95 LT1 Erratic voltages EEhack

[Opti]

I just swapped a 96 Camaro to an obd1 pcm so I can use EEhack and Tunerpro. I was doing a quick datalog, and wanted to do a power balance test. The power balance test wont run because the TPS voltage is bouncing erratically from .4 to .8 which is in the acceptable range, but it thinks there is constantly fluctuating throttle between 0-4%. I noticed multiple other sensor bouncing erratically. I know 3 or 4 of them share a pcm ground.
My first assumption is a ground issue on the car, but before i go hunting and checking all the grounds wanted to see if it may be a signal issue, or if someone had run into this before.
Also the laptop was not plugged into a power source.

[steveo]

pull the belt and try again to prove your alternator rectifier isnt shot.
the tps should be dead on consistent
if it still does it, you have a shoddy ground, or maybe extreme rfi from an arcing plug wire or something?

[Opti]

pull the belt and try again to prove your alternator rectifier isnt shot.
the tps should be dead on consistent
if it still does it, you have a shoddy ground, or maybe extreme rfi from an arcing plug wire or something?
Thanks I'll try that

[Opti]

I pulled the alternator and its connectors. No change.

So the issue is I have erratic signal from many sensors. The obvious ones are map, tps, a/c pressure, ect, iat and ac evap temp. It will even do it with key on engine off. They dont swing far but swing a couple degrees or percentages always. I noticed the negative battery terminal was a little loose. I tightened it up and it helped a little. I checked the main lug on pass side strut tower and cleaned it up, I found a couple broken grounds at the coil. I may have broken them though in dissassembly. Im going to repair those grounds, but even with them off, I still get the erratic signals.
I also unplugged pretty much every sensor or actuator I could find and checked one at a time to see if something was backfeeding. No change. In fact when the erratic sensors are unplugged they remain erratic, just at a different reading.

Didnt have time to get under car and check the grounds on the bottom. Will do that when I repair the grounds on the coil.
Also havent check for an arcing plug wire, I will but I think it might be unlikely because the wires are relatively new and the vehicle runs very well.
Anything else to check out when I get back to it?

[NomakeWan]

Sounds like the reference ground for the PCM itself is floating if unplugging the sensors doesn't resolve the fluctuation.

[Opti]

Thats what Ive figured but the PCM ground is accessible from under the car, and I worked in the rain as long as I could stand. Waiting for a dry day to get under it.

[Opti]

Cleaned up the engine grounds today. Didnt change it. Still does it key on engine off.

Screenshot of datalog KOEO


\

https://i.imgur.com/XWjbuB6.jpeg

Any other recommendations would be greatly appreciated

[MO LS Noobie]

With everything plugged in, key on but engine does not need to be running, and the voltage fluctuating attach a accurate voltmeter to the negative battery terminal and the positive lead to the grounds at the PCM. This is called a voltage drop test. If you have a fluctuating reading of even 0.010 volt you have a bad ground condition that needs to be cleaned up. Do not forget the negative battery to body grounds, Maine ground to engine block, and engine block to grounds.

If you're voltage drop test of the negative side is okay, it must be an internal PCM problem, not likely the only other suspect. Good luck

[steveo]

other stuff can cause this, like corrosion in the key switch contacts or fuse block it's a really old car. go through everything.

[Opti]

Not excited about this. Tried grounding PCM straight to lug on right fender. No change. TPS doesn't act up until engine is running, everything else fluctuates KOEO including AC Pressure, ect, o2s, map and I think a couple others.

Currently chasing circuits from PCM to ground.

[steveo]

weird. im not excited for you either. can you post an eehack log of key on engine off for like 20 seconds?

here's your ecm wiring http://shbox.com/1/4th_gen_tech2.html#1995_pcm_pinouts

briefly how it works, the ECM actually expects noisy power and ground input.

things labelled 'sensor ground' are the ecm filtered grounds, they are not connected directly to chassis ground but have internal filtering. they're solid black.

then there's the 5v reference voltage wires, they're solid grey, they spit out a cleaned up 5 volts

even with fluctuating input voltage or a bit of ground noise, you should get fairly solid 5v out and that's what your TPS etc use

if you want to verify your harness you need to disconnect your ecm and all sensors and start probing things... actually isn't hard but i would test with like a 5 amp load, that way if you have a splice that's on the way out, it'll burn it the hell out so you can find it.

[Opti]

Here is a ~20 second datalog. Key on engine off. Without the additional ground strap to PCM i used for testing, all oe engine grounds are connected.

https://filebin.net/jikd6uiffaxaries

[steveo]

okay yeah you have major issues. i would suspect something on the power side because there are multiple redundant grounds that would have to fail for this to be so bad.

look at your ECM input voltage, it's unstable and way below 12v. a resting vehicle with a good battery would be 12.5v. every connection is suspect, you will just have to isolate the issue.

[Opti]

Any recommendations on where to start, or what to look for. I'm not much of an electrical guy, and most of the diagnosis I have done has been limited to ground side and just making sure voltage is getting to things, more indepth than that and I'm a little out of my depth.

