PerryB

Yeah, the 944online sensor harness has the same modification. I have tried to start with the O2 sensor connected and disconnected. No difference.

On the 944online harness, pin #23 is de-activated. All the sensor grounding is done through pin #5. Looking at the DME internal wiring, all those ground pins are connected inside the DME.

Perry

PerryB

Specsalot,

I have sent you a PM.

Perry

PerryB

Freinds,

For anyone still following this thread...

Yesterday was a very strange day. I have continued to try to diagnose my car's no start condition. During my poking around in the engine compartment, I noticed that there was a vacuum line that had become unhooked. This is a line that takes off from a port underneath the throttle assembly and goes between the cylinder head and the intake manifold to a thermoswitch at the back of the head on a pipe that leads to the heater valve. From this switch, the vacuum line then goes to these two vacuum pods mounted on the driver's side inner fender, near the firewall.

I didn't think that this would be the source of my problem but knowing that a vacuum leak could cause a no start, I plugged it back in, re-hooked up the DME and KLR, and went to see if the car would start. What happened next completely shocked me. The car cranked over normally, but then "kicked back" like a car that has over-advanced timing. I could smell fuel. I pulled out a spark plug and it was slightly damp with fuel. I re-installed the spark plug and tried to start the car again. Again, I got a few "pops" as it seemed that the car was getting both spark and fuel, but seemed to be miss-timed.

I then noticed that during cranking the tach seemed to bounce--not like normally as a result of the reference sensor firing--but erratically. Additionally, there seemed to be a weird, crackling noise from the engine that corresponded to the tach needle bounce. After some investigation I narrowed the noise down to the injectors firing and the idle air control valve rapidly openning and closing! This would only occur when I initially turned the key to the "On" position, and would only last for about 5 seconds, but happened every time I turned the key.

After some more investigation, I was able to recreated this condition by applying power directly to the Red/Blue wire at pin #2 of the 14 pin connector mouted near the brake booster. It would only happen when I would just begin to touch the jumper from the + battery terminal to the male end of the bullet connector at this spot. Once I had made solid contact, the tach bounce, rapid injector firing, and the ICV cycling on/off would stop.

According to the wiring diagram I have, this wire supplies power to the injectors, then the injectors fire by being grounded throught the DME.

What is going on here?? I don't think that the vacuum line had anything to do with this. In fact, I purposely unhooked the line and the condition stayed the same.

Is it possible that I have a shorted cell inside my battery? I have tested the battery and it puts out the correct CCA, has about 12.6 volts after charging, and delivers about 11 volts to the DME during cranking.

Perry

specsalot

Perry - Great Progress - You are definately on to something with this find.

CONNECTION T21 TERMINAL 2 may have poor contact in the socket. The male pins are split and can be spread slightly to improve contact pressure in their female recievers. Give all contact points a good cleaning if you haven't done so. There may also be a ground in the DME harness connected to this terminal.

As you point out CONNECTION T21 TERMINAL 2 supplies the common side of your injectors, your timing valve and your KLR Unit (at KLR TERMINAL 6). So not having B+ where it belongs (bad connection or grounding) may be your issue. Many of the DME Harness Jetronic Connectors experience issues (broken wires / insulation failure) inside the boots just beyond the plug connectors). You might want to pull back some boots and see whats happening. During my no start in 2010, I found insulation failure and weak wires at all my injector plug connections. Noises that sound like "crackle" always make me think of arcing / sparking / grounding. If you're going to test the harness for gounds be sure to unplug your electonics first.

The vacuum line you reference is part of a system that captures fuel vapor from your fuel tank and disposes of it. Good catch. It all helps.

Nothing in your battery comments suggest you have a battery issue. I got similar values with my battery which is less than 6 months old.

Good Find - Good Luck - You're almost there

- Paul - aka "specs"

Van

When you're measuring battery voltage, are you touching the terminals of the battery, or the cable connectors?

