Get a better experience by installing our free app!
Not now
Install
Get a better experience by installing our web app!
Hide
How to do it?
In Safari, tap on the menu bar. Scroll down the list of options, then tap Add to Home Screen.(If you don’t see Add to Home Screen, you can add it. Scroll down to the bottom of the list, tap Edit Actions, then tap Add to Home Screen.)
dr bob wrote: Sun Feb 25, 2024 2:34 pm
Great video!
Guessing that switch has the same internals as the one in my 928 from Behr. The fluid in the capillary tube expands with heat, and the business end inside the little box is plumbed to the end of a little beryllium-copper bellows. Warm, the expanding fluid extends the bellows and actuates a common microswitch. Cold, the bellow shrinks with the fluid, and the switch opens. There's a small amount of adjustment available at the microswitch mount to move the actuation range. That range is determined by the 'deadband' in the microswitch itself. Colder-but-not-quite-freezing offers lowest vent temp, with the switch closer to the bellows. I had mine set for 34º open, and it would close at about 39º.
But... That bellows work-hardens and the action against the switch gets soggier and soggier with use. Eventually, the shrinking fluid with cold is no longer able to draw the less-flexible bellows smaller, and the switch stays closed. The only real solution is replacement of the whole thing.
Thanks Dr. Bob! Someone else on YouTube made that same comment about the adjustment screw, so I tried again after adjusting the screw and was able to confirm it the switch does open in a glass of ice water when adjusted counter-clockwise to the highest temp setting. I updated the video and will post that shortly. I was motivated to dig in because I could not find anything about this part on the internet -- which is a very rare occurrence these days for the 944. Nice to have Carpokes-exclusive tech info too, at least for now.
Thanks for the great video Tom! I took a look in the footwell, and it appears this is doable on the car without the evaporator assembly out of the car. I was just too nervous to start yanking to pull the tube out on mine, as didn't know if that would break anything.
And thanks for noting that part number. Mine had a non-Porsche part number on it so I couldn't find replacements. It seems these are available on ebay for around $20 used. Worth replacing as opposed to bypassing if its that cheap. Mine is on the way!
TurboD wrote: Wed Feb 28, 2024 6:35 pm
<...>
It seems these are available on ebay for around $20 used. Worth replacing as opposed to bypassing if its that cheap. Mine is on the way!
My casual advice: Invest in a new one.
Used ones are pretty much a sucker deal. since the reasons they are available used tends to be that the bellows got too stiff to function reliably. As-new, the switch function offers about 5º of deadband. As the bellows get work-hardened, that deadband grows. That starts with switch-always-closed. Adjustment as in Tom's video regains the anti-icing function but sacrifices some cabin cooling as the evaporator needs to get a bit warmer and warmer before the switch recloses.
----
The 928 uses a tiny compressor clutch relay mounted on a board in the control head. The contacts aren't sufficient for the clutch current, and more importantly don't take to the inductive 'kick' voltage spike that happens when the relay opens. The freeze switch lives in the circuit between those relay contacts and the clutch coil, so it too faces the same spike and arc-welding in the switch contacts. Porsche added suppressor diodes in a separate module, but those deteriorate with use and age to the point where they are pretty worthless. My 'black 928 living in Los Angeles with no tint' solution includes a third-party electronic temp switch, two relays in series, and a time delay circuit that bypasses the freeze switch contact for the first 15 minutes of compressor operation. The car is converted to R-134a and slightly under-charged, and will easily draw the evaporator temps to single digits if left to itself. So on a hot-soak restart, coldest available evaporator temps work to quickly pump heat out of the cabin. Before any significant icing can block evaporator airflow, the switch bypass is lifted and the switch cycles normally again. The electronic switch is adjustable for open temp and deadband. The danger is only that a short stop like fuel while touring, and having the vent temps easily cold enough to freeze fingers on the steering wheel. I can live with that. I need to do a whole write-up on that. I started it a couple summers ago now but it was abruptly derailed by an aggravated host...
Since moving to cooler Oregon, I have adjusted out most of that start-up delay. No hot-day downtown meetings anymore. Blessings counted regularly.
TurboD wrote: Wed Feb 28, 2024 6:35 pm
<...>
It seems these are available on ebay for around $20 used. Worth replacing as opposed to bypassing if its that cheap. Mine is on the way!
My casual advice: Invest in a new one.
Used ones are pretty much a sucker deal. since the reasons they are available used tends to be that the bellows got too stiff to function reliably. As-new, the switch function offers about 5º of deadband. As the bellows get work-hardened, that deadband grows. That starts with switch-always-closed. Adjustment as in Tom's video regains the anti-icing function but sacrifices some cabin cooling as the evaporator needs to get a bit warmer and warmer before the switch recloses.
----
The 928 uses a tiny compressor clutch relay mounted on a board in the control head. The contacts aren't sufficient for the clutch current, and more importantly don't take to the inductive 'kick' voltage spike that happens when the relay opens. The freeze switch lives in the circuit between those relay contacts and the clutch coil, so it too faces the same spike and arc-welding in the switch contacts. Porsche added suppressor diodes in a separate module, but those deteriorate with use and age to the point where they are pretty worthless. My 'black 928 living in Los Angeles with no tint' solution includes a third-party electronic temp switch, two relays in series, and a time delay circuit that bypasses the freeze switch contact for the first 15 minutes of compressor operation. The car is converted to R-134a and slightly under-charged, and will easily draw the evaporator temps to single digits if left to itself. So on a hot-soak restart, coldest available evaporator temps work to quickly pump heat out of the cabin. Before any significant icing can block evaporator airflow, the switch bypass is lifted and the switch cycles normally again. The electronic switch is adjustable for open temp and deadband. The danger is only that a short stop like fuel while touring, and having the vent temps easily cold enough to freeze fingers on the steering wheel. I can live with that. I need to do a whole write-up on that. I started it a couple summers ago now but it was abruptly derailed by an aggravated host...
Since moving to cooler Oregon, I have adjusted out most of that start-up delay. No hot-day downtown meetings anymore. Blessings counted regularly.
Seems like the kind of thing a microcontroller could control with aplomb, with a relay, micro temp sensor, and maybe with one or two engine inputs.
Seems like the kind of thing a microcontroller could control with aplomb, with a relay, micro temp sensor, and maybe with one or two engine inputs.
My less than elegant but way functional solution uses a 556 dual-timer IC, a couple transistors, a potentiometer and a few discrete components, all mounted on a little prototype perfboard. Two small relays to handle the clutch coil current. If you don't suffer the same black car in hot climate blues, the better sensor switch with just a simple relay is even more elegant in its simplicity and relatively bulletproof reliability.
The microcontroller, on the other hand, offers a BT interface so you can continuously monitor evap temps on your phone.
officer: It looks like you are using your phone in traffic, sir!
driver: No, just checking AC evaporator temperature in this heavy traffic, officer.
I wanted to offer some follow-up on parts and pieces I used, on the chance that they might help with the problems shared here. I dug back through notes and found the adjustable anti-freeze switch I used. I bought this one on Amazon on a whim when I saw it on a 'deals' listing for under $5 each. (a while ago now...)https://www.amazon.com/gp/product/B003R43E0O As I type this in 3/24, the listing price is only about $10, so certainly cheap enough to buy and play with. Cheap enough then that I bought an extra for the spares bin. (pic below)
A key part of using this switch is that it needs to feed a relay for the AC compressor clutch, rather than directly switch clutch coil current. In my 928, the factory uses a too-small relay in the control head, and the current from the relay passes through the anti-freeze switch on the way to the clutch. The factory fit that anti-freeze switch in the cowl area, accessible under the hood and the rain shield fitted over that area at the base of the windscreen. To make this new adjustable switch work in my car, I added two small Omron SPST relays in series after this anti-freeze switch. The relays are common in many Japanese cars these days, and are much more compact than the Bosch 53B relays that were factory fitted for similar duties the 928. The two-in-series arrangement helps with contact arcing when the relay opens and the clutch coil magnetic field collapses. I added four suppression diodes in parallel to ground at the clutch coil terminal "just to be sure", probably overkill with the two relays in series. Omron G8HL-H71 on the relays, 12V 120 Ohm coils, 40A contact rating.
Dimensions of the base of the box part are 39mm the long side, with the sensor tube coming out the end. Width of the base is about 29mm, with a couple little plastic tabs sticking out a little further. The tube is about 2' long, with a nominal 2.5mm OD.
The car is in winter hibernation storage right now, on tall stands and under some serious protective covering (movers pads under a regular car cover), else I'd share pics of my installation. The car will come out of storage in six weeks or so, and I can add pics then if anyone is interested.
Funny to see this. I bought the non-adjustable one from Amazon when I saw that it was a mere $4.01
I tested it and it seems to work about the same as the original 944 version at the mid-point of the adjustment range. I briefly contemplated modeling an adapter bracket to 3D print, but wasn't sure the 'demand' would be there. The little bulb on the end is a little too big to fit in the 944 evaporator, but I was able to file it down enough to fit without compromising the capillary tube. At $4.01, for people planning to use them in a 944, you might buy a couple in case you break into the capillary trying to file down the bulb.
** caution - sarcastic non-technical socio-economic commentary included. Proceed at your own risk! **
Makes you wonder some on big-picture trade issues. On an immediate level, you have to wonder how these are manufactured, shipped to port in China, containerized there, floated to the U.S., unloaded by union longshoremen, transported in the container to a sorting and distribution center, shipped to Amazon distribution, inventoried and catalogued, pulled, packed, shipped, and delivered to my front door... for under $10. And every hand that touched it along the way made money doing it. Makes it tough to compete.