[steveo]

- not an electrical guy
- almost 30 year old fuel injected car

you must change one of the above things

honestly you might just try keeping the logger running running a graph in eehack and start wiggling wires and connectors until something changes. you gotta start somewhere...and something is very broken

also a big long power jumper wire might be helpful. you want to focus on the PCM IGN circuit, if you can find a place to jack power into that wire that fixes the problem you can move backwards from there to find the fault.

everything is suspect, gm really cheaped out on the wiring in these cars. the harness runs to the passenger footwell to some connectors and the body control module, and continues to under the dash where it connects to your fuse panel, ignition switch, etc.

look for anything that's been previously repaired first.

[steveo]

oh another thing to check is the fuseable links in the engine bay, i have seen them corrode. you can run a jumper from positive battery terminal to the red junction box to bypass them and see if that helps.

[kur4o]

It is also possible that the pcm have some inside problems[water infiltration]. Does the 96-97pcm acted like that. You can bench test the pcm.
5v regulator inside pcm may went south. Also measure the voltages and confirm they act like that. There could be some glitches in data communication too.

I`ve been having voltage drop at pcm, and nothing completely cured it. Adding extra thick wires to fusebox, separate wire to fuel pump and so on. Voltage reading at pcm never match the voltage at battery. At the end I had to add a separate wire to a spare a/d channel to read voltage at injectors, so offset being used are as accurate as possible.

The one wire that goes from fuse block to pcm is really weak. You can also check spark wires for arching, running engine in a complete darkness.

[Opti]

Alright alright well that gives me a place to start and some things to look at. I'll start chasing it tomorrow. If I find anything I'll post back.

[In-Tech]

kur4o is correct, it could be inside the ecu. The suck is chasing shite you can't see, lol
I built a cap bank with switches so I could figure out what it needed, then soldered in the capacitance. Figured it out once the harness was out to replace ecu with a holley ecu to talk with the holley dash.
Turned out to be the noise from the o2 circuit had chafed a bit :mad1:

[x11 nut]

Been chasing bouncing TPS voltage for months, a buddy that is an old timer GM Master tech told me to replace the temp sensor and it fixed it. Old temp sensor was still reading in the logs but it shares ground with TPS. Not sure it is the same as your issue as you have it going with more sensors.

[In-Tech]

Been chasing bouncing TPS voltage for months, a buddy that is an old timer GM Master tech told me to replace the temp sensor and it fixed it. Old temp sensor was still reading in the logs but it shares ground with TPS. Not sure it is the same as your issue as you have it going with more sensors.

Yes, the analog ground chasing can be a biatch :( My little cap bank can work on either, but mainly used on the 12v ign sw side.

[steveo]

you could unplug every sensor to see if your input voltage stabilizes

[kur4o]

All these sensors use 5volts reference from pcm, If a sensor is bad, it can screw or fluctuate bad enough the 5 volt source. Really good idea to start unplugging stuff.

[Opti]

I unplugged everything I could get to. About the only thing fluctuating I didnt unplug is the evap temp sensor and o2 sensors. I dont know what the evap temp sensor looks like and I just forgot about the o2s.

I only had a couple minutes so I couldnt get much done but here is what I did.

I tossed the obd2 computer back in and it fires up and runs fine. I do have a super high idle on the obd1 pcm that i assumed was unrelated to swapping the computer and I noticed the erratic readings when I tried to diagnos the high idle.

I check voltage to the obd1 PCM right at the connector (both orange wires) and its steady at 12.06V, it will move to 12.07 every couple seconds but not the frequency like the readings im getting on eehack
I checked voltage out on the MAP sensor (a reading that is fluctuating) and it is as steady as the voltage into PCM.

I did open the obd1 pcm to look for obvious signs of water intrusion or damage. Didnt see anything.

Does this mean I likely have a bad OBD1 PCM1?

[NomakeWan]

Your multimeter might not be fast enough to detect (or, more likely, display) the fluctuations in voltage. Problems like this are where an oscilloscope come in handy.

That said, if you do confirm that the voltages external to the PCM are stable, including the ones that the PCM itself are generating (such as the 5V for the sensors), then process of elimination suggests the PCM is indeed bad. Maybe an ADC circuit, or one of the ADC's reference inputs. It's still able to communicate over the bus, so the whole PCM is not bad, and it's able to generate valid serial data. But clearly something related to the ADC loop is failing.

[kur4o]

You can check the ground strap between 2 boards inside pcm, I read somewhere they can pose such problems. Or pcm can`t filter noise from engine it hits the ADC.
Tom H might give some clues what may be the hardware cause.

[Opti]

I should have a little more time tomorrow for testing. I'm going to go back and unplug everything and see if it still fluctuates.

Also will check the a voltage output from pcm.
I think I've got a buddy who has a 94-95 pcm laying around I could swap in and check.

[Opti]

Local LT1 guy had a Bbody pcm he let me borrow to try.

New PCM fixed it, nice steady readings now, also fixed the high idle.

[steveo]

awesome! never seen that before
these ecms are usually bulletproof

[NomakeWan]

Sweet! Great diagnostic skills, and glad you've got it nailed down!