My dad used to have a jaguar, and that sucker would only start if the battery posts were *clean*! Otherwise, the slightest little bit of corrosion between the post and the cable would reduce the current and cause a no-start condition.

specsalot

Van - Got to Love those British Cars. Had an MG for years - Thats how I learned to love wiring issues

Perry - Here is what my diagram shows:



CONNECTION T21 TERMINAL shows up in this picture as the rectangular box labeled F49. This is B+ if all is OK. Box 3 is the weld point supplying B+ to the common side of your injectors and timing valve. Rectangle B48 is B+ to your Idle Speed Positioner (valve). Sounds like B+ is not arriving intact on a consistant basis.

Proper voltage between KLR pins 6 and 14 corresponds to SHOP MANUAL TEST POINT 2 (page 28-19). The shop manual suggests a problem with the DME relay. Try bridging DME RELAY TERMINALS 30, 87 and 87b. See if this stabilizes voltages and produces a start.

Most likely you have an issue with wiring rather than components, but the DME Relay is a known issue. I think you replaced your DME relay, but perhaps there are connection issues between the relay blades and the socket they connect to in central electronics. Doing a 3 way jumper takes questions about the DME relay itself off the table for testing purposes. Funny electrical sounds for 5 seconds (after key on) suggest intermittant DME relay operation.

When I was working my recent "no start" after every start attempt, the turbo water pump would run for ~30 seconds after key off. [This is what it's supposed to do]. In the past when you were trying to start, do you remember your turbo water pump running after cranking the engine? Is it operating now (~30 second) after crank attempts? My turbo water pump is loud enough that I can hear it run against the defening silence of a no start (windows down - hood up) You can also see the coolant swirl in the coolant tank from the turbo cooling return when the pump is on.

PerryB

Vans, Specsalot,

Thanks for your imput. Thanks for sticking with me!

The battery suggestion was a long shot. Years ago I had a Bosch injected Alfa that cranked just fine, but would not start. Eventually, we found that it had a weak battery that passed most of the test we put to it, but failed to get the computer to fire the ignition or injectors. I thought that this might be the case here...

I have done a VERY thorough job cleaning the B+ and B- connections. I also thoroughly cleaned the ground connections on the back of the engine block and the top of the clutch housing when I did the clutch job. I also cleaned the additional positive connections at the positive battery terminal.

The "crackling" sound is the injectors firing rapidly--I'm sure of this--not actual electric arcing, also the ICV rapidly cycling on and off.

It seems like the injectors are getting the +12v that they need, but that the DME is grounding them erractically. When it does this it also gounds the ignition which sends the tach needle bouncing and occasionally fires a spark plug causing the pop. The power to these injectors comes through the DME relay, from the 87 pin. Whether the power flows through the DME relay, or by me jumping the 30 and 87 points, or by me supplying B+ power directly to the male connection at terminal 2 (with the male pin removed) at the 14 pin connector, the condition repeats itself.

Specsalot, if you could send me the entire page of the '87 Turbo wiring diagram I would greatly appreciate it. I can only see/download the upper left corner that appears in the post. If you can email it to me at my email address I would greatly appreciate it.

The turbo cooling pump does run for about 30 seconds after cranking, as it did back when the car was running.

When I spliced in the new sensor harness all the connections at the DME looked clean and good--no bent pins, loose connections, frayed wires, cracked insulation, etc. This particular car is very clean and unmolested. It looks as if the computers had never been removed prior to this. There was no evedence of the battery acid leaking through to this compartment. Also the car was originally from LA and I bought it from a guy in Denver, so I don't think that the car has seen much moisture. There was no corrosion on the electrical connections anywhere on the car.

I just returned from the garage. When I turn the key to the run position and bridge the 87 and 30 ports on the DME relay female plugs, the injectors rapidly cycle, and the coil fires, again all erractically and only when the bridge is initially made.

With the key in the run position and the male pin at terminal 2 from the 14 pin plug is unplugged, I can get the coil to fire every time I BREAK the connection between B+ and the male pin.

I am completely baffled by this. Again, it seems as if the DME starts to ground the injector circuit and the coil circuit when they get power. However, THE DME IS NOT GETTING ANY POWER FROM THE 87 PORT ON THE DME RELAY. Why is the DME grounding the coil and injectors under these conditions?

Perry

specsalot

I will e-mail you what I have for diagrams. I have no doubt the problems are as you describe. Having injectors fire and the coil randomly sparking without the engine cranking is a perplexing situation.

PerryB

Friends,

Thanks to anyone still following this thread.

I have done another round of testing that have yielded bizarre results. I have removed the KLR computer and tested the power between terminals 6 and 14 on the KLR harness. They have battery voltage with the key on and when cranking. I have tried to start the car with terminals 9 and 16 on the KLR harness bridged, without any success.

Now here's where things get strange. I again tried to do test point 15 form the DME/KLR test manual. I hooked up the oscilloscope to terminal 9 on the KLR harness and to ground. On the scope's screen a flat line appears. Then, as soon as I apply power to the DME by jumping the DME relay terminals 30 and 87 the line jumps up. I believe that this now indicates that the DME computer is grounding that circuit. As I rapidly insert and withdraw the jumper the line will rapidly alternate between its high and low positions on the scope's screen. While doing this the coil will begin to fire ON ITS OWN for about 5-10 seconds when I leave the DME energized. As that happens the scope shows a square-toothed pattern (as called for in the manual) and I can hear the coil rapidly firing. If I pull the coil wire off the distributor cap and stick a grounded spark plug into its free end, I can actually see the coil rapidly firing the spark plug during this 5-10 second interval. When this happens the tach need raises up about 2mm up (still below 1000rpm). The square tooth pattern is clearly the DME grounding and ungrounding the coil circuit. All this occurs without the engine cranking.

Wait, it gets better. Then a high pitched whine occurs, the pattern on the scope goes nuts, oscillating so rapidly that the pattern is just a blurr, and the tach needle pegs itself past 7000rpm! This contiues until I unplug the DME jumper. I then try to recreate this condition, but cannot. I can still get the square tooth pattern to appear by rapidly energizing/unenergizing the DME with the jumper.

Okay. From reading the wiring harness, I think that the DME recieves the analog sine wave pulse from the reference/speed sensors, converts it to a digital square tooth pulse, sends that from DME pin 21 to both the tach and to the KLR pin 9. The KLR then sends that signal to KLR pin 16 back to DME pin 32, through the DME to DME pin 1, then to the coil. This alternating square tooth signal is what grounds the coil, getting it to fire.

The DME does not seem to be either converting the analog signal to digital, or sending the signal to the KLR (or both). The DME only seems to convert and send the signal to the KLR randomly when I apply power to it in a rapid on/off manner. More importantly, the DME seems to do this WITHOUT ANY IMPUT FROM THE REFERENCE/SPEED sensors--all these symptoms occur without the engine cranking. In my mind all this seems to point to a bad DME.

Could the tach cause this problem? Could there be some kind of short in the tach that prevents the ignition signal from getting to KLR pin 9?

Given that the DME is brand new, is there something that could have damaged it during installation?

I will greatly appreciate any thoughts or suggestions for further testing.

Perry

PerryB

Oh yeah. One more symptom that may have nothing or everything to do with this. The electric window relay (G7) gets hot during this testing. The windows work normally.

Perry

flagtie

I'm still following. Shorting tachs, wire harness problems, etc. I'm just an NA guy but this all seems pretty crazy for just a 6 week clutch job. I'm sticking with my earlier idea. Send those boxes to someone to put in their car and test.

