no spark

[armandodiaz]

I wish it was that easy. I replaced the DME relay 2 months ago in a poor idle diagnoses. I also ruled out the DME relay when I jumped the connector bypassing the relay. It still would not start.

Here's an update on what I spent my day doing:

1. Bypass DME relay, heard the fuel pump run
2. Bypass the alarm module (not too sure about the bypass I've tried bypassing the module a few months ago and never got it to run without it. This time I jumped +12V that goes from the Alarm Mod to the DME). Also went through the alarm wiring diagnostics.
3. Disconnected battery unplugged and replugged computer.
4. Sanded down all the ground points in luggage and engine compartment.
5. Checked for spark at the spark plug and at the coil wire (no spark)
6. Checked fuel pressure; didn't have a gauge but the fuel rails got pressurized every time I cranked the engine.
7. Check voltage at coil; found power at both wires with ignition on.

These are the possible problem areas:
1. Coil
2. Reference and/or speed sensor
3. Alarm module
4. Computer

Unfortunately these are all parts that can not be tested and I see no way of point it down to only one. I guess the only way to fix this one is to replace parts one by one (cheapest one first). Hopefully I'll get lucky.

[xlr8]

I also had a no start problem recently and fixed it by replacing my nearly new DME relay.

Another interesting link

Good Luck!

[robhamster]

I have an '84 cab and I recently went through the no spark routine. I narrowed it down to the DME computer and tried to find a shop that had a machine that they could plug it into. The two dealerships and three Porsche specialty shops all said no such machine existed and then one night, after my third glass of wine, I was sitting in the garage looking at my crippled little car and realized that Porsche did make a machine that one could plug a DME into ........ another 911. I found a shop that had a proper year "testing machine" around and took the DME there. For $20, they told me it was bad. I checked new and aftermarket and then consulted the experts here who suggested that before I send my old DME off to take it apart and look for bad solder joints. I used a magnifying glass and sure enough I found a bad solder joint. I fixed it with an iron and some solder and was off and running for about $0.12, including electricity for the soldering iron (oops, I forgot the three glasses of wine, add $8.00).

Before you spend any money on a DME, I would suggest taking your apart just to check for bad joints.


Good luck. I worked on this problem for five weeks with my car. I know how frustrated you must be.

[Lorenfb]

Check the three corner pins (of the four) of the DME connector for +12 volts.
Check the speed sensor (8 & 27) for 1.0 volts A.C. and the ref sensor (25 & 26)
for .10 volts A.C. while cranking.

If the injectors are working but no spark the DME is bad assuming no mickey
mouse alarm affecting the spark. The Porsche alarm affects the fuel and not
the spark. Use a test light to check for pulsing at the coil, the minus terminal.

[armandodiaz]

Thanks Loren and everyone else.
I will run those checks tonight after work.

[armandodiaz]

Originally posted by robhamster
I have an '84 cab and I recently went through the no spark routine. I narrowed it down to the DME computer and tried to find a shop that had a machine that they could plug it into. The two dealerships and three Porsche specialty shops all said no such machine existed and then one night, after my third glass of wine, I was sitting in the garage looking at my crippled little car and realized that Porsche did make a machine that one could plug a DME into ........ another 911. I found a shop that had a proper year "testing machine" around and took the DME there. For $20, they told me it was bad. I checked new and aftermarket and then consulted the experts here who suggested that before I send my old DME off to take it apart and look for bad solder joints. I used a magnifying glass and sure enough I found a bad solder joint. I fixed it with an iron and some solder and was off and running for about $0.12, including electricity for the soldering iron (oops, I forgot the three glasses of wine, add $8.00).

Before you spend any money on a DME, I would suggest taking your apart just to check for bad joints.


Good luck. I worked on this problem for five weeks with my car. I know how frustrated you must be.

Where did you find the bad solder joints? Was it on the board itself or was it a joint connecting the board to the plugin adaptor?

