Hard Cold Starting 3.2 Carrerra

[Greg Greer]

My 1986 3.2 Carrerra has been hard to start from cold this season. It usually fires right up after winter. Now, it takes a lot of turning over before sputtering to life, and idles rough at first. After a few minutes or so and blips of throttle, it idles and runs normally out on the road, will restart easily after running, but has been very difficult to cold start 3 times in a row now.
What are likely culprits?
Thanks always for sharing your hard won knowledge!

[Arne2]

The car is probably not holding pressure in the injection over longer periods. I'd look for leaking injectors or leaking back flow valve in the pressure regulator or fuel pump.

[Tom]

My 1986 3.2 Carrerra has been hard to start from cold this season. It usually fires right up after winter. Now, it takes a lot of turning over before sputtering to life, and idles rough at first. After a few minutes or so and blips of throttle, it idles and runs normally out on the road, will restart easily after running, but has been very difficult to cold start 3 times in a row now.
What are likely culprits?
Thanks always for sharing your hard won knowledge!

Could be lots of things, of course, but I would be tempted to check the cylinder head temp sensor. I have not owned a 3.2 in years (and miss it) but its engine management/DME is very similar to the 944, which I am familiar with. The DME (ECU computer) is virtually identical, but since it is air cooled, your car has a cylinder head temp sensor in lieu of the DME coolant temp sensor. But the DME doesn't know the difference and expects the same values on both cars. Below is a an excerpt from the factory manual for your car, and a more detailed explanation for the 944. Also see an article on Pelican, linked below, suggesting the ground wires are known to go bad on these sensors.

If the system were not holding fuel pressure, I would expect the motor to resume smooth running fairly shortly after firing up (like under 30 seconds) once the fuel pump is pressurizing the system. Is it? And if it sat all winter, probably worth going over more basic maintenance possibilities too -- ignition, filters, vacuum leaks, fuel quality (run it low and fill it up with fresh gas), etc. The temp sensor may not be it for sure, but would be high on my list of things to check first.

CHT-sensor.gif (322.65 KiB) Viewed 1886 times

Here is a link to the Carpokes test procedure for the 944. The test on page 2 of this would be identical on your car.
viewtopic.php?t=695


And here is the Pelican article I mentioned.
https://www.pelicanparts.com/techarticl ... sensor.htm

[Greg Greer]

My 1986 3.2 Carrerra has been hard to start from cold this season. It usually fires right up after winter. Now, it takes a lot of turning over before sputtering to life, and idles rough at first. After a few minutes or so and blips of throttle, it idles and runs normally out on the road, will restart easily after running, but has been very difficult to cold start 3 times in a row now.
What are likely culprits?
Thanks always for sharing your hard won knowledge!

Could be lots of things, of course, but I would be tempted to check the cylinder head temp sensor. I have not owned a 3.2 in years (and miss it) but its engine management/DME is very similar to the 944, which I am familiar with. The DME (ECU computer) is virtually identical, but since it is air cooled, your car has a cylinder head temp sensor in lieu of the DME coolant temp sensor. But the DME doesn't know the difference and expects the same values on both cars. Below is a an excerpt from the factory manual for your car, and a more detailed explanation for the 944. Also see an article on Pelican, linked below, suggesting the ground wires are known to go bad on these sensors.

If the system were not holding fuel pressure, I would expect the motor to resume smooth running fairly shortly after firing up (like under 30 seconds) once the fuel pump is pressurizing the system. Is it? And if it sat all winter, probably worth going over more basic maintenance possibilities too -- ignition, filters, vacuum leaks, fuel quality (run it low and fill it up with fresh gas), etc. The temp sensor may not be it for sure, but would be high on my list of things to check first.


CHT-sensor.gif



Here is a link to the Carpokes test procedure for the 944. The test on page 2 of this would be identical on your car.
viewtopic.php?t=695


And here is the Pelican article I mentioned.
https://www.pelicanparts.com/techarticl ... sensor.htm

Thanks for the suggestion! We went in today to test and uncovered another mystery. Our cylinder head temperature sensor has only one prong. The female/wiring harness plug has two slots. It's clear that we didn't break a prong off during the unplugging process. The plastic inside of the cylinder head temperature sensor is smooth and solid where the second prong should be. We tested it by putting the positive lead on our one prong and grounding the negative lead. Nothing. So, we assume (knowing full well what that means) that the sensor is kaput. The replacement sensors we see for sale all have two prongs. This isn't extremely worrisome, as the female end has two slots. Still, it's a mystery. Why does ours have just one prong? After almost 38 years of a single prong, is our DME going to freak out if we replace the sensor with a dual-pronged model?

