No Start, Potential ECU Issue

[19Eclipse90]

Last week, the family '92 Stealth TT began really giving us issues being very temperamental not wanting to start. I was on the phone with my dad (who is actually with the car) running through component test procedures earlier this evening when we ran into a concerning result with the cam position sensor harness: CAS, Crank and CAM Angle Sensors - 3SI Wiki.

Everything was kosher until the third step in the harness inspection -- measuring the voltage at pin 1 with the sensor disconnected and the key in the "On" position. As the link states, the voltage reading should be about 5V; instead he was measuring around 10-11V a great majority of the time. At those voltage levels (and the sensor reconnected), the car will not start whatsoever. Somehow, someway, the voltage found its way to 5V again and the car did start when the sensor was reconnected. As far as I know, and I tried very hard to confirm with him over the phone, nothing was changed. Shortly after that and again without any changes, it didn't want to start so the sensor was disconnected and the reading was back around 10-11V as before.

I know pin 1 runs directly to pin 68 on the ECU and that the ECU should provide the 5V; I just want to make sure that points directly to an ECU problem (blown resistor, short, something like that) and there aren't any inner workings/electrical magic I'm overlooking that might cause this? FWIW, the ECU was pulled and the capacitors look fine, nothing looks like it has blown up, and there is no fishy smell.

Thanks for looking and the insight!

[steve]

Seeing 10-11v on 68 of the ECU would indicate that the voltage regulator on the ECU has problems and I'm surprised the corrected itself and still worked. I would have expected the CPU chip to blow if it's seeing 10v. I can see it far more likely to survive if the CAS was bad and pushing voltage down the wire instead or a short in the harness.

You can test the CAS by itself if you provide 12v and ground to power it and current limited 5v for the signal lines. Then when you turn the CAS it will pull the 5v down to ground as it rotates. The 5v power has to be current limited or you'll destroy the CAS. Something like 1k resistance between the CAS signal pin and the 5v power should be safe.

[19Eclipse90]

I relayed the information and was told that there had been wires shorting out and since repaired previous to me talking with my dad last night. He also said that when he went to pull the ECU (should have pulled the negative battery cable, obviously) that sparks shot out from somewhere down there. Either that was enough to save the CPU chip or there really is some damage that he just couldn't see.

In any case, we didn't trust the ECU very much and with you (Steve) suggesting there likely is an internal problem, we are confident it is in need of a rebuild. Nearly 20 years old and 215k miles later, I'd say it deserves it. I will be going over the ECU wiring myself before reinstalling the unit to ensure there aren't any issues there even though my old man swears it's fixed. I'll post results when it comes around to it.

Thanks for your input and advice, Steve. Your expertise is always greatly appreciated!

[19Eclipse90]

Got the ECU back and installed after checking the wiring and, surprise surprise, it works like a charm.

Thanks again, Steve!