[robhamster]

You have to take the top off by prying the metal tabs aside and taking out the screws on the bottom that hold the circuit boards in. I found the bad solder on point at which one of the pins that attach to the DME harness connects to the circuit board. There are a bunch of curved connectors that go from the insert for the harness into and on to the ciruit board. I would start there. Good luck

[Early_S_Man]

Armando,

The bad solder joint in the DME affecting the coil signal is usually at the power transistor along one of the board edges. Be sure that you have cleaned both coil connections!

The coil can be tested independently from the DME by unscrewing the negative lead to the DME ... use a test jumper lead to ground, and a new, grounded spark plug wuth a 0.040" gap as a load. Each time you touch the jumper to ground and release ... you should get a healthy spark!

[armandodiaz]

Update:
I followed Loren's advice and checked voltage at the computer plug for the ref and speed sensor. Just to be sure I also jumped them with an LED. It was tough to get any kind of voltage reading from any one of the sensors, but I did get a faint light from the speed sensor and no light from the ref sensor. I'm just wondering wether the .10 volts would be enough to light the LED.
I took the computer apart and one of the first things I noticed was lots of corrosion on the backside of the circut board. Both myself and my wife looked extensively for any cracked joints. Neither one of us could find anything. I took a small wire bush and cleaned the corrosion and blew some air to clean it off.
I hooked up the computer back to the car and with the manual checked some of the computer outputs properties and made sure power and ground were going to where they sould be.
For now I'm ruling out the computer and I'm ruling out the Alarm mod. I'm focusing my attension on the ref sensor 1st and the coil 2nd. I'll problably change both ref and speed sensors because I'm already there and to give me peace of mind. If that doesn't work, I at least got the toughest ones out of the way.

[Dan Owens]

Test your coil first as Warren suggested. It's simple and quick and doesn't take a lot of brain power. I'd also look at the speed and reference sensor connector plugs and the ground from the body to the engine manifold.

[armandodiaz]

Originally posted by Early_S_Man
Armando,

The bad solder joint in the DME affecting the coil signal is usually at the power transistor along one of the board edges. Be sure that you have cleaned both coil connections!

The coil can be tested independently from the DME by unscrewing the negative lead to the DME ... use a test jumper lead to ground, and a new, grounded spark plug wuth a 0.040" gap as a load. Each time you touch the jumper to ground and release ... you should get a healthy spark!

On the coil test, that sounds like a good test but wouldn't I have to disconnect the neg side and breifly (very) jump it to +12V? The gound is being supplied by the spark plug, there is a constant +12V on the positive side.
This is one of the things that is baffling me. I do have a constant +12v on the positve side, +12v on the neg side while cranking and of course ground at the spark plug. Everything that is necessary for a spark, but still nothing.
I also don't seem have any voltage form the ref sensor. If the ref sensor was bad, would there still be +12V at the neg side of the coil? Maybe I should just buy all 3 things (coil, ref and speed sensors) and get it over with.
Just trying to figure out the logistics of what controls what has me . My left eye is already twiching and I'm going to have to find a shrink after this.

[Lorenfb]

You should not spend money until you find the problem!
Also, don't open the DME unit until you're sure that it's the problem,
you may introduce another problem now or later. You got bad advice
about taking it apart! You wouldn't start taking your engine apart,
if the car didn't move until you found the problem.

The system is very simple to troubleshoot.

Did you do this:

Check the three corner pins (of the four) of the DME connector for +12 volts.
Check the speed sensor (8 & 27) for 1.0 volts A.C. and the ref sensor (25 & 26)
for .10 volts A.C. while cranking (no LED lights, use a multimeter).

If the injectors are working but no spark the DME is bad assuming no mickey
mouse alarm affecting the spark. The Porsche alarm affects the fuel and not
the spark. Use a test light to check for pulsing at the coil, the minus terminal.

As I said, use a test light (like an icepick with incandescent bulb) to check
for pulsing at the coil. If the coil is bad but the DME is O.K., you should
have pulsing at the coil (unless it's shorted - very rare).

You can use a small LED tester to check for injector pulses. If the injectors
are O.K., but no spark, then the DME is bad. Check out this web site
(www.systemsc.com) on the Diagnostics page under "No Start".

Those tests will localize your problem.

