Yes I agree with Alan, there is a break in the power feed some pictures of the engine harness might be good ,
maybe the harness was damaged where it comes through the timing belt cover.
Maybe the positive terminal on the battery is suspect as well,
I would also guess that you inspected the power lead from the battery to starter,
and the power wire isnt rubbing on the swaybar .
And the starter connection is also tight
Also check the 3 power wires that go to the top of the CE panel,
maybe they got mixed up somehow

 

A quick check of the voltage at the relay socket Terminal 30 for the starter and EZK relays may help determine the source of the loss of power. The EZK pulls power off one of the small leads at the + terminal of the battery. The starter relay pulls power off the jump post connector. If you have no voltage at either I would suspect the connection at the battery + terminal. if you have EZK relay voltage but not starter relay voltage then suspect the supply chain from battery to starter to alt to jump post.

In either case you could visibly have no power to the system and an engine that would continue to run after started. If the former the entire system would be operating off the alt generated power. If the latter the engine management would actually be operating off the battery while the rest of the car would be running off the alternator.

 

Alan thank you for confirming my suspicions. The ground cable has been replaced. I will spend more time with the alt-starter cable. The main positive cable is so well supported and protected that I cannot see that being the issue. On the other hand, the alt-starter cable runs through the active area relatively unsecured and has a lot of opportunity for damage. I still don't understand why it is so random and so quickly self-healing. It's not like I hit a bump and something happens. It's totally random both ways.

And since the alternator continues to drive the car when the battery drops out with voltage and tach gauge anomalies, I think that eliminates part of the engine harness.

 

This would be my lead suspect too.
You could do various socket/terminal testing, but you KNOW where you will likely find the lesion.
Go with the horses, and don't hunt the zebras.

 

Only way to test this is when the next random event happens, quickly connect my multimeter as you describe. So far, the only things I managed to do were measure 1 to 2 volts at the jump post while the battery was 12.7 across the posts but with no spark unfastening a known good ground cable from the rear deck. So, the ECUs could still be drawing 12V from the small wires despite no sparking when changing the ground connection? I suppose their current draw is so small that could be the case.

I believe I've ruled out the battery post connection by inspection, cleaning and manipulation and battery replacement, all of which had no effect. And since the car runs on the alternator, doesn't that rule out the alternator to jump post?

And Maine, I'm right there with you about looking for horse rather than zebras. Thanks.

 

[QUOTE=Bill Ball;10785151]Only way to test this is when the next random event happens, quickly connect my multimeter as you describe. So far, the only things I managed to do were measure 1 to 2 volts at the jump post while the battery was 12.7 across the posts but with no spark unfastening a known good ground cable from the rear deck. So, the ECUs could still be drawing 12V from the small wires despite no sparking when changing the ground connection? I suppose their current draw is so small that could be the case.QUOTE]

The relays themselves would be open so even if you have voltage at T30 you don't have any current flow so therefore no spark when disconnecting the battery.

 

Thanks Simon.

 

Yep. Get her up. Disconnect at starter and alternator. Snake out the line. A little other disconnecting may be involved, but you know where you are. The look of the sheath may give the immediate answer. If not, slice it, and inspect the cable itself. If for some reason it is not there, you have now definitely excluded it. Should take an hour and 3 cups of coffee. Best Wishes.

 

This is true - but the interior lights alone are enough to generate sparks on reconnection....

Alan

 

Any chance the positive battery cable is corroded
under the insulation at the battery terminal end
(like what happens to the ground straps)?
Might be worth opening it up to have a look. If
nothing else, it would eliminate another possibility.

 

It might be. I still don't get why the problem changes so capriciously. I would expect a more consistent failure with corrosion like that.

 

How about running a new +ve wire ( in parallel) from the battery to the starter, and a lower gauge to the jump post - yes, a pain to do, but doable with 2/0 and 2g from HD. I did this on the $900 84 euro i rescued because it had intermittent power loss issues, the +ve cable at the starter had a bad crimp and it was moving around in the crimp itself - I pulled the cable hard at the starter and left the crimp terminal behind .

 

Neil - that prompts me to look more closely at the crimps. Thanks.

 

Finally

 

Yes I'd have been looking there too...

Alan