intermittant non-start issue is driving me nuts

[andyt11]

The car is an 88 3.2 911, stock except for a Steve Wong chip and an exhaust.

For probably three driving seasons now I've had maybe twice a year where the car turns over but wont fire. Twice it happened after the car wash, the other times there was no water involved. I have never got to the bottom of it, because it fixes itself during the time I'm doing diagnostics.

This last time I didn't even get into diagnostics. The car drove out of town for an hour to a restaurant last night, but after the meal it would only turn over, it wouldn't fire and run. In the end we came home on a flat deck. The very next day today, I tried to start it. Same result, wouldn't fire up. So I pushed it around in the garage because it was in the way. When I was done, I turn the key just for fun, and the car started. It's infuriating, because now it behaves perfectly, and when I get into it with the multi meter, everything will check out just like in previous years.

This has happened five or six times over the last three years. Each time I find all the sensors etc check out as they should, coil is fine, DME relay is fine, idle valve is doing its thing, got spark, got fuel, etc etc... but at some point during diagnostics the problem vanishes. So of course everything checks out.

I do have a suspect main HT lead, I found it had failed/disintegrated last year and i re-crimped one end and put it back into service. It was late in the season. I've ordered new ones now as I suspect the lead is failing. Apart from that everything on the car checks out.

There is one other thing I noticed during the years, and that is when the car is in a non-starting condition, I can get it to fire really roughly for a few seconds if I unplug the AFM. But I think doing so floods the engine because you can really smell the fuel if you do that.

My plan right now it put the new HT leads on c/w with a dizzy cap and rotor. The ones that are on are not exactly in the best shape and it's time anyway. The car will run after that i'm sure, but then it runs right now. I'm unsure how I can know if I've fixed the problem.

So looking at this latest case, what can fail at random, and then be cured by being pushed around the garage for ten minutes?

[88911coupe]

In my case random no starts ended up being due to one of the RPM or reference sensors failing intermittently. I'd been under the impression they either worked or failed but in my case the resistance would be in spec and wildly out of spec depending on when I checked it. Ischmitz is the one who guided me through what to check and when. After replacing the buggy sensor it starts fine and consistently.

[strikee]

Im with 88911 on this one. I had the same thing. Would go to start the car and it would just turn over. Nothing. Walk away for five minutes. Shake the car side to side and bam, it would start.

I replaced both the reference sensors (with the BMW ones) and it has started first time every time since. One of the most rewarding DIY jobs I have done.

Don't be fooled into thinking its a five minute job. My reference sensors were pretty much stuck in the mounts. Lots of really good threads on how to do this.
Ed.

[T77911S]

replace the 2 sensors as above.
either replace the DME relay or pull it out and resolder the joint
ignition sw is another possibility although I think most of the tome with this it will not turn over

have a spare spark plug with you. when it will not start check for spark.

pulling the plug on the AFM probably puts the DME in a "default" mode. if it does, this mode will make it run richer. ( a lot of EFI cars will do this. I don't know for sure if the Porsche does or not). this mode is a "safe mode". since it does not know how much air is coming in it goes to a default setting of very rich.

[kyngfish]

+1 to everything above. My bet is its the crankshaft position sensor/ref sensor. you mentioned that it happens after a visit to a restaurant? Does it happen more when the car is warm?

Look around on these forums. The speed and ref sensors are the same part just used differently. If you find the Bosch part number you can get them for 60 each instead of 150. Replacing rhem should be easy peasy as long as you don't have to regap the bracket and if they aren't stuck.

[CCM911]

In my case it was the ECU. Ingo Schmitz had me up and running in no time.

[ivanuf]

It also happened to me and it was the battery. It had the 12V but the cranking power was not enough to turn the engine. Once I would not start at all. I left it on the driveway plugged to the battery tender and it crank fine the next day, that Is how I figured out it was a weak battery.

[andyt11]

I thought they were either failed or not failed too. Perhaps I should reach out to Ischmitz and ask him what his thoughts are.

[ivanuf]

Nope, we sell batteries in our store. A simple volt meter will only give you voltage but you need to verify the CCA (cold cranking amperage) this values for my SC needs to be in the 600+ range to start the car. A reading of 12V with a CCA of 400 from a good battery will not move my starter or start my car.

[DRACO A5OG]

+1, if cold voltage is less than 12.5 volts, she most likely will not start and CCA is most important for our 911's

[andyt11]

By the sounds of things, as much as I hate to throw parts at a car, I should just spend a couple hundred bucks and swap out the two sensors and enjoy the rest of the summer. And a new DME relay for good measure.

This is such an infrequent condition (once or twice a year) but it's making it so I can't trust the car. Won't be taking it on another date in the near future that's for sure.

Although they only happen once or twice a year too

[DRACO A5OG]

What is the voltage cold?

I have no issues, but I always put her on a tender just to keep the battery healthy and de-sulfating which kills batteries no matter the age.

[andyt11]

Right now I'm not sure. Its probably down from all the cranking on saturday. But as I said, it started and ran on Sunday, about 10 minutes after failing to start on Sunday.

I will put it on charge this week.

[Joe Bob]

If hot start, partially stuck solenoid or bad check valve in the fuel pump exit neck. Solenoid, whack it like a RH step kid, check valve replace. DME relay and sensors as suggested above. Those are the big four. Also, one that stumped me, worn fuse contacts from spinning to break ocean air corrosion.

[mysocal911]

Originally Posted by DRACO A5OG

+1, if cold voltage is less than 12.5 volts, she most likely will not start and CCA is most important for our 911's

Actually, most all vehicles should start in very cold weather when the voltage is as low as 9-10 volts.
Remember, there're climates other than what's found in SoCal, e.g. northern Europe, where the temp
is in the minus degrees C in the winter. Given that, how could any automotive manufacturer require that
a car's battery be at least 12.5 volts for adequate starting under all temps?

[andyt11]

Doesn't make any odds of the car is warm or cold, wet or dry. it's just as random and goes away just as random. Smells of fuel though. DME relay clicks twice. Car performs the same on a spare relay too.

When the car is refusing to start I get very definate tach bounce on cranking.

I read that the tach is given its signal from the reference sensors. If that's true, it would perhaps indicate that the sensors are fine and that I have no spark. I didn't have tools to check when it last broke down. And now the car runs if course.

Could the coil be failing intermittently? Would I get tach bounce if I didn't have a spark?

[DRACO A5OG]

Please check your voltage when cold or go to the auto parts store and have them check your battery and charging system, it should be free if a chain store. Make sure they check the CCA.

[mysocal911]

Originally Posted by andyt11

When the car is refusing to start I get very definate tach bounce on cranking.
I read that the tach is given its signal from the reference sensors.

Correct.

Originally Posted by andyt11

If that's true, it would perhaps indicate that the sensors are fine and that I have no spark.

Again correct. So you need to actually pull the coil wire from the distributor and check for a spark
(place wire 10mm from engine case).

Originally Posted by andyt11

Could the coil be failing intermittently?

Yes.

Originally Posted by andyt11

Would I get tach bounce if I didn't have a spark?

Yes.

[andyt11]

Next time it fails (probably a year from now) I'll check for a spark. Although I'm sure last year when I went through this we did check for a spark and it was present.

Last year it was kaput for a while and I was able to check the resistance on the position sensors and cht sensors. I'm sure we checked for spark, I'm sure we checked the DME relay socket too, and the relay.

[andyt11]

And Draco, I'll check voltage and charge battery tonight. It may be low, but that said it went from not running to running in about 5 minutes on Sunday.

And of course now it's running again there isn't much to diagnose.