Whatever you do. Good luck. I'm rooting for you. I know what a no start is like. It turned out I had 2 bad DME's that's why I'm biased.

konakat

Wow, I don't even know what to say, except find someone else with a turbo and swap DMEs before you drive yourself insane. That is some crazy stuff.

Have you opened up the DME? It seems like you would be able to see any damage that might cause that kind of craziness.

PerryB

Thanks for staying with me Konakat, Flagtie.

Crazy, yes, an understatement.

No, I won't open the DME. Specialized ECU put a kind of tamper evident tape onit that if removed or torn, invalidates their warranty.

Besides, I have no idea what to look for or how to test the guts of the DME.

Flagtie, I keep trying to think what work I did with the clutch job that could have caused all this. You're right in saying that it's a crazy outcome just for letting the car sit for 6 weeks while I did all those repairs. One of the things that I did was to replace the windshield washer hose. Maybe that's what did it!

Stick with me and this thread, if only for the schadenfreude effect.

Perry

specsalot

Perry - I've inserted a few comments in red in your quote above. Squinting at the documentation and based on limited reading my VERY limited understanding is as follows:

The KLR receives:

1. Throttle postion input (terminals 21 through 24)
2. Trigger signal from DME 31 at KLR 24 (not sure what this does)
3. Ignition signal from DME 21 at KLR 9 (this also drives your tach)
4. Manifold Pressure (Vacuum) - physical input
5. Knock Sensor Input (KLR 11-13)

The KLR outputs:

1. Ignition Input KLR 16 to DME 32 (this signal probably incorporates shifting ignition timing based on sensing knock)
2. Full Load Output signal KLR 18 to DME 3
3. Signal to the timing valve.

The basic functions of the KLR are to:
1. Listen to the knock sensor and adjust ignition timing if it hears engine knock.
2. Informs DME when the driver is demanding Full Load (i.e. pedal to the medal)
3. Manage over boost situations via the timing valve.

The DME does the following:
1. Creates a crank shaft position info stream (from crank sensors)
2. Calculates required fuel and creates injector pulse by grounding injectors.
3. Sends a timing signal to KLR which is modified and returned to the DME creating ignition pulses by ground the coil via DME transistor networks.

You are describing a system that is being driven into self-excitation by toggling power to the computers and various sensors ON and OFF. It's very hard to draw conclusions about what may be wrong. What can be said at this point is that the core circuits in the DME that fire the coil are operational. You are also showing that fuel injection seems to be operating when this system is self-excited. This implies that core circuits in the DME which pulse your injectors are also working. The fact that the Motronic DME is not in normal process (i.e. cranking the engine) sending the square wave over to the KLR, implies that the DME may not be processing the crank sensor information.

We know you've read the inputs with an o-scope. How do those inputs measure (for comparison purposes) using a volt meter? IIRC, Specialized's video suggests reference mark sensor must be below 1 VAC (with values typically 0.2 volts). Speed reference must be at least 1.6 VAC (typical ~ 2 VAC). I would be nice for an apples to apples comparison (i.e. VAC RMS) of your crank sensor outputs to values identified in Specialized's video. Most likely they will match up well.

When I struggled with a my no-start in 2010 (after my clutch job), I had beautiful spark, but NO INJECTION. Turned out it was a poor connection with the DME not receiving the speed sensor signal. You remain at a point where you are getting neither spark nor injection when everything is plugged up normally and you crank the engine. The challenge is understanding what the things you report mean. Keep testing / posting results. We're all pulling for you to get this solved.

PerryB

Specsalot,

Thanks again for your thorough reply.

This weekend I made a point of going through the DME test plan one step at a time. Here are the results...

Test point #1--Grounds and connections. I have removed and inspected the ground connections at the firewall as well as the top of the engine block and clutch housing. All are tight and clean. I also removed the connections at the AFM, throttle switch, DME temp sensor, speed and reference sensors, the cycling valve, the grounds in the interior fender wells, the grounds under the dash, and the main harness connections to the DME and KLR. All clean and good.

