midoko
15+ Year Contributor
- 36
- 27
- Oct 5, 2010
-
Clarksburg, MD and Winchester, VA,
Virginia
My 90 Eclipse started having an issue with the tach breaking down (needle falling to 0) when it rolls above about 3500 rpm. Odd thing I noticed, the tach "revives" as soon as I lift off the pedal.
I've read tons of threads, with reported fixes by replacing the filter, bypassing the filter, updating the ECU, coil & transistor, et al. Someone suggested measuring a known good filter to see what the cap and resistor values were. Assuming mine was bad, I dug in. It was mounted here on my intake.
The epoxy potting was really cloudy from exposure, so I wet it and was able to make out the values: 2.2K ohm resistor and 2.2 nf cap.
I measured the resistor from the connector and it was still barely in spec, 2.29K. I couldn't get anything on the cap because, for the extremely low value caps, my meter needs it to be plugged into some holes on the meter. I reasoned the cap must be bad, so decided to open it up. I used a heat gun to soften the epoxy, then carefully dug the material out. I only heated it just enough to soften the glue. But I had to reheat it multiple times because it cooled off pretty quick.
I clipped the cap out and it measured high also, around 2.4nf I think. So it was not crazy out of spec. I had some of those caps, so I soldered one in, and tried it out. Too bad, no joy! Next I decided to try the bypass trick. Still didn't work. Back to the forum to finish reading an old thread by Obsidian. The bypass worked for him. Near the end, Steve (wiseman) reasoned the filter may only be there to clean up the signal when/if it gets dirty or to reduce interference.
"It possible that the noise filter is just that, a way to reduce the electrical noise produced from this signal so that it doesn't effect the radio rather than some sort of conditioning required for the gauge cluster. That would explain why just shorting the two pins makes the tach work as does a resistor between the two."
I read tales on other forums where people had upgraded their old points distributors to HEI, and it had messed up their tachs. They too where sharing schematics of various noise filters to clean up their erratic tachometers. Mitsubishi engineers anticipated some noise, or saw it, because the svc manual schematic shows shielding on both sides of the filter (blue).
So now I'm thinking the tach gate is my issue all along. If the caps in that are bad, then the conditioning they're supposed to perform is not happening! I'll try to measure it tomorrow. And as far as the people where the bypass worked, maybe the issue they had was that their noise filter was open, or they didn't have a filter, so installing the jumper just closed the circuit and their tachs worked.
On a side note, my car has a 91 ECU, and a 90 coil, transistor and tach. Only other 91+ sensor related item is the IAC. But I think I read that nobody could get the tach to work with the 91+ ECUs without also adding 91+ coil, transistor and gauge pack. Is that true? Because, until recently, my tach had been working fine for 10s of thousands of miles.
You must be logged in to view this image or video.
I've read tons of threads, with reported fixes by replacing the filter, bypassing the filter, updating the ECU, coil & transistor, et al. Someone suggested measuring a known good filter to see what the cap and resistor values were. Assuming mine was bad, I dug in. It was mounted here on my intake.
You must be logged in to view this image or video.
The epoxy potting was really cloudy from exposure, so I wet it and was able to make out the values: 2.2K ohm resistor and 2.2 nf cap.
You must be logged in to view this image or video.
You must be logged in to view this image or video.
I measured the resistor from the connector and it was still barely in spec, 2.29K. I couldn't get anything on the cap because, for the extremely low value caps, my meter needs it to be plugged into some holes on the meter. I reasoned the cap must be bad, so decided to open it up. I used a heat gun to soften the epoxy, then carefully dug the material out. I only heated it just enough to soften the glue. But I had to reheat it multiple times because it cooled off pretty quick.
You must be logged in to view this image or video.
You must be logged in to view this image or video.
I clipped the cap out and it measured high also, around 2.4nf I think. So it was not crazy out of spec. I had some of those caps, so I soldered one in, and tried it out. Too bad, no joy! Next I decided to try the bypass trick. Still didn't work. Back to the forum to finish reading an old thread by Obsidian. The bypass worked for him. Near the end, Steve (wiseman) reasoned the filter may only be there to clean up the signal when/if it gets dirty or to reduce interference.
"It possible that the noise filter is just that, a way to reduce the electrical noise produced from this signal so that it doesn't effect the radio rather than some sort of conditioning required for the gauge cluster. That would explain why just shorting the two pins makes the tach work as does a resistor between the two."
I read tales on other forums where people had upgraded their old points distributors to HEI, and it had messed up their tachs. They too where sharing schematics of various noise filters to clean up their erratic tachometers. Mitsubishi engineers anticipated some noise, or saw it, because the svc manual schematic shows shielding on both sides of the filter (blue).
You must be logged in to view this image or video.
So now I'm thinking the tach gate is my issue all along. If the caps in that are bad, then the conditioning they're supposed to perform is not happening! I'll try to measure it tomorrow. And as far as the people where the bypass worked, maybe the issue they had was that their noise filter was open, or they didn't have a filter, so installing the jumper just closed the circuit and their tachs worked.
On a side note, my car has a 91 ECU, and a 90 coil, transistor and tach. Only other 91+ sensor related item is the IAC. But I think I read that nobody could get the tach to work with the 91+ ECUs without also adding 91+ coil, transistor and gauge pack. Is that true? Because, until recently, my tach had been working fine for 10s of thousands of miles.
Last edited:
), I swapped in my spare GVR4 ECU, and it worked fine. I took my 90 ECU to Dave to fix. Meanwhile, the 91 ECU, it's still installed. Yeah, so that's why your posts about requiring a tach fix with the 91+ ECU set off bells. I don't know if pins #6 & #14 are swapped in the harness, or whether the car will even run if they're not??? But it is and has been running. I don't drive it often, so I didn't connect the two events (i.e., ECU swap and tach breaking). So today I opened up this 90 ECU and it's not like the one in Jeff's internal fix install procedure, presumably indicating it is a 90. The whole point of this story is: What if the reason my tach isn't working, is because there is a 91 ECU (GVR4 spare) in my 90 Eclipse??? I'm going out