PerryB

And it still doesn't start.

flagtie

OMG. That's just brutal. I am very sorry.

This just can't be some unsolvable out of this world thing. You can't just give up though. If he'll be so gracious as to help more, then send Paul those computers again. I'd really like to know if they start his car. My bet is they don't.

specsalot

Perry - What was going on during this last start attempt.

Spark?

Fuel?

Anything different from prior no-starts?

konakat

I agree with Flagtie. Before driving yourself crazy, see if Paul can try them again. I'm wondering if you have a fault that Specialized is not able to detect on their test rig.

Lecrok

PerryB, I assume you have this from the description of your troubleshooting so far : http://arnnworx.com/DME_KLR_Test_Plan.pdf

PerryB

Everyone, thanks for your replies.

Lecrok, I have the official Porsche diagnosis test plan and have followed it step by step to the best of my abilities. For the complete description of my work to date, confer with Post #45.

Specs, I still have no spark or injection. I put my oscilloscope on the KLR harness at pin #9 and it shows that I still get a signal to the coil that the DME is grounding the coil rather than alternating power and ground that would get the coil to fire. The DME does not generate the square tooth signal showing that it is converting the analog signals from the speed and reference sensors into the digital pulse.

I talked at length with a tech at Specialized ECU Repair today. He was sympathetic. He offered to have me send out the computers, have them installed into a car and send me a video of the car running (or not running). I mentioned that I had installed an aftermarket wiring harness between the speed and reference sensors and the DME. I told him that the new harness combines the shielding ground for these two sensors at pin #5 in the DME harness. (They were previously seperated at pin #5 for the reference sensor and #23 for the speed and 02 sensors). The tech didn't like this. I explained that according to the internal DME wiring diagrams that I have, these two pins join inside the computer. He explained that they test and rebuild their components to work with the stock harness and that in their work they used to combine grounds, but found that this generated inconsistent results. I asked him if having the harness set up as it is now (with the shield grounds combined on pin#5) would damage the computer, he said no.

So, my plan of attack now is to see if I can convert the new harness back to using seperate pins for the ground shielding. This will not be easy. The wiring harness at the DME connection is fragile, with small gauge wires and tiny crimp connectors that are very tightly packed into this brittle plastic gang plug. As I replaced the old harness with the new one, I was very aware that I could crack the plug, or screw something else up as I reassembled the harness.

I'm interested in other's opinions on this plan to re-seperate the ground shields for the speed and reference sensors.

Again, thanks to everyone for their sympathy and help.

Perry

specsalot

Although we purchased different reference / speed sensor harness kits, it appears that they were built the same way. I had mixed feelings while installing the kit because of this wiring change. But in the end it yielded acceptable results. Electrically the kit looks equivalent to OEM, except now both sensor shields are common externally, not just internally within the DME. Both crank sensors share a common ground landing on pin 5. My install required that I cut the chassis ground (brown wire) off of the original pin 5 and resolder it onto the new harness (which paired both crank sensor shields as bare wire) destined for pin 5. My instructions also required that I cut the black wire (old speed reference sensor ground) on pin 23, removing the old reference sensor ground from the circuit. I assume you had to do the same things.

Are you sure you have a good solder joint where you re-attached chassis ground to the shared crank sensor grounds at pin 5? If this joint is bad or somehow became disconnected, it could impact the ability of your electronics to interpret info from both crank sensors.

In my harness kit the speed sensor turned out to be miss-wired. Sensor Pin 1 was connected to DME pin 8 instead of 27. I think this was an assembly mistake. I solved this by just swapping plug pin's 27 and 8. The documentation references the criticality of having correct initial slope on the reference mark sensor output, but makes no similar reference for the speed reference sensor. I swapped the pin's anyway to match the intent of the wiring diagram.

The electrical shielding on both crank sensors functions to keep any nearby conductors from inducing stray voltages into the sensor leads. You wouldn't think it would make all that much of a difference how they were connected as long all shields were grounded. Feedback from Specialized is kind of eye-opening. The implication is that it does make a difference. It makes sense to reconnect them per OEM wiring diagram.

