PeteL

You need to double check electrical connections. I had a no-start that kicked my butt, until on a whim, I pulled the boot back from the connector for the DME temp sensor. About a half-inch of wire was exposed from just behind the connector and touching...shorting out. Have not replaced the connector yet, but have ensured the wires no longer touch, and the car starts. Sometimes it is stupid stuff like that, and it takes forever to find, and perhaps parts needlessly replaced.

PerryB

Mr. Laches,

Thanks for the sympathy and support. I will check the DME temp sensor connector right now. If it turns out to be that I owe you a six-pack of whatever you want!

I have been very busy with work, etc., and will not be able to devote a large block of my time and mental energy to the car until this Friday.

If you think of anything else, anything that I might have done during the clutch job please tell me.

Don't worry about asking if I've check the basic stuff--could be that I've missed something obvious.

Perry

flagtie

I question those DME's. The best way to test the DMEs without potentially sacrificing another one would be to send both of yours to another 951 guy. Have him try your DME's in his car.

konakat

Good call. A short or an open on the temp sensor can cause a no-start.

PerryB

Thank you to everyone that has responded.

I have re-tested everything.

The speed and reference sensors are putting out the correct voltage patterns, as desribed in the DME/KLR manual. I am fairly confident that my oscilloscope displays are accurate with these tests.

The power to the DME is correct as measured at both pins #18 and #35 and with ground hooked to pin #5. Both pins have battery voltage during cranking and with the ignition switch in the run position.

There is continuity between the DME connector and the KLR connector as measured through pins #21DME and #9KLR, #31DME and #24, #32DME and #16KLR.

There is continuity between the #1 pin on the DME connector and the coil negative side, the positive side gets battery voltage with the key in the run position and when cranking. All the ground connections on the DME #23, 5, 19, 10, 28, 17, 16, are good. DME pin #4 (start signal) gets battery voltage when cranking. DME connections #22, 9, 7, 6 to the AFM are good. DME coolant sensor pin #13 has 3,500 ohms (correct for temps between 15C and 30C--the temp in my garage). There is continuity between the DME pins #14, 15 and the injectors.

Now, here is something that did not pass testing. On page 30 of the DME/KLR test manual, Test Points 14 and 15, the DME is supposed to put out a square wave signal to the KLR. I tested this by hooking up the harness to the DME, then connecting KLR harness pin #9 to the positive terminal on the oscilloscope and the negative on the oscilloscope to chassis ground. Turning the key on, I got the flat line on the oscilloscope to move up (it remained flat) on the screen. When cranking, the line stayed flat. Also, when connecting KLR pin #24 to the positive on the oscilloscope and the oscilloscope negative to chassis ground, I got no change to the display--the manual calls for a spike in the display. I am less confident in my set up of the oscilloscope for these two tests.

It seems that the DME is getting all the correct imputs--power, ground, voltage waves from the speed and reference sensors--and has continuity to the coil, injectors--but isn't putting out the correct signals to fire the coil or the injectors.

What those two test tell me is the the DME is not working even though I just put in a rebuilt one.

Any one want put my DME and KLR into their 951 to test the computers for me?

Perry

specsalot

You've read my thread. My car presented tach bounce, fuel injection, but weak spark (and eventually no spark). I fought sending back my electronics as long as I could. If you do send them off, send off both units.

PerryB

Specsalot,

I thought that our virtual paths might cross!

Yes, I did closely read your thread, and, although your symptoms were different from mine, I have realized that I may end up solving my problem by sending both computers out for testing/repair.

Since I'm not a big believer in coincidence I have to ask why doing the clutch job caused $700 in damages to my computers. Obviously I want to know to satisfy my curiosity, but also to make sure that I don't do it (whatever "it" is) again. I'm not a member of the sh*t happens club. Usually I (or someone else) cause(s) sh*t to happen and before I spend more of my hard-earned money I need some understanding of the cause, progression, and solution of this problem.

I don't want to sound ungreatful for your input and the help of the others on the forum. What has happened in the last two months is that I repeatedly think I have the solution, spend money on parts, spend time on installing them, then, and this is the worst part, get optimistic that this latest "fix" will get the car running, only to be disappointed.

So my great worry is that I will send out the computers and they will either test out as okay (meaning that I'm still completely stumped) or that they will need to be rebuilt (meaning that I am out another $350, and the car STILL might not start).