[Tom]

The original 1-wire sensors are grounded only by the threads in the head, and are known to fail. Porsche upgraded to a 2-wire sensor that has a dedicated ground wire, instead of relying solely on the sensor-to-head threads. Once Porsche supersedes a part like that, you generally can only find the new part at dealers and auto parts retailers -- so no surprise you are only finding the 2-wire version.

Why does your car have a 1-wire sensor and a 2-wire connector? I don't know for sure, but it is common on the 3.2 cars. My guess would be that Porsche just wanted to use up the parts it had -- probably the 1-wire sensors -- but planned ahead for the more reliable 2-wire sensor. Some claim that's why the 912 and 924S even existed -- to use up parts leftover from the earlier cars. But that's speculation mostly -- what I do know is that your car is not unusual. A lot of them came like that. Makes for a very easy upgrade anyway.

Apologies if this is too basic for you, but it's always a good idea to make sure your multimeter readings are correct before buying parts. If you touch the probes together, do you see nearly zero ohms? If you touch them both to the area where you are grounding it, do you see nearly zero ohms? If you test something else to read resistance, are you seeing the expected value? Not doubting your results at all, or assuming you did it wrong, but it's something I've seen people do before so thought I'd mention it -- especially with lower-end multimeters where you have to manually set the resistance range.

[Greg Greer]

The original 1-wire sensors are grounded only by the threads in the head, and are known to fail. Porsche upgraded to a 2-wire sensor that has a dedicated ground wire, instead of relying solely on the sensor-to-head threads. Once Porsche supersedes a part like that, you generally can only find the new part at dealers and auto parts retailers -- so no surprise you are only finding the 2-wire version.

Why does your car have a 1-wire sensor and a 2-wire connector? I don't know for sure, but it is common on the 3.2 cars. My guess would be that Porsche just wanted to use up the parts it had -- probably the 1-wire sensors -- but planned ahead for the more reliable 2-wire sensor. Some claim that's why the 912 and 924S even existed -- to use up parts leftover from the earlier cars. But that's speculation mostly -- what I do know is that your car is not unusual. A lot of them came like that. Makes for a very easy upgrade anyway.

Apologies if this is too basic for you, but it's always a good idea to make sure your multimeter readings are correct before buying parts. If you touch the probes together, do you see nearly zero ohms? If you touch them both to the area where you are grounding it, do you see nearly zero ohms? If you test something else to read resistance, are you seeing the expected value? Not doubting your results at all, or assuming you did it wrong, but it's something I've seen people do before so thought I'd mention it -- especially with lower-end multimeters where you have to manually set the resistance range.

Thank you for the explanation. Your refresher on the multimeter was not too basic. We use it regularly enough to be comfortable, infrequently enough to prefer to check the instruction book every time. We went back to triple-check the sensor, and, yup, kaput! Fingers crossed that the new one fixes our problem.

[cpgaylor]

curiosity, what was the fix? dme relay, 02 sensor??? thx

[Greg Greer]

Hi cpgaylor and everyone else,
The cylinder head temperature sensor was the culprit! After replacing it, the car started right up in that magical fuel injected way.
HOWEVER, this project threw me for quite a loop. The new sensor would not take into the cylinder head threads after hours upon hours of trying to get it to start. It was very difficult to get a good alignment of the sensor and cylinder head threads in that tiny space. Also, my special deep sockets ended up bumping against the engine tin access hole after only 1/2 to 3/4 of a revolution and throwing off the alignment of sensor to cylinder head threads. I was really starting to panic and think this new sensor was not going back in. Anyway, after taking some time off from it, and getting more advice from the community, I ended up beveling the start of the sensor's threads, then running a die wheel M10 X 1.0 up and down the sensor's threads, and re tapping the cylinder head threads just a turn or so. Then, the sensor took with only my fingers turning it. I really could not maintain these cars without all of your shared knowledge, Thanks again everyone!