No spark tried everything - Ingo even loaned me his DME??

[4flyboy]

Will do, appreciate the offer.

Scott

[86 911 Targa]

Scott,

Have you tried placing a diode across pins #25 & 26,
and pins #27 & 8?

Also, as per my original suggestion, look for
12v at pin # 86 of the DME relay.

good luck,

Gerry

[4flyboy]

Thanks Gerry,

And everyone else who's given suggestions. I know that Alex is going to try a jumper and we are working through every single post to find the cure. Alex is unable to post except through my account, I'm sure he'll be back with results.

Gerry, how would i know if there is a factory alarm or ignition disable on the car? Were they included on all 3.2's?

Scott

[ischmitz]

The factory alarm cuts power from the ignition switch to the DME relay. As long as both stages of the relay stay on and as a result the fuel pump runs during cranking this is fine and not an alarm issue.

[4flyboy]

Understood Ingo,

We'll cross that one off the list, appreciate it Gerry.

Scott

[86 911 Targa]

Originally Posted by ischmitz

The factory alarm cuts power from the ignition switch to the DME relay. As long as both stages of the relay stay on and as a result the fuel pump runs during cranking this is fine and not an alarm issue.

Maybe I missed something, are both relays active?

Gerry

[techman1]

Scott,

How long have you owned this car? How original is all of the wiring? I recall an issue for an older porsche or BMW; the previous owner had put in a hidden kill switch. I think it was grounding one of the sensor signals. This does not sound like an issue, as you have scoped out all of the DME sensors.
So, maybe not grounding one of the sensor inputs to the DME, but maybe an output? What signals do the DME output that if grounded, would cause these symptoms? Or maybe not a kill switch, but a bare chaffed wire grounding accidentally? Signal to coil? Just throwing out ideas...

[86 911 Targa]

Originally Posted by 4flyboy

Thanks Gerry,

And everyone else who's given suggestions. I know that Alex is going to try a jumper and we are working through every single post to find the cure. Alex is unable to post except through my account, I'm sure he'll be back with results.

Gerry, how would i know if there is a factory alarm or ignition disable on the car? Were they included on all 3.2's?

Scott

https://www.google.com/imgres?imgurl=http%3A%2F%2Fforums.pelicanparts.com %2Fuploads25%2F911_alarm21460643051.jpg&imgrefurl=http%3A%2F%2Fforums.pelicanparts.com%2Fp orsche-911-technical-forum%2F910180-how-does-factory-alarm-disable-car.html&docid=4PcP5cjNVHVyqM&tbnid=1rs7GyhkkTdV8M%3A&w=567&h=352&bih=813&biw=1258&ved=0ahUKEwiH56TR-8_NAhVMyWMKHTW0C6MQMwg9KBUwFQ&iact=mrc&uact=8
^^^^^
Here you go.

If you do not have the alarm option, the 12v to pin # 86 of the DME relay
is sourced by a bridging connector in place of the alarm module.

Best,

gerry

[al525i]

It's fixed. I'm still not sure exactly what is wrong but the shielded wire that runs in the harness from the speed sensor to the DME is bad somehow. I replaced the entire speed sensor harness and it runs fine... More in depth write up to come

[ischmitz]

Good news.

I am still baffled how this wouldn't have been diagnosed right away by putting the oscilloscope onto pin 8 & pin 27 inside the DME connector while cranking. If the harness or sensor or gap is compromised the speed signal is bad. Even a DVM in AC setting would show less than the 2V AC (maybe nothing at all).

[86 911 Targa]

Originally Posted by ischmitz

Good news.

I am still baffled how this wouldn't have been diagnosed right away by putting the oscilloscope onto pin 8 & pin 27 inside the DME connector while cranking. If the harness or sensor or gap is compromised the speed signal is bad. Even a DVM in AC setting would show less than the 2V AC (maybe nothing at all).

Or the Diode test which would have picked up the harness/connector issue..............

Gerry

[mysocal911]

Post #2:

Originally Posted by al525i

This seemed like a speed or reference sensor problem so I performed a resistance test on the sensors. They were within spec. Next I broke out the oscilloscope. The speed sensor waveform looked great but there was some noise in the reference sensor output.

