DME blues

[r lane]

Has anyone encountered a case of numerous DME failures on a single car. In my case, 87 Carrera, I believe something in the engine management system is causing damage to the DME. 2 flywheel sensors, 02, cylinder head, TPS, throttle microswitch, injectors, ignition. What am I forgetting? Anyone run across this. Thanks, Bob

[Fly911]

The DME relay does two things... It has two internal relays, one supplying 12V to the ECU and ignition coil, the other supplies 12V to the fuel pump. None of the engine sensors will have any influence on the DME relay. It's just like two on/off switches.

However, the DME relay have on a large number of cars caused problems, my own 1987 Carrera included. The relay is cheap, both on price and quality. I have permanently installed two DME relays under the driver's seat, so if one fails, I can swap the connector over to the other one.

The RPM and TDC flywheel sensors may cause problems, but are not DME related. They are also cheap, and can be purchased from our host here on Pelican.

All 911 Carreras are over 30 years old, and it's probably wise to change out all engine sensors, and fuel injectors. It's probably also wise to upgrade the ECU with a chip for modern fuel injectors. This will improve drivability, fuel economy and power significantly. Scarseller here on this forum has developed a new ECU chip, in combination with a MAF sensor and modern injectors, with surprisingly good results.

I went a different route with my Carrera, with a Canems ECU (still with the DME relay, though...) and individual throttle bodies from Jenvey (sold as a complete kit from Canems). The Canems ECU is a direct plug in to the ECU multi pin connector, and programmable to whatever you want. And with full engine and sensors diagnostics.

As for your DME, buy one or two spares!

[Quicksilver]

The sensors can't cause damage to the DME but a short can.

What did you do to the DMEs when they failed? A good repair shop should be able to tell you what failed which will tell you where to search on the car.

[GH85Carrera]

The standard answer to most all electrical weirdness is clean and repair all the grounds. That the grounds apart, clean the connections, and make sure they are all tight. It is a pain in the butt to go through and find them all, but over the years the metal connectors just get some surface corrosion from the moisture in the air. It is true for any car that is 30+ years old.

I used to clean just about any ground connection I came across in the general work on the car as a standard practice. A ground can't be to good, just good enough to always work.

[Rodsrsr]

Originally Posted by Quicksilver

The sensors can't cause damage to the DME but a short can.

What did you do to the DMEs when they failed? A good repair shop should be able to tell you what failed which will tell you where to search on the car.

Correct, this is why you always test a questionable DME in a know running car and not a confirmed working DME in a questionable car.

[famoroso]

Solid state...

911 Solid-State Relay — Focus 9 Technology, LLC

They also offer a pump-prime version that enables you to hear the fuel pump running prior to cranking the engine.

[ischmitz]

Interesting - the OP refers to the DME (the computer under the drivers' seat with a 35-pin connector). Some respond referring to the DME relay (a small black box the size of a matchbox). The DME relay has a bad reputation due to some quality issues and does fail from time to time. Carrying a spare is a good idea since it might help you out in a pinch when the alternative is a flatbed home. They cost somewhere around 25$ and you should have one in your glove box. The DME itself is pretty robust given it's age. Aside from compromised solder joints due to age and vibrational stress there aren't many failure modes unless something external was done to it. Other than that the 3.2 DME is really robust and rugged. Occasionally other components do fail but can easily be diagnosed and replaced with a good oscilloscope and some basic electronics skills. It would be helpful if the OP clarifies what exactly failed and what was done to repair the failure.

The discussion about whether to plug a known good DME into a car under test or use the DME under test in a known good car has not straightforward. Fact is that a short test either way will not damage anything. But there is no preferred way as either test bears risks:

- A prolonged exposure of a good DME to car with issues (shorted injector harness, shorted ignition coil, shorted DME relay input, or shorted ICV) will eventually lead to component failures inside the DME. The low-side drivers for the coil and injectors use current-limiting but are not long-term stable when outputs are shorted to VBat. However, a short test (turn ignition on, crank and check if it does start) is perfectly fine and safe.

- A prolonged attempt of trying to start a good car with a damaged DME can damage things in the car. If the fuel or ignition outputs of the DME failed short the DME loses its ability to control current. This can damage the ignition coil or the injectors and a flooded engine can hydro-lock with fuel dripping out of the exhaust pipe...

[mysocal911]

Originally Posted by ischmitz

Interesting - the OP refers to the DME (the computer under the drivers' seat with a 35-pin connector). Some respond referring to the DME relay (a small black box the size of a matchbox). The DME relay has a bad reputation due to some quality issues and does fail from time to time. Carrying a spare is a good idea since it might help you out in a pinch when the alternative is a flatbed home. They cost somewhere around 25$ and you should have one in your glove box. The DME itself is pretty robust given it's age. Aside from compromised solder joints due to age and vibrational stress there aren't many failure modes unless something external was done to it. Other than that the 3.2 DME is really robust and rugged. Occasionally other components do fail but can easily be diagnosed and replaced with a good oscilloscope and some basic electronics skills. It would be helpful if the OP clarifies what exactly failed and what was done to repair the failure.