What I was really hoping for was that someone would suggest something that I had overlooked that would cause this problem.

I will keep posting to let everyone following this thread know what happens.

Again, thanks to everyone who sent me their thoughts and suggestions and support.

Perry

specsalot

I fully understand. I'm not a fan of unjustified spending. I had a "no start" after doing my clutch in 2010. I "guessed" it was the speed reference sensor and replaced it. Like you, I slotted the bell housing for easier assembly. I set the gap using feelers with the bell housing out of place. After the fact, the issue was just poor connection at the speed reference sensor plug on the DME harness. [I know because I later swapped back the sensor I though was "bad" and the car ran fine.] During that "no start" (in 2010) I had plenty of spark, but no fuel injection.

There's nothing about doing a clutch that should cause a computer problem. Here is an excerpt from the manual describing safety rules with respect to the electronics:



As long as you feel like you haven't crossed any of the above lines (safety rules), it is reasonable to believe there that your work shouldn't have caused computer issues. The crank sensors generate their own AC voltages. When I tested mine using a VOM (per specialized video) they measured ~ 2.2 VAC on the speed reference sensor and ~ 0.2-0.4 VAC on the reference mark sensor. A VOM measures an "average" RMS voltage. I played around with the gap a bit and was able to get the speed reference up to ~ 5.5 VAC, but in the end I reset the gap (0.8 mm washer glued to a spare sensor) to factory clearance. Here are my "silly" questions as I'm sure you've probably already worked through these items:

Question 1. I'm not sure how you can measure 12V P-P out of either flywheel sensor?

There is usually a spacer washer which sits under rear most sensor. It is show as ITEM 32 in the drawing below:



The PET catalog lists two different P/N's for this washer



There are also two different P/N's for the support bracket (ITEM 12 in the diagram above)



I assume the thicker washer goes with the 944 101 321 07 bracket.

Question 2. Any chance you've placed the spacer washer under the wrong sensor?

FWIW, I have a bracket that requires the 2mm washer. So I guess it's a newer style bracket. The relative registration of the two sensors is based having the correct washer under the correct sensor. Also the sensors need to be bottomed in the support bracket. The typical process is to set the gap of the speed reference sensor. This insures that the reference mark sensor is registered correctly as well. If you computer isn't getting valid signals from both it will not operate. Setting the gap with everything together is a PITA. Since you've moved yours (after you set it with feelers), its probably worth setting again by gluing a washer on a spare sensor and resetting the gap. Your voltages should look something like mine using a VOM to check them. No tach bounce pretty much suggest your electronic may not be getting proper signals.

The sleeve (ITEM 31 above) should be sitting more or less flush with the raised surface of the support bracket.

specsalot

Why my car "died" this past November is something I'm still not certain about. After looking over things, I had 2 of 4 plug wires with very poor attachment to the plug tops (retaining clips rusted / missing). The plugs tops showed lots of tracking. This probably caused the initial issue with the electronics. Poor plug connections is kind of like violating safety rule #12 above.

Hang in there - You will figure it out.

PerryB

Specsalot,

Wow, those replies look like they took a lot of time to compose. Thank you for helping me out. I also notice that you posted both in the wee hours of the morning. I hope that you work the graveyard shift and didn't stay up late just so you could lessen my ignorance!

I have not seen that list of forbidden electrical procedures. I don't have a specific memory of violating any of them, with the possible exception of number 4. I may have tried to start the car with my battery charger still attached and running. Again, I don't specifically remember doing that, but it is possible.

As for your specific questions...

2. The sensor mounting is correct in all aspects. I have installed the figure "8" washer under the speed sensor. While both sensors are new, all the mounting parts are the same and installed in the same positions as when I parked the car last December. They are correctly gapped using the procedure on the Clark's Garage website.

1. My oscilloscope measures peak to peak voltage in a sine wave pattern. After calibration, I set the scope so that from the lowest spot on the screen to the highest is 40 volts. During cranking, the sine wave goes about 6 volts above the central horizontal line (the x-axis) to about 6 volts below. Again, this is my eyeball measurement, it may be a bit more or a bit less, but it is definitely greater that the minimum of 2.5 volts as required by the DME/KLR test plan. If I use the next lower calibration scale, the top of the screen to the bottom is 4 volts. Testing using this scale creates a pattern that clearly maxes out the screen space. The same thing happens when testing the reference sensor, although the display is a single spike that occurs once each time the flywheel rotates, rather than a sine wave pattern. Both displays are exactly what the DME/KLR test manual shows should happen during testing and I did both tests at the DME connection.

