> I'm a little confused.

Not half as much as I usually am.

> My 1986 740T wiring diagram shows the fuel pump relay being operated

> by the ECU, not the pumps.

Yes, the ECU provides a switched ground for the pump relay, which operates the pumps indirectly. I supposedly knew that, just wasn't thinking clearly. Now that I'm awake and actually have a schematic in front of me, I'm thinking that a general voltage surge would be about the only thing that could kill the pump relay part of the ECU.

All I really know is that on my failed unit some of the transistor and diode solder points showed signs of overheating and that some ECU failures reportedly just affect fuel pump operation. In my case it was a more extensive failure. Tracing these hot points back through the circuitry to their source was a bit beyond my skill level.

> I could see the transistors having problems if the injectors had

> problems, perhaps.

Yes, that's a much better theory. The injectors are pulsed to ground directly through the ECU transistors which have to carry more amps than those for the fuel pump relay. An internally shorted injector might overload the ECU, but I don't think that's a likely failure mode. The injector gets it's power from the radio noise supression (injector) relay. On LH 2.2 systems (but not LH 2.4) there's also a ballast resistor between these. If a surge or short started at or went through one of these then you could have problems. One such possibility would be a worn injector relay where the internal contacts are starting to arc. If I'm right then replacing that relay may be a preventive measure. Another would be an old ballast resistor that is crumbling and partially shorting internally.

> As for soldering, I have never had any problems with soldering

> components, afterall, these things are passed over a wave soldering

> machine.

Yes, but that's hot and fast so it doesn't get a chance to build up excessive heat in the component. If you use a wimpy soldering iron and have to linger with the heat to get a melt or if you're trying to remove a component and have to repeatedly apply heat (especially if you don't have a de-soldering tool) then more heat gets a chance to conduct into and build up in the component risking frying. I suppose it's more of a precaution for us neophytes -experienced types with a proper iron that can get the job done fast have less to worry about, at least that's the rule of thumb I was taught.

> I would still like to know if anyone has pulled an

> inoperative ECU apart and checked the solder joints.

Yep, first thing after checking for a dirty edge connector was to inspect for dirty/cold/cooked solder joints. These are fairly high quality boards so apart from the odd hot spot there was nothing. Anyway, I re-soldered 50 or so points just on spec -didn't help. On a second attempt to actually pull a few suspicious components out of circuit for testing I noticed a couple of the traces at the hottest looking solder points separated easily from the substrate. A solder bridge or jumper wire would repair that -still nothing.

I was able to get a Bosch re-furbished unit without having to do a core exchange. So, being the inquisitive type, I cracked the new one open and compared it side by side with the old one. As I recall, the board itself had a different number, but looked the same to me. Almost all transistors, diodes and major caps were new and, if anything, looked beefier. A few that I checked had different Bosch house part number from the originals. As I recall, I was able to cross reference one new transistor and it had a slightly higher rated amperage than the generic cross reference of the original one.

The old one has been sitting in my parts bin for over a year waiting for the day I've got nothing better to do than attempt a re-build.

It would sure be interesting to hear the story from the techs that actually re-build these things. Probably something like, "Re-build 'em? Heck, we just chuck the board and re-use the case."