[armandodiaz]

Loren I used a digi meter checked for AC volt and also jumped it with the LED. I read in an old thread that even with a meter its tough to notice the voltage change. That's why I also hooked the LED. Tonight I will try to hookup an Analog meter I have. The old threads say its easier to see current moving with the needle.
I didn't metion it but yes, there's +12V on three corners and grd on the fourth.
I honestly didn't specifically look at the injectors. I'm assuming that if the fuel rails are getting presurized everytime I crank the car the injectors must be firing. I unplugged the fuel rail about three different times and each time there was pressure and fuel.
Using the test light is how I found out there was +12V on the neg side of the coil when I cranked it.
This is why I'm trying to decide weather its the coil or the ref sensor. In one hand I wasn't able to get a reading from the ref sensor (sounds like a bad sensor, right?). But in the other hand I'm getting constant +12V on the pos side of the coil and cranking +12V on the neg side (sounds like a bad coil). So which is it? I would think if the ref sensor was bad then I souldn't get anything on the neg side of the coil, don't you think?

[Lorenfb]

Fuel pressure is NOT enough. The injectors MUST pulse.

You'll have +12 at the negative side of the coil until the DME starts pulsing
the coil. The DME needs both the ref & speed sensors. Some think that
one is for spark and the other for fuel, NOT true. The ref determines TDC
and the speed determines engine RPM and relative crank angle position.
Both are needed for spark and injection.

You must have your meter set to A.C. volts! The web site posted should
provide values, i.e. ohms & volts of the sensors.

[rick-l]

Originally posted by armandodiaz
This is why I'm trying to decide weather its the coil or the ref sensor. In one hand I wasn't able to get a reading from the ref sensor (sounds like a bad sensor, right?).

Those sensors are like mini permanent magnet altornators. They output a small AC (altenating current) waveform. The speed sensor puts out energy after every tooth on the ring gear goes past it. The reference sensor only once per rev so it wil be harder to see the ref sensor output. Did you say one will light a light bulb I would assume that one is good.

[But in the other hand I'm getting constant +12V on the pos side of the coil and cranking +12V on the neg side (sounds like a bad coil). So which is it? I would think if the ref sensor was bad then I souldn't get anything on the neg side of the coil, don't you think?

The DME is a constant current source to the igniition coil. When the DME is sitting there doing nothing it sinks current from the coil (I think 0.5 amps??). When the DME wants a spark it suddenly interrupts that current and the built up magnetic energy in the coil breaks down the air in the plug gap.

[rick-l]

Rethinking this coil problem
Since v = L di/dt the voltage across the coil would be zero or you would see 12 volts on both sides of the coil if;
1) DME lead open or power transistor bad.
2) DME sinking current properly but no spark pulses.
Does your meter have an AC current setting?

The other thing that just occured to me is maybe the DME does not sink current until it senses crankshaft rotation (more software).

[Lorenfb]

The output stage of the DME is NOT a current source. It's a switch with a current
limit of about 8 amps. The coil has very low resistance, less than an ohm (.5-.7).
The DME does not switch "on" until the proper time to set the dwell, i.e develop
the current.

Bottomline: The negative side of the coil will be +12 until pulsed by the DME.

[armandodiaz]

I'm using the AC setting on the multi meter. But like you pointed out, since the ref sensor is only creating .10V per rev, I believe this is too little voltage and the engine is turning too fast for the meter to pick it up. The light I got was from the speed sensor which has a stronger signal more frequent pickup (I'm satisfied with those results). On the neg side of the coil I do get a pulsating light when the engine is cranking. I will test the injectors tonight with a test light.

[armandodiaz]

Loren
On my eariler diagnostics the neg side of the coil had no grd and no pwr until cranking occured then there was pulsating a +12.

[Lorenfb]

The .10 volts A.C. is the RMS voltage & not the peak which is over 3 volts
for the ref, if you viewed it on a scope.

So, you do have pulses on the coil while cranking, which indicate that both
the ref & speed sensors are O.K. If the flashes on the test light are bright
at the coil & still no spark then the coil is bad. The spark should jump 25mm.

If you do have spark at the coil, next you must check for injector pulses.