Some electrical questions - You've already discussed these in various posts - I assume the answer to all of the questions below would be YES B+ volts - Just asking again based on what these points supply.

1. When you turn your key on, do you see voltage on pin 86 of the DME relay?
2. Key on should be providing B+ to both terminal 2 and 3 at the 14 PIN connector (T21) located on the firewall aft of the brake booster?

Comment:
Pin 3 supplies B+ to the DME computer; Pin 2 supplies B+ to the injectors, timing valve, and KLR unit. If either one of these is not getting required voltage, this may be why your car won't start. Back in an earlier post you made, I inferred that T21 Pin 2 was not seeing B+ like it was supposed to. At that time you were applying B+ directly to that point which resulted in your system self-exciting (random coil / injector firing for a few seconds). Your old electronics exhibited this same self-excitation behavior briefly when I tested them in my car following "key on" events. But my observations were not as dramatic as what you saw.

I have re-read this thread looking for things missed. It looks like you have been very through and careful. The only thing which stands out beyond the above questions was the comment (post 40) about one of your window relay getting hot during your testing. This has been a real ordeal, which you've worked very hard at solving. Keep working at it and posting observations made during your efforts. The answers are there waiting to be discovered.

-Paul

PerryB

Specs,

Thanks for your reply.

Yes, I have B+ at DME relay socket 86, and at pins 1, 2, 3, 5 on the 14-pin connector with the key on.

I'm thinking of sending the computers to Specialized ECU Repair for them to plug them into a 951 that they have. If they can then see that the car doesn't start, they will know that there is something wrong with the computers, if it runs, then I know that there is something else wrong with my car.

Perry

specsalot

That makes good sense. I'm sorry this is proving to be such an ordeal. I guess it goes with owning anything that is borderline classified as vintage. As fat as I'm concerned these cars are special. Ahead of their time and worth the effort for the joy the bring to drive. Keep us posted.

PerryB

Thanks, Specs.

I took some time today to connect the ground shield wire for the speed sensor back onto female pin #23 at the DME harness, as per the original layout. Big PITA.

Still no start.

Perry

PerryB

For anyone that is still following this thread...

I sent both of my computers back to Specialized ECU Repair for them to retest and to put them into a test mule car that they have. They said that their car started and ran normally. I received the computers back yesterday and I also installed a new battery.

I have injection but no spark, so the car still does not start. I will hook up my oscilloscope back up to the output of the DME and to the output of the speed and reference sensors.

Perry

flagtie

Definitely still following. Good luck.

Alex Sol

arggggggghhhhhhhhh !!

Alex Sol

i really don't have much to add other than... some empathy... you have been so incredibly thorough and i am going through this thread looking forward to the 'simple' fix.... keep at it.. solution cannot be too far away and happy motoring to follow....

the wires on my 1987 951 are so brittle and every time i looking into the engine bay or under or around i find stripped wires, bare wires, cracked wires on grounds, and lots of corrosion. i use basic electrical tape to repair but have seen some very high end electrical tape that would provide more (and proper) insulation.

i live in Toronto and drive the 951 in the winter on nice days... previous owner(s) did not take great care of this car and raced it at one point...

my 951 has been out of commission because i dropped the funny looking bolt into the inspection hole (note: cover that hole) and am a few hours away from buttoning up the intake to test the motor.

i had no start issue also - no spark and replaced the speed and reference sensors (gapped using the glue a .8mm washer to the end of the new sensor then adjusted the bracket accordingly - removed the sensor, removed the glued washer and replaced) and expected to be a one hour job that has turned into a 20 hour fix... at least i'm getting to know the 951 pretty well...

you are sooooo close. look forward to the solution!

PerryB

Thanks to anyone that is still following and thanks to Flagtie and Alex Sol for your empathy.

I have retested the speed and reference sensors with my oscilloscope and they check out, just as they have ever since I first tested them.

I tested the signal from the DME to the KLR and it again failed this test. If Specialized ECU is right and the DME is working correctly, then there must be some reason why the DME does not fire the coil. What that reason is, I have no idea, and if anyone readng this does, I want they thoughts on this problem.

Perry