The discussion about whether to plug a known good DME into a car under test or use the DME under test in a known good car has not straightforward. Fact is that a short test either way will not damage anything. But there is no preferred way as either test bears risks:

- A prolonged exposure of a good DME to car with issues (shorted injector harness, shorted ignition coil, shorted DME relay input, or shorted ICV) will eventually lead to component failures inside the DME. The low-side drivers for the coil and injectors use current-limiting but are not long-term stable when outputs are shorted to VBat. However, a short test (turn ignition on, crank and check if it does start) is perfectly fine and safe.

- A prolonged attempt of trying to start a good car with a damaged DME can damage things in the car. If the fuel or ignition outputs of the DME failed short the DME loses its ability to control current. This can damage the ignition coil or the injectors and a flooded engine can hydro-lock with fuel dripping out of the exhaust pipe...

and an over-voltaging alternator.

[darrin]

Originally Posted by mysocal911

and an over-voltaging alternator.

GOOD call --

OP -- have you ruled this out? A failing voltage regulator can cause transient voltage spikes -- The symptoms I observed (but didn't know related to the issue when it happened to me many years ago) included "dimming" headlights and dash lights (the "dimming" was actually normal voltage after the very bright light caused by high voltage had dissipated), random seat belt warning light illumination and (since I was so VERY clueless back then) sulphur smell and dampness around the battery (caused by it boiling over due to high voltage). The high voltage spikes compromised some of the soldier joints in my DME computer, which a knowledgable person fixed for me by going over the circuits and reflowing the failed soldier joints.

[jpnovak]

What is failing in the DME? Have you had one looked at to determine? Is it injector circuit, ignition circuit, input circuits, etc?

[mysocal911]

Originally Posted by darrin

GOOD call --

OP -- have you ruled this out? A failing voltage regulator can cause transient voltage spikes -- The symptoms I observed (but didn't know related to the issue when it happened to me many years ago) included "dimming" headlights and dash lights (the "dimming" was actually normal voltage after the very bright light caused by high voltage had dissipated), random seat belt warning light illumination and (since I was so VERY clueless back then) sulphur smell and dampness around the battery (caused by it boiling over due to high voltage). The high voltage spikes compromised some of the soldier joints in my DME computer, which a knowledgable person fixed for me by going over the circuits and reflowing the failed soldier joints.

Most of the time that destroys the injector driver transistor, and can also cause the injector I.C. to fail.

[ischmitz]

Originally Posted by jpnovak

What is failing in the DME? Have you had one looked at to determine? Is it injector circuit, ignition circuit, input circuits, etc?

Do we even know if it's the DME or the DME relay the OP is referring to?

[kach22i]

Originally Posted by GH85Carrera

The standard answer to most all electrical weirdness is clean and repair all the grounds. That the grounds apart, clean the connections, and make sure they are all tight. It is a pain in the butt to go through and find them all, but over the years the metal connectors just get some surface corrosion from the moisture in the air. It is true for any car that is 30+ years old.

I used to clean just about any ground connection I came across in the general work on the car as a standard practice. A ground can't be to good, just good enough to always work.

Good suggestion about a bad ground.

I try to start with the simple basics and eliminate as many possibilities as possible.

I also backtrack what I've done, did I put the part on right, did I bump or break something else while doing the installation?

And bad bootleg parts are a possibility I suppose.

EDIT:

The over-voltaging alternator theory may be the winner.

[kach22i]

Originally Posted by famoroso

Solid state...

911 Solid-State Relay — Focus 9 Technology, LLC

They also offer a pump-prime version that enables you to hear the fuel pump running prior to cranking the engine.

My neighbor has some of these on his 1976 with 3.2L in addition to the stock cylindrical ones found on my 1977.

Thanks, I wasn't sure what it was.

[darrin]

seems we all (folks with 3.2l at least) run into the overcharging alternator situation at some point . . . .

[ischmitz]

Originally Posted by darrin

seems we all (folks with 3.2l at least) run into the overcharging alternator situation at some point . . . .

Fortunately, that’s one of the easiest conditions to diagnose with a cheap multimeter.

[Quicksilver]

As ischmitz pointed out we aren't exactly sure where the problem is so . . .

Let's ask the question I generally reserve for "broken computers":

"Exactly what are you seeing that leads you to believe it is broken?"

[wjdunham]

Originally Posted by darrin

seems we all (folks with 3.2l at least) run into the overcharging alternator situation at some point . . . .

Is there a product that can be put into the circuit to prevent the overvoltage situation that is suitable for automobile environment. Something that would indicate when the condition occurs so the alternator can be changed out, and will limit the voltage to 12V?

[mysocal911]

Originally Posted by ischmitz

Fortunately, that’s one of the easiest conditions to diagnose with a cheap multimeter.

Not so when it's intermittent.

[r lane]

Heard back from ecu Docs and am sending them the unit again. They said they will go over their findings with me. Until now, what failed within the unit had not concerned me as I thought it was simply an internal component failure due to use and age, not damage from outside. I have one of the cars at the shop, will want to get the other one over this week for comparison. Thanks again for all of the input. Will keep you posted. Bob