951 AC Conversion

My dyno-sheet-or-it-didn't-happen Goal

I have had several 944s over the last 15 years – none of them had working AC.  My summer 2018 goal was to get both of my two P-cars (951 & 928) to perform as well as any modern car. This was definitely a stretch goal. I also want to use R134A so I can more easily maintain my cars.

So my metric was get outlet temps of about 32 degrees Fahrenheit on a typical Missouri summer day (88f ambient and 75% humidity). There are lots of internet stories out there of how all I needed was a drier and a can of gas. When I see these stories, I always look for how success was measured. I know that good goals always have good metrics. Without metrics I always end saying silly things like, "It'll  freeze-your-fingers-off."  So, I think of my outlet temp metric as a dyno-sheet-or-it-didn't-happen type of thing.

One other thing about my goal - I didn't reach it (yet). You’ll have to jump to the end of the story for that.

Porsche 951 on work stands
Fig.1 - Ready to begin AC maintenance

And so it begins... I bought this crimper – it is Chinese and appears to have fairly good overall quality and made consistent and good-enough crimps.  It doesn’t leak hydraulic fluid and easily reaches crimping pressures of steel fittings.  I had the tool shipped to me from China and I suspect it is made in the same factory as those marketed here in the USA by Master Cool.  I bought the hydraulic version (with a remote pump) because I thought this version would work better in the tight constraints of the engine bay and thereby avoid removing the AC hard lines from the engine bay; a task that could require removing the engine. This meant that I would need to crimp at least two fittings while situated in their normal location. 


Hydraulic fitting crimper
Fig.2 - Hopefully, this crimper can work on fittings in the engine bay.
Hard to reach AC crimps
Fig.3 - In situ crimps needed on Porsche 951.


I’m particularly interested in using the low-cost parallel flow condensers that are now on the market.  I think that is the key to getting R134A to deliver. I pulled the old condenser and right away see rust or some kind of debris inside it. I wasn’t too surprised – I’ve had a few hard lessons in the AC efforts of the past. This old condenser is contaminated and has to go. And I will have to inspect and deal with cleaning up the inside of the AC system. 


contamination found at the old condenser fitting
Fig.4 - AC system contamination found at the condenser.
Getting a better look at the AC system contamination.
Fig.5 - Getting a better look at the AC system contamination.


I ordered up this parallel flow condenser to replace the old OEM condenser – it is half the weight of the original. These condensers are available at countless AC suppliers. It actually seems to be made with pretty good quality. It arrived with caps over the fittings and the unit was holding a vacuum. 


Size comparison of original to new condenser.
Fig.6 - Size comparison of original to new condenser.

I ordered the nearest size condenser that I could find to that of the original condenser. At first I wasn’t sure this would fit but mounting this new condenser worked out well. I have the bracket patterns and hardware sources so you can copy/improve on my effort if that’s what you want. The hardest part was accepting that I had to cut out the original mounting pins from the radiator frame. Once those were gone the mounting problem issue fell by the wayside.

Preparing for mounting of new condenser.
Fig.7 - Preparing for mounting of new condenser.
Brackets and hose considerations for new condenser..
Fig.8 - Brackets and hose considerations for new condenser.

The vibration isolation mounts thread into an existing hole in the radiator mounting frame (one on each side) that has a plate nut already installed. Just be sure your radiator is centered to avoid contact/damage with the isolator's stud.

Condenser mounting isolators.
Fig.9 - Condenser mounting details.

Here is the brackets sketch, I cut these by hand. The steel was very thin perhaps .06" which is fine as  the new condenser is very light. If you use this, you might want to mark the small holes directly from your condenser.


Sketch of condenser mounting brackets
Fig.10 - Condenser mounting brackets sketch.


This parallel flow (PF) condenser rejects a fantastic amount to heat but introduces some downsides too. First, it’s passageways are so small, it can act like a filter and be easily clogged. This might explain why some say this type of condenser cannot be flushed.  And last, I’m convinced it is physically smaller in capacity and that alters the ideal amount of oil and refrigerant prescribed for this car. I'm thinking the OEM prescription will have to be adjusted downward to use this condenser.  Keep in mind too that oil is not compressible and that any more oil than is needed for the compressor’s long life takes away from the overall efficiency of the AC system. The point is that I don't want to go overboard on the amount of oil I use.