Test point #2--Voltage to the DME and KLR. I have tested for both voltage DME pins 35 and 5, as well as 18 and 5. They have battery voltage with the key on, and when cranking. With the KLR pins 6 and 14 I get battery voltage with the key on and while cranking. The DME, KLR, and fuel pump all get power.

Test points #3 and #4--Speed and Reference Sensors. I have spent more time with these two sensors than anything else on the car. They are brand new, and so is the harness that connects them to the DME. All my tests have been through this harness at the DME connection. They each have about 880 ohms. During cranking the speed sensor develops about 2.5 AC volts. During cranking the oscilloscope shows a sine wave voltage of about 12 DC volts for this sensor.
During cranking the reference sensor develops about 0.12 AC volts. During cranking the reference sensor sends out a single repeating spike with an initial positive slope of about 12 DC volts. Every test shows that they both are working correctly and are sending that correct information to the DME.

Test point #5--Ignition System. All the HT ignition wires have the correct resistance, are free from cracks and shorts. The distributor cap is used, but shows no evidence of carbon tracking. The rotor is firmly attached to the camshaft. The coil has the correct primary and secondary resistance. The ignition switch sends 12 volt to the positive side of the coil with the key in the on and start positions.

Test point #6--AFM. All the wires from the DME to the AFM have continuity. The AFM gets 5 volts measured by removing the back cover of the DME wiring harness connector and probing the back of female pin connector 9 and ground. Probing the back of female pin 7 and ground I get 255 millivolts with the flap closed, and increasing voltage as I move the flap. Using an old-style needle volt meter, I then tested and found that the voltage increased smoothly up to about 4.5 volts with no dead spots. I also tested the air temp sensor in the AFM: 3,500 ohms--correct for the temp in my garage.

Test point #7--Fuel Pressure. Low 30's to mid 30's.

Test point #8--Idle Actuator. Not part of the test plan for a no start.

Test point #9--Throttle Switch. Ohm reading between pin 2 and ground is .8 ohms with the throttle closed that then goes Open as soon as the throttle is moved. I get 5 volts between plug terminals 1 and 2 on the connector. Using my needle voltmeter I get about (analog scale) 500 ohms at KLR terminals 22 and 23. As I open the throttle the resistance increases smoothly without breaks to a maximum of 4,500 ohms.

Test point #10--DME water temp sensor. Again, about 3,500 ohms--correct for the temp of the garage.

Test point #11--Injectors. Injectors will fire when powered and grounded. There is continuity between the injector connectors and terminals 14 and 15 on the DME.

Test point #12--Knock Sensor Blink Code. The car needs to be started and driven before the KLR stores error codes. As per the test plan's suggestion I have bypassed the KLR by bridging pins 9 and 16 with no positive results.

Test point #13--Vacuum/Pressure Sealing. I have checked the tightness of the hose clamps that secure all the main intake hoses. I have traced out all the smaller vacuum lines and find no problems. During cranking, the AFM flap moves as the engine the draws in air.

Test point #14--Altitude Compensator. Not part of the test plan for a no start.

Test point #15--Ignition Signal to KLR. Failed. Confer with my previous posts.

Test point #16--Oxygen Sensor. Again, not part of the test plan for a no start.

These results represent much time and learning on my part, but I am still completely stumped.

Specsalot, I have no idea where the connction in the passenger's foot well is for the tachometer. I have looked and not found anything. I don't think that the tach has anything to do with this problem but seemed a possibility as it as well as the KLR should get the ignition signal from DME pin 21.

All results point to a bad DME, especially test point 15. However, is there some other cause that would prevent the DME from sending the ignition signal to the KLR? All the wires that connect the DME to the KLR have continuity.

My next step?

Thanks for all the support and suggestions from everyone on the list, especially Specsalot. Without this encouragement I would have pushed the car off a cliff by now.

Perry