why does a good DME relay not work?
 

Here is the series of events:
The car has been running great. Installed the new set of battery cables, including the supplemental + to the fuse tray. When that was finished, the car would crank but not start. Messed withe the fuse tray and the car started and ran. Set the fuse tray in place and started to put on the cover and the car died. This happened 2 or 3 times. In short, jiggling connectors under the fuse tray, though not sure which, appeard to make a difference. Right now, the car isn't starting no matter how much I jiggle (the connectors).

A little more methodically, found that there is spark always, without fail. What quits is the fuel pump. Replacing the DME relay with a brand new one makes no difference. Since the cover comes off the old relay easily it was no problem to verify on the bench that both coils in it are good. For certain, in the car the relay section that energizes the fuel pump fails to close and fuse 34 remains dead. If I force the relay contacts closed, the fuel pump comes on. The - side of the non-working coil is supposed to ground via pin 20 of the DME. I'll be checking for continuity in the wire harness tomorrow.

Any ideas on what could have been disturbed by installing new battery cables? Anyone seen this problem before?

this is a fine example on why i hate the dme relay

sounds like the battery cable install went fine, id trace the wire again just to make sure everything is snugly connected, you might want to check the speed and reference sensor, might have a vaccumm leak, or the AFM might be bad

The under hood (later 944) fuse and relay assembly has been known to develope bad relay bases. The fingers that press against the pins on the relays and fuses come unsprung. You can pull the whole inners up so you can get to the underside and check your wire connections. The relay bases can be taken out individually too.

I have had the entire fuse tray apart, all 7 layers of if. Cleaned all the connections with scotchbrite and wiped on a film of light oil. Did resistance checks on the bench from inside the DME relay to all related harness connections underneath the fuse tray- all good. Still no fuel pump. The relay, of course, is good: both coils work and when the contacts are closed manually the pump comes on.

It looks like the pump relay coil is supposed to ground thru the DME pin 20. What inputs does the DME need to enable the pump relay while the car is cranking?

Thanks,

During starting the signal to the starter coil is split off to the DME. During running the signal from the speed sensor (the one that is over the starter ring teeth) keeps the fuel pump running.

The spark is from the reference sensor -- "sees" a pin on the flywheel in front of the starter ring.

Possibly a bad connection at the sensors or the sensor tip has metalic crud or it's adjustment has drifted. A bad connection can be the connector does not latch properly to the other or even a cracked connector that has broken the inside wiring.

I've taken temporary measures to get the car running:

inside the DME relay, jumpered pins 85 and 85b, so the pump relay coil shares the hard ground connection of the main relay coil. With that the car starts and runs perfectly. This tells me that whatever the speed and ref sensors are supposed to do, they must be doing, otherwise the engine would not run smoothly as it does. I understand there is now risk of a fiery death in case of accident since the DME can't shut off the fuel pump any more. I intend to keep looking for the root cause of the problem.

What is really puzzling is that after my son installed the new battery cables the car ran fine as always. Then he installed the new supplemental cable and the problem began. I've checked the sensor connections and they are secure and undisturbed. Seems like it should be something simple. Why doesn't the DME pull down pin 20?

Thanks for all the suggestions

With the ignition on there is voltage at the DME connector corresponding to pin 20. Is that what you mean?

Inside the DME, pin 20 connects to a CD4049UBEX hex inverter buffer, which is obsolete (Harris). I am trying to find its data sheet to compare and determine if a CD4049UBE (TI) would be a suitable replacement.

Thanks,

Here is the wrap-up on this problem:

We replaced the CD4049UBEX (obsolete) chip in the DME with a CD4049UBE and plugged in a fresh DME relay without my workaround jumper modification (see above). Car starts & runs great!

Cost: $0.40 and a bazillion hours troubleshooting the car and researching that the new chip is compatible with the old.

Thanks to all for your input.

How much was the "fresh" DME relay?

Still think it was a bad DME relay but then you "fixed" everything at once so who is to know.

I think my troubleshooting was pretty methodical.

That fresh DME relay is the spare out of the glove compartment and substituting it for the old one did not fix the problem initially. There was 12V all the way to pin 20 of the DME but no path to ground thru the DME to complete the pump relay circuit. That is why the car would not start.

Since neither the old nor the new relay would turn on the pump I temporarily modified the old one to provide an alternate path to ground for the pump coil so the car was drivable. The point is that only the unmodified fresh relay would demonstrate that the chip replacement was effective.

I replaced the chip and the unmodified relay works fine.

Gene

You reported that both took place at the same time. It appears from your follow up not to be the case.

Aging solder connections in the older DME's have been a problem. One of the last resort fixes is to reflow the DME solder connections. Another has been burnt through PCB traces because the grounds from the engine to the chassis were left off. Gounding then goes though the DME which is not designed to take the load of the starter.