To adjust the OEM prescription of refrigerant and oil, my plan was to slowly add my R134A gas while watching my outlet vent temperatures. This method is fairly common because people are always having to guess a little about how much oil is in a system when a system is refilled.

So with the condenser mounted, I started planning the new hoses. Some of the hoses have new fittings at both ends but those that attach to the compressor are using the old compressor fittings (this becomes a real problem - more on this later). 

All are to be fitted with new "reduced outside diameter" barrier hose.  Why?  I like it; its more flexible, has some tolerance to twist into a tight space and has a much smaller bend radius.  I'm not sure but I suspect that since the hose wall is thinner that I will be able to reach higher clamping pressures too. Maybe there are downsides but I haven’t found them yet. Besides, of the seven die sets that came with my crimper, the ones for #6 hose are limited to reduced OD hose. And since both of my cars use #6 hose, so accepting this hose was a forgone conclusion. I ended up using reduced OD barrier hose for ALL of my hoses – and (so far) with no regrets.


Modeling the condenser's fittings and hose routing
Fig.11 - Modeling the condenser's fittings and hose routing.

 Note: all of my Adel clamps were positioned to use factory drilled holes. 

More modeling of the condenser's fittings and hose routing
Fig.12 - More modeling of the condenser's fittings and hose routing.


There are some need-to-know things about crimping. The basics are available on U-tube for new-hose-to-new-fitting and I can show you how I did the new-hose-to-old-fittings. My old AC compressor was junk so next, I bought this compressor from Rock Auto. It is rebuilt by Denso and I think that is important.  These compressors look so nice that I would guess they are new rather than rebuilt. Cost was about $440 and comes with a warranty.


Which AC compressor to buy
Fig.13 - New AC compressor info.

I also bought a new filter drier and a high/low pressure switch from Paragon-Products.com. My model year (87) has a dedicated high pressure switch and a low pressure switch as well. The Paragon switch is meant to replace them both. I put in the new switch and retained the old high pressure switch anyway. I figured I could jumper it if it gave me any problems. I should also mention that the 944 drier is a universal model but the one I got was slightly smaller than my original mounting clamp. I cut a small piece of old radiator hose to insert under the clamp to fix this fitment issue. 


Which AC drier info
Fig.14 - New AC drier info.

I also got 8 oz of PAG-46 from Amazon.  PAG-46 oil is required for the compressor’s warranty and is stipulated in both the Toyota Technical Service Bulletin and the Porsche Technical Bulletin for how to convert to R134A. Both documents include systems that use my compressor model but the Toyota TSB (AC002) prescribes a no flush method.  I wanted to avoid flushing, if possible, because it can be so difficult to get the last of the flush chemicals out of an installed evaporator. So anyway, that is how the Toyota TSB got onto my reading list. The Porsche procedure (TB 9501) assumes I have a working R-12 system and can run it to stir up the original mineral oil and thereby extract/drain it easily. Not true in my case. My system has been dead for years.  

More on fittings - The fittings and hose that I needed I got from Coldhose.com.  There are hundreds of supplier’s but Coldhose has a nicely organized ordering system that helped me visualize/identify what I needed. These fittings are complete fittings with integrated beadlock ferrules.  This general type is used on the 944 but perhaps not as robust as those on the original 944.  In my case, I chose  steel beadlock fittings because I wanted to use steel, rather than aluminum, and I wanted to use reduced OD barrier hose as mentioned earlier. Maybe aluminum is better, I just made a choice and went with it.  See the cold hose site and "Fittings/oring/female oring" for a selection table. 

So to start on the hoses, I needed to determine what size hose is needed. The 944 uses #6, #8 and #10 and I fell into using this AC hose lingo rather than keeping track of the actual inside diameters. Hose is cheap; er, I mean inexpensive. I bought 4 feet of each and that was plenty. More details on hose length will emerge when I later mocked-up all of the hose placements.  