This possibly was not performed at the ECM connector causing a mis-directed troubleshooting effort,
resulting in wasted time & money on parts, e.g. ECM test & mod (28 pin)?

[rick-l]

Originally Posted by ischmitz

I am still baffled how this wouldn't have been diagnosed right away by putting the oscilloscope onto pin 8 & pin 27 inside the DME connector while cranking.

Post 16 perhaps????

No spark tried everything - Ingo even loaned me his DME??

[al525i]

I'll put up a screen shot of the wave pattern later but yes, the scop test was performed at the dme connector between pins 27 &8 for the speed sensor on channel one and pins 25&26 for the reference sensor on channel two of the scope .
The wave forms looked great.

That's why I say I don't understand why the solution was the solution.

The shielded wire for the speed sensor passed every single test I threw at it yet was keeping the car from starting. I know motronic systems very well and this is the first time one has thrown me for a ride in several years

[yelcab1]

Congratulation on the find and the fix. That is the most important thing.

[ant7]

+1

Originally Posted by yelcab1

congratulation on the find and the fix. That is the most important thing.

[al525i]

Here is the waveform pattern from when it wouldn't start. As noted earlier. The scope leads were connected to the DME side of the harness at the appropriate places. the top line in yellow is the speed sensor. the Green is the reference sensor.

Ingo is a major resource to everyone here and I greatly appreciate the time he's spent talking with me both online and over the phone. His help in this cannot be overstated



Everyone else had some great ideas and I'm sorry I wasn't able to respond before due to the forum locking me out. All the basic stuff had been done (several times) before it got to the point of posting here. The post suggesting the scope being isolated could be a problem sparked the idea to slave in test wires for the speed and reference sensors so thank you rick-l!

[rick-l]

Where did you reference (ground) the O-scope?

[ischmitz]

Good point Rick. What's throwing a red-flag on the screenshot is the varying amplitude of the speed signal. This tells me something is funky. I wonder how this looks now that the cable is replaced. Was this measured with the DME connected or disconnected?

The o-scope used looks like a Windows PC-based USB version. If this is on a laptop plugged into AC with a charger or a PC the mains earth and GND are likely tied together. Never connect the GND lead of the test head to anything but GND, else there will be damage. Ask me how I know...... Further, when you connect the GND leads from channel 1 and channel 2 on a multichannel o-scope to different pins on the DME you essentially pull both to GND and short them together in the process. Inside the DME pin 8, 25, 26, 27 are all different and none should be shorted to another. So measuring both with the DME connected already throws a curve-ball to the flywheel decoder.

For a real measurement of a single differential signal (such as the speed signal) one would use channel 1's input to pin 8 and channel 2's input to pin 27 with both GND leads from the test heads connected to GND on the DME. Then use channel math (ch1-ch2) to get the true differential signal. Because of that I prefer a Fluke 123 hand-held on its internal battery for this type of work. I can connect its GND lead to anything and get a true differential measurement without having to worry about GND loops etc. It acts just like a digital voltmeter.

Ingo

[mysocal911]

Originally Posted by rick-l

Where did you reference (ground) the O-scope?

The referencing of ground is not an issue when probing and the ECM is not connected. Furthermore,
each of the two pins of both the reference & speed sensors "float" together (8 & 27 - 1K ohms) and
(25 & 26 - 1K ohms) when not connected to the ECM and provide a low impedance voltage source
while cranking and measured separately.

But, since many lack the knowledge of using an O-scope beyond its very basic features, the two sensors
should have been evaluated using a VOM with the scale set to VAC. The ref sensor typically measures
about .10 VAC and the speed sensor measures about 1.0 VAC while cranking. This was pointed out
in post #15. There's no need to make the testing overly complex, e.g. "true differential measurement",
when a simple VOM measurement suffices! Furthermore, using a VOM allows the measurements while
the ECM is power-up.

So, another simple 911 3.2 troubleshooting effort that became overly complex and mis-guided for
the 911 owner wasting his time and money.