Two starting issues -- related or not?
 

I've got a 1995 993 engine in my 1972 911 tub.

Last week, I had an intermittent starting problem that has gotten worse. When it started, maybe one in ten start attempts would have the thing crank and crank with no (apparent) spark or gas smell. Like anyone would, I suspected the DME relay, but replacing it did not change the problem. And now it's happening maybe eight out of ten times. So I've got to fix it.

My suspicion is my crank position sensor. Does anyone have a simple way to test the voltage coming out of it? I suspect the only way is to slide a needle into the plug connector in the engine compartment and test voltage there?

Now, for maybe a year or so, I've had an occasional problem where the solenoid will 'click' but the starter will not crank. In my not-too-bright head, I always thought I had a 'dead spot' somewhere with the starter motor, but I never really thought about it very much. The solution has always been to either bump start it just enough to move that 'dead spot' a little -- at which point it cranks and starts just fine. Alternately, sometimes it's enough to just click and re-click the solenoid action with the starter switch, which seems to give it enough of an effective light rap to allow it to crank normally. The clicking problem happens maybe one in a hundred times.

Neither problem seems to have anything to do with the level of battery charge.

Is there any chance that these two things are related?

And do you think I on the right track with suspecting the crank position sensor? And is there any fix for an old one outside of replacement?

I've also thought it could be the engine ground or the fuel pump.

I wouldn't think the two problems are related, unless there is some potential for questionable wiring harness locations. My 74 has just such a questionable harness and as I fix/cleanup/replace portions here and there little electrical gremlins seem to go away. Any chance you have something chafing?

A poor/intermittent connection(***) at the starter solenoid might cause both symptoms.

*** INTER-CONNECTION, actually.

Heavy duty wire from the battery to the starter solenoid, smaller wire from there to the alternator and ALL engine compartment accessories.

Crank, but no start, battery connection to solenoid okay, connection to engine bay, NOT.

Just a "click", heavy duty connection at starter open.

Thanks, guys.

I've fixed the non-start issue, although I wouldn't describe it as a repair and wouldn't say it's not going to be a problem again. I started by going through and checking ground connections -- both by emorying and reattaching and also using an old set of jumper cables as a makeshift bridge. It was one of those times when having a lift seems like the greatest thing ever.

But I still had the crank and no apparent gas/spark. So I took the driver seat out to reconfirm the grounds at the DME and triple-check the DME relay. Still nothing, until by chance I gave the top of the DME a tap while cranking the engine. Zoom it came to life, steady idle -- just like it used to.

So I took out the DME, took off the cover and made sure the chip(s) were seated. Back in and I still can't get it to fail. So my fix was just a tap -- which isn't very reassuring. But at least it points to a location -- and a target for a tap the next time I get a no-start.

The bad news is that I might have a DME repair/rebuild in my future.

The good news is that I hadn't ordered a new Crank Position Sensor yet, and didn't have to mess with checking the clearance on the one I've got.

I plan on taking a look at the starter wiring before I put the seat back in. My car has a panel where the back seat used to be, so getting to the starter doesn't mean hugging the transaxle. :)

Hey Jack, agreed the tap would make me nervous, too. And unfortunately, intermittent faults are the hardest to track down.

In your case I am wondering how things add up? If the DME is sensitive to vibrations due to an internal fault I would expect effects when driving, too. E.g. a slight miss at some point. Your knock should not be very different from normal vibrations transferred from the car chassis to the DME housing.

The DME is actually pretty stupid when it comes to starting vs. running. If it sees a correct flywheel signal it is running and issuing spark and fuel signals. If the flywheel signal isn't present the fuel/spark signals stop momentarily.


Cheers,
Ingo

When I have experienced issues like this they were either ground related OR DME sensor related.

Ingo, it turns out you're correct. After a consecutive seven or eight easy starts, it went back to the crank-only condition, and this time tapping on the DME box made no difference at all -- which leads me to believe it was just a coincidence that it went into the functioning mode after my initial tap. Since then, it's cycled through a functioning and non-functioning cycle again.

So I'm back to my suspicion that my crank position sensor is sometimes failing to send the requisite volts back to the DME for it to enable the coil (etc.) during startup. Either that or my fuel pump is failing -- although I'd think that would not fail in such a clear on-off fashion. When the car does fire up now, there's no stumbling (like there was when my MAF sensor was bad, which mjshira saved me on).