I suppose you can buy the original compressor fittings and maybe the compressor fitting adapters too but I did not locate a source for those. So, I used the old fittings. For the crimps to the old compressor fittings and for the hard line fittings,  I had to cut the old ferrules off to get the old rubber separated (using a grinder and small, thin abrasive cutting blade) and then clean-up the remainder of the fitting to remove any grit that might damage my new compressor.

Removing the old crimp collar
Fig.15 - Preparing for crimp.
Removing the old crimp collar (in-situ crimp 1)
Fig.16 - Removing the old crimp collar (in-situ crimp 1)
Removing the old crimp collar (in-situ crimp 2)
Fig.22 - Removing the old crimp collar (in-situ crimp 2)

So on the compressor fittings and the two hard line fittings, I’m keeping the core of the old fitting but I have cut off it’s crimped ferrule (aka collar). That means I needed new ferrules. I got these from an EBay seller in a variety pack.

A variety pack of crimp collars needed to use with original compressor fittings
Fig.17 - A variety pack of crimp collars.


I used the hose to find the correct collar (the hose should fit within the collar with a snug fit). Then I had to work with the fitting and the collar to get the collar to just barely fit over the bead of the fitting. On a few of them I had to open the hole a little by drilling.  It would be a really good-to-have if we could get the two pieces (the original fitting's core and the new collar) to mechanically lock together. If the beadlock fittings were all the same design, that would be easy. That is not the case though, so I'l have to think about that.  For now I'm happy to just get a good crimp and make the fitting, collar and hose as one connected piece.

Adapting a crimp collar
Fig.18 - Adapting a crimp collar.

Caution: I accidentally dinged one compressor fittings in a vice and had to order a salvaged replacement from PlyhammersParts.com. The fittings are made of very soft steel. Keep that in mind when cleaning them up.

Re-using a 30 year old fitting means they have to be made as perfect as possible. The compressor warranty papers made a big deal of this, showing pics of a damaged compressors that were not covered by their warranty due to contamination. A gun cleaning kit was my tool of choice and with some elbow grease it worked pretty well but still there were small imperfections (deposits, stains, tiny pits). 

The in situ crimping part is fairly easy except that the hard line has a formed bubble and it is very hard to push and hold the hose over that bubble while it is crimped.  Before I made any real crimps I got my condenser mounted and compressor (with fitting adapters and fittings) in place and the old filter drier mounted. I also had the coolant reservoir removed so that I could see the two hidden AC hoses. This way I could mock-up the hose placement to be sure I had the right hose length.  I could also mark on the hose with witness marks so the rotational orientation would come out right. I think I avoided some errors with patience in getting a good look at how everything fit together; however, there was one hose that has some twist.  

For most of the hoses I clamped my crimper into a vice but the two hard line fittings had to be crimped while working within the engine bay.  As I said, this is what I had in mind when I bought this hose crimper. It may have been the best option but still it’s hard to do with just two hands.

First in-situ crimp.
Fig.19 - First in-situ crimp.
Bottom view of first in-situ crimp.
Fig.20 - Bottom view of first in-situ crimp.
Close-in view of in-situ crimp.
Fig.21 - Close-in view of first in-situ crimp.

I mentioned contamination earlier and so I want to get as good a look at each fitting in the system to see how far it had progressed. I was then still in a quandary about the evaporator flushing. To me it seemed like I either had to  pull the dash or not do the flushing.  So next I opened up the fittings for the expansion valve. I found it was clean. So I changed the orings and put it back together. I bought the correct insulating tape to recover the expansion valve - of course it comes in a roll that is 10 times more than I need. 

The expansion valve was super clean, so I just put it back together with new orings and oem cork tape. I also cleaned out old deteriorating foam and debris from the evap fins. I used strip calk to get the cover resealed and it was all back together in a couple of hours.

Expansion valve inspection.
Fig.22 - Expansion valve inspection.
While I'm in there....
Fig.22a - While I'm in there....

So finally, I’m putting the whole system back together.  Now I used the new filter drier which I had kept sealed up so as to safeguard the desiccant beads from moisture. Also I used new orings and carefully oiled each one using the PAG-46.  My AC books are adamant on this, I guess it helps the oring to slip into sealing position.

Next was my AC system’s first vacuum test.  That went fine.

I filled and burped the engine’s cooling system and took care of a small wiring problem associated with the 12 vdc compressor on signal.  I thought I was all ready to go for refrigerant charging but what I should have been thinking about was things I should have on-hand to troubleshoot when the charge-up doesn’t go as planned.  A laser thermometer would have been good. I should have my gauges setup with short shut-off valves for fill hoses (I had them but one of them leaked and one of my hoses had a worn-out seal so those were set aside for repairs). A review of the seals in my old hoses and maybe a plan on using the mandatory R134 quick disconnect fittings for the high and low side fill ports would have been good things to do too. All stuff that came back to bite me on both of my summer 2018 AC projects (my 951 and my 928).

I actually did expect some charge-up problems but I didn’t get far with them. With about 200 psi on the high side I suddenly heard a pop and realized I had a big leak. And naturally, per Murphy’s Law, it was at a fitting that is the most difficult to access - the in-situ crimp #2 that I made for the drier hose to evaporator hard line. It’s the little hose under the coolant reservoir.

I made a quick visual estimate of how much oil I lost in that leak and set about getting the coolant drained - again.  Once the leaking fitting was exposed, I theorized about what went wrong and spent some time thinking about how I could go forward.  My failure theories aren’t worth much but I knew this was a serious threat to my fix-the-AC-or-else project.  I could order new fittings and I could crimp up a new hose as a basic re-try or I could pull the engine, remove the hard line and have a pro-AC shop crimp it for me – an almost certain fix.

Or maybe, there was another way. There had to be because I really didn’t want to pull the engine and my faith in this budget crimper was waning. This is one of those moments when I felt I had to take time to review everything. 

First pic of Leak 1.
Fig.23 - First pic of Leak 1.
Closer look at Leak 1.
Fig.24 - Closer look at Leak 1

I actually enjoy these outside the box opportunities – too much I suppose. Many times my Dad and I have stumbled into these situations and the crazy ideas were fun to laugh about. His perspective on innovating was molded by his growing up during the depression and later in the Navy where every idea was on-the-table.  People don’t seem to think like that anymore and I wonder sometimes how the Wright brothers would handle the internet eye-rolls of today. I’m getting carried away. My shop IS kinda like a bicycle shop but I’m NOT the Wright brothers and this is just a leaky hose problem.  Anyway, I looked again at the crimp quality and along the way, it occurred to me that with the thin wall of the reduced OD hose and the relatively small outside diameter of this #6 hose that there had to be some sort of clamp that could do this.  This is what I tried - a cluster of mini t-bolt clamps. And it worked and held at pressures as high as 350 psi. It worked so well I used the same fix in the same place on my 928 (drier hose to evaporator hardline) and its still working there. 

Leak 1 - repaired.
Fig.25 - Leak 1 - repaired

So I’m ready for system fill round 2.  

Adding refrigerant oil.
Fig.25 - Adding refrigerant oil

I had a new set of Chinese gauges that I wanted to try but there were difficulties so I used my ancient Robinair setup that I inherited from my Dad.  I should know a lot more about vacuum readings but all I really know is that I’ve had success if I can get the gauge to hit the lowest measurable point when vacuuming out the system – see pic. I think that is -28 psi.  My pump run time was about an hour (using my cheapie HF vacuum pump) and once closed off, the system held overnight. Remember though, holding vacuum is not the same as holding pressure.

Note on vacuuming out the system.
Fig.26 - Note on vacuuming out the system.
Close up on vacuuming out the system.
Fig.26a - Close up on vacuuming out the system.

I started the fill using the suction port on the compressor - all went well right from the start – sucked in the can slowly, ambient was 80 to 85 and humid. There was some cycling which I don’t understand as I had the new low pressure switch jumpered.  The cycling gradually stopped over the course of about 5 minutes. I started checking the temperature at the main outlet with the blower on high but forgot to put the recirc on. Outlet temps were in the high 30s – which seemed really good. Max pressure was 180 and I stopped filling after just one can (11 oz). It was an easy place to stop and review. The condenser was definitely radiating heat - wow! The rebuilt Denso compressor is runnign smooth quiet. Even the DME auto adjust of the idle speed is working (My DME is programmed with Lindsey/Rogue's A=Tune). My pressures were 180/30. I used my new Chinese sniffer to look for leaks but found none. I was feeling good but knew there had to be more problems ahead.

Well, it only took six weeks to get to this point - my first test drive with the new AC.  I had some little things to clean-up like adel clamps for the compressor to condenser hose a permanent repair of the compressor electrical connector problem. I found a little oil on the outside of the compressor fittings (a bad sign) but I wiped that off thinking I could check it after some use. I put on the under-pans, batwing and aft aluminum pan. I got the car down off of the stands/ramps. I backed it out and the AC seemed really cold; even freeze-your-fingers-off cold.

I drove it about 30 miles that day. Ambient was 82 and humid enough to not be comfortable. Idling outlet temps were just south of 40 d Fahrenheit. That seemed acceptable.   As soon as I got rolling (40ish mph) my outlet temp came all the way down to 30f.


It’s hard to get a good pic of the tiny thermometer when moving. The outlet temp cycled a little between 30 and 38 but as soon as I turned on the re-circ it would hold at 30 as long as I was moving. Now I know its early but Dayam! This feels like a win. It didn't even occur to me that that little thermometer (one of Dad's) was over 50 years old.


First drive with unbelievable outlet temp.
Fig.27 - First drive with unbelievable outlet temp.

There is one thing though that bothers me (beyond the using only one can of R134A) and that is the question of why hasn’t my freeze switch disengaged the compressor – the compressor runs constantly. And it’s a pretty big load on horsepower too. According to my thermometer, I’m reaching 30f so perhaps my freeze switch is bad. After several days of driving the car I decided I would adjust the freeze switch and maybe get it to kick off the compressor at 35f. I tried that but no change. More time went by and I was feeling pretty good. Days passed.

The 2018 Missouri summer heat was steady and at one point I used the car to run an errand. In the process, the car was parked in the open and the interior got heat soaked. When heading back home it took a good 15 minutes for the interior to cool down. I knew then that something wasn’t quite right. It turned out my thermometer for measuring outlet temps was way off. With another thermometer, my low 30’s temps turned out to be 42f.  I wrote in my notes… "So it turns out that it’s a good thing I didn’t tell anyone about my miraculous low temp AC system."

So you might think I would jump right back in to this project and get on with tuning whatever needs done to get my outlet temps lower but by this time I was way down the road on my other car and summer was turning to fall. I also had a new milkshake problem on my 951 and that took all of my focus. I managed to get past the milkshake only find my radiater leaking. I had to get that behind me too before it got cold. So anyway, to this day my 951 is just as it was in late September of 2018. Its freeze-my-fingers-off cold but really only gets down to about 42f. Update: 11/29/19... I lost cooling in early summer 2019. Analysis - the re-used factory, compressor fittings were leaking.

944 ownership - it just gets better and better.
Fig.28 - Milkshake: 944 ownership just gets better and better..
New radiator needed in spring of 2019.
Fig.29 - New radiator needed in spring of 2019... arg

PS  So, what to do next?  Obviously the single can of refrigerant that I used is a red flag and the freeze switch needs a revisit. But before I make changes I want to know more about what is wrong. I bought a book,Just Needs a Recharge by Rob Siegel. This guy is definitely on my frequency. He is a BMW guy and covers all the bases of bringing back your long dead AC system.  Hopefully, Rob will help me get to the next level with my AC conversion.  

Copyright © 2022 ArnnWorx Specialty Tools
Credit Cards Welcome
Your IP Address is:
warobbo template
Please come back later
Free SSL