Given this test and the fact that both sensors are new, pass the ohm test, use a brand new harness to connect to the DME, are gapped correctly, and show the correct displays on my oscilloscope I really have to conclude that they are not the problem.

Unless there is some test that I have not yet performed, or there is something that you (or anyone else reading this thread) can think of, I will most likely contact Specialized ECU Repair about testing both their rebuilt unit and my KLR next week.

Again, thanks for your help and support. I will greatly appreciate any other thoughts about this problem.

Perry

specsalot

Perry,

I work rotating shift and was on night shift last night.

Sounds like you've run this down fairly well. I did a year of "vocational electronics" in HS in the early 70's and used o-scope's on a daily basis. I believe your numbers. If you look at the DME schematics (available on Clarks Garage), you will see that the two signals are fed to a single IC package which does some kind of A/D conversions. Here is a related thread (from pelican) with some interesting links in it.

http://forums.pelicanparts.com/porsc...tly-stuck.html

Your Motronic DME should be the following part numbers:

Bosch: 0 261 200 075
Porsche: 951.618.121.08

This is Specialized core business, but its possible they may have sent you the wrong replacement (not likely).

When my DME failed the first time, damage was listed as cracked solder joints, a failed memory module and a failed voltage comparator. Second repair was transistors and capacitors which drive (ground) the coil. Electrical problems are difficult to run down, thats for sure. The tech assured me that a failure in the KLR is what smoked the DME a second time.

I'll be sending you a PM.

specsalot

Looking at the wiring diagrams for an 87, your tach is fed from DME PLUG PIN 21 via T1 which is located in the passenger foot well. The interconnection between Motronic DME PLUG PIN 21 iand KLR PLUG PIN 9 is fed via a shielded cable. The shielding connects to KLR PLUG PIN 10. Disconnect your DME / KLR plugs and see what kind of resistance you measure between KLR PLUG PINS 9 and 10. Perhaps your shield has been compromised grounding the KLR signal. Another possibility is that the line to the tachometer may be grounded. Try reading KLR PLUG PIN 9 to ground and see what you get. If it shows grounding disconnect T1 in the Passenger Footwell and see if it clears. If so, plug everything back up and try a start.

According to the schematics there is a Knock Control Diagnosis Plug (T26) also located in the passenger footwell. With everything plugged in, you can measure the ignition signal against ground directly off PIN 3 (1.0 GN/YE) of this plug. You've already tested via the test plan. Here's another approach to the same point. Potential issues fall into one of 3 areas:

1. Bad inputs into electronics
2. DME harness issues
3. DME/KLR electronics failure.

Having a PhD in the obvious (above) and 50 cents buys a cheap cup of coffee. I didn't go this deep level on my car. I did the best diagnostic I could on #1 and #2 and ultimately settled on #3 above. Specialized are electronics experts but not so much so on individual vehicle harness wiring. Beyond what happens on the circuit boards themselves, Specialized typically don't volunteer much useful info. This is where the TEST PLAN and experienced folks here carry the day.

specsalot

FWIW - Here is T26 schematically. On the 88 diagrams its labeled Knock Control Diagnosis Plug T26. On the 87 diagrams is only labeled "Test Connector". On 87 terminal 3 is 1.0 GN/GE on 88 its 1.0 GN/YE. In any case - same same. Good Luck

PerryB

Specsalot,

Thanks you all your work in helping me get to the bottom of this problem.

I will send you a PM later after doing a few more checks.

Perry

specsalot

The Lindsey Reference Sensor Harness Kit (LR-944-HAR-REF) I installed came with two different instruction sets because of the way they designed the kit. The difference in installation is based on whether the stock O2 sensor is being retained or not.

The schematics show that the oxygen sensor and the speed sensor share a common shield ground which lands on DME PIN 23. Try unplugging your oxygen sensor. Perhaps there is a problem with the new sensor or its ground shield.

Keep fighting the fight. You will figure this out