenser translates to being able to operate at a lower charge level for lower evaporator pressure while still maintaining liquid at that expansion valve. To the TXV, it just looks like a cooler day.


The max-performance goal is to have a charge level just high enough to maintain liquid state at the expansion valve on the hottest day, and not much more. Less is better when it comes to the AC system charge and pressure levels, as a lower condensing pressure means lower pressure and therefore colder evaporation temps. More heat is absorbed to boil the liquid. This is also a good time to remind folks that any partial pressure of non-condensing vapor (read: air) in the system will seriously mess with system performance. The air takes up space and capacity in the compressor and condenser. Most importantly, since it never condenses to liquid, it offers heat instead of cooling since there's no heat used to flash it to vapor. It fools the expansion valve and raises the suction-side pressure without doing any real cooling.

TL;DR, and maybe TMI --

I converted the 928 to use R-134a very soon after it came to me almost 30 years ago. It had too many leaks so it was all apart for new hoses and seals anyway. There's a slew of interesting math and measurement needed to find the 'perfect' charge amount in advance. But since the drier has a sight glass, I could sneak up on that perfect charge level slowly, watching system pressures with electronic gauges and a few thermocouples, plus more for ambient and cabin vent temps. With the freeze switch disabled, cabin vent temps are in the mid teens (Fahrenheit) cruising in mid-90's ambient temps. The freeze switch helps avoid evaporator icing and offers a little frozen-fingers relief along the way. The slight under-charge also protects against a particular 'bubble' in the pressure/temperature relationship that was happening around 95º ambient temps and sitting in stop-and-go L.A. traffic. The 928 has two hefty electric fans, but even with those the condenser is challenged in that extreme but alas too common condition back then.

If you have a chance to look at a Mollier diagram (a Mollier diagram charts the temperature, pressure, and heat carrying relationships for not-perfect gasses/vapors) for R-134a, you can see that bubble starting around 140º. That's the 'phenomena' that causes overpressure events in poorly converted R12 systems, especially those have been (over-) charged to "no visible sight-glass bubbles plus some more just in case" levels with R-134a. That's a very common habit that R12 systems generally tolerate, but offers no forgiveness in R-134a systems. There's a reason why modern R-134a cars no longer include sight glasses.

[Pauerman]

Tom,
Did you end up using the factory expansion valve with the parallel flow condenser?

I did. Never thought much about that. Am I missing something?

I was asking because since you did the Griffiths parallel condenser upgrade, I was wondering if they required using a different / newer expansion valve that would be specific to R134a

[cda951]

@Tom , thanks for sharing your experiences---I read the first post about installing the eBay parallel-flow condenser, was just about to pull the trigger on it for a customer's '89 951, then I scrolled down and saw the post about the leak that occurred the next week! His original condenser in poor shape, but I guess a new factory replacement part will be used. We've converted 3-4 customer 951s to R134a and it's ended up working pretty well. A quality parallel-flow condenser would be a great addition, though.

@dr bob, that's a great explanation, and it tracks with my experience with R134A conversions on customer Porsches----the ones with half-assed conversions often had higher-than-ideal high side pressures to maintain "normal" low-side pressures. Re-doing the entire system with more efficient heat exchangers (which often means no less than THREE separate condensers in an air-cooled 911) makes a world of difference in terms of system performance and reasonable system pressures. I miss our old-school dual R12/R134a refrigerant machine, it allowed one to adjust the charge level on the fly with the engine/compressor running to perfectly dial in the system. Both of our modern Robinair R134a and R1234yf are more automated, and require a complete evacuation and recharge in order to adjust charge level.

I have also had the "pleasure" of tearing into the dash of few 928s to replace all the leaking Mercedes vacuum pods and make the climate control system work properly. The 928, more than any other car I've experienced, absolutely needs the HVAC system set to "recirculate" mode in order for the A/C to blow cold air, due to the location of the HVAC outside air intake and its proximity to heater hoses and engine bay heat.

[Tom]

@Tom , thanks for sharing your experiences---I read the first post about installing the eBay parallel-flow condenser, was just about to pull the trigger on it for a customer's '89 951, then I scrolled down and saw the post about the leak that occurred the next week! His original condenser in poor shape, but I guess a new factory replacement part will be used. We've converted 3-4 customer 951s to R134a and it's ended up working pretty well. A quality parallel-flow condenser would be a great addition, though.

@dr bob, that's a great explanation, and it tracks with my experience with R134A conversions on customer Porsches----the ones with half-assed conversions often had higher-than-ideal high side pressures to maintain "normal" low-side pressures. Re-doing the entire system with more efficient heat exchangers (which often means no less than THREE separate condensers in an air-cooled 911) makes a world of difference in terms of system performance and reasonable system pressures. I miss our old-school dual R12/R134a refrigerant machine, it allowed one to adjust the charge level on the fly with the engine/compressor running to perfectly dial in the system. Both of our modern Robinair R134a and R1234yf are more automated, and require a complete evacuation and recharge in order to adjust charge level.

I have also had the "pleasure" of tearing into the dash of few 928s to replace all the leaking Mercedes vacuum pods and make the climate control system work properly. The 928, more than any other car I've experienced, absolutely needs the HVAC system set to "recirculate" mode in order for the A/C to blow cold air, due to the location of the HVAC outside air intake and its proximity to heater hoses and engine bay heat.

In defense of the eBay seller, he's shipped me a new one the minute he saw the video, no questions asked and even though long past the return period. He shipped out a replacement same day, trusting I'd send he leaky one back (on his shipping label). So, while the first part was a disappointment, the seller was very quick to stand behind his product. My old Mercedes has been hogging my lift (and garage time) last week, but I'm hoping to try the replacement condenser this week and will report back. If it is air tight, I'd go back to recommending it. If it leaks, then I'll get a new factory unit and give up on parallel flow for now. Fingers crossed.

[dr bob]

Hi Chris --

I've had the distinct pleasure of replacing all those vacuum actuators in my 928 too. All but one got more modern silicone diaphragms, and twenty-lots years later they are still doing well. My only 'episode' with the system was a leak in the diaphragm of a heater control valve. The system is almost completely reliant of vacuum for operation, and even with the HCV tied closed, the small vacuum loss was causing some really odd symptoms. While the whole dash was out for that work, I resealed the flaps and doors in the air box with some slightly better materials. Split silicone hose for flap and door seals, modern poly foam tape for the rest. Interesting having to do it all, but certainly justified with the excellent system performance I've enjoyed since.

Following on to your mention of needing the fresh-air door closed for decent AC -- I can hear that door opening and closing as the cabin temps cycle around setpoint. There are certainly times I wish the compressor would cycle, of the temp blend door had a more linear effect on actual vent temp. Or there was an automatic fan speed control tied in. Meanwhile, the system uses a binary on-off for the fresh air door It's, um, interesting.

Tom, the same Behr design team did the climate control in upper-trim 1980's MB cars. The vacuum actuators (all but that one) are MB parts.

Since migrating north from California a bit over ten years ago now, the AC in the 928 sees much less use. I cut the cable tie on the HCV within a few months or arriving here. 40-50º temp swings between morning and night demanded a little heat in the morning, cooling in the afternoon. Prior, it was a black car with untinted greenhouse windows in Los Angeles. Serious AC is mandatory. The difference between my 911's (mid 60's to mid 80's versions) AC vs. the 928 is massive. Porsche recognized that they'd have to make the system a super performer if they wanted to compete in the MB market. So they went to the source of MB cool.