So tomorrow I'll try to test the voltage coming back from the crank position sensor. Does anyone know what the correct voltage range should be?

I don't know of what other sensors would produce this condition. If anyone has other suggestions, I'm all ears.

Jack,

you need use an oscilloscope to the CPS pins in the 55pin connector and then crank. The O-scope should show at least 3Vpp while cranking. See here for the WSM page. For a quick check you measure resistance between the "hot" pins 1 and 2 of the black reference sensor plug. I just measured mine and it is 535 Ohm (my engine is currently out of the car)

http://forums.pelicanparts.com/uploa...1388382475.jpg

Let me know how it goes. If all else fails I have a 964 loaner DME for testing....

ischmitz, can Jack also check the voltage with a VOM set to AC?

It's not helping your diagnosis, but when you mentioned smacking the DME it made me thing of this
http://www.derekmohrdesign.com/image...omMANdment.jpg

I guess a DMV set to AC is a reasonable start. First, I would check for resistance to see if the sensor is open. Then a DVM set to AC to see what you get.

You won't see if there are drop-outs but for a no-start it should give you some good information.

Ingo

Check the interconnect at the starter solenoid?

All right. I'm going to remind those of you who don't already know: I have about a 5-year-old's grasp of electronics. So let me explain what I'm doing on and those of you with real eductations can hopefully guide me right.

There are three wires from the sensor, a white, a brown and a thicker black wire.

http://forums.pelicanparts.com/uploa...1388430207.jpg

When I check resistance between the white and the brown wires, I get a reading of 515 ohms.

Then, when I crank the starter, I get an AC reading that varies between 1.9 and 2.5 volts, which I would conclude is faulty, based on Ingo's 3V number.

So the next question would be whether I should replace the sensor or try to figure if it's incorrectly spaced. I'm inclined to just replace it, since it's not failing in a consistent manner, and nothing has moved the sensor relative to the crank as far as I can see.

Should I just hit it with some penetrating oil and order a new sensor?

And wwest, I also do want to check the solonoid and starter. But (again, because I'm not very smart about this stuff) could you give me a couple of steps to do that?

And one more wrinkle: I opened up my rear seat panel and discovered that the bolt on the stud for the big wire into the starter was... kind of loose.

It looks like the stud is also a feed for five 12V feeds headed back toward the engine compartment - 1 thick, 1 thin and 3 medium red wires. I cleaned everything up, tightened the bolt down, and (here we go again -- it's like dating a German chick) the engine started right up.

It's started three times in a row, now.

So it could be that the movement there was interrupting/restoring power to downtream stuff and could have been messing with one loop or another in my startup symphony. Or not. The thing with an intermittent failure is, well... it's intermittent.

But the whole problem has cycled through pretty reliably before, so I'll keep testing the thing to see if I should go ahead and order a new crank position sensor or if I'm done and the old mongrel is good again.

Stay tuned for more. :)

I am not a Porsche owner, so this is coming from my associations with others like yourself and reading innumerable threads on related subjects. If the sensor replacement does not work, could the ignition switch be the culprit? The intermittent nature you discussed seems to be a common element in all of the "it was the 8#$%*! ignition switch" threads I've gone through. If you already have a new switch, then never mind.

Good luck. I have a hunch you will track it down in no time.

Jack, the sensor is most likely fine. With the DVM you measure the RMS voltage of the sensor. With an oscilloscope you measure peak-to-peak. With a sine wave the peak-to-peak value is 2.8 *RMS. See here:

http://forums.pelicanparts.com/uploa...1388438227.jpg
(amazing how smart google is..... :))

I believe the loose +12V is what caused your issues. It is probably what supplies power to the injectors and coil packs and with that supply compromised I can see an intermittent no-start.

Ingo

So far (again) it looks like you are correct, Ingo. I keep trying it every few minutes. Previously, this would have led to a non-working cycle and then a working one. But it's been consistently good.

So it might have just been a simple loose bolt on a key 12V connection. And that might solve the intermittent clicking/no-crank problem, too.

Exactly. I have a whole wall dedicated to 'electronic diagnosis tools.'

http://12-gaugegarage.com/photos/fil...-1024-full.jpg

Dozens of starts now, and both issues seem to have been fixed. Tightening a loose bolt isn't going to make me any kind of internet Porsche repair hero, but it sure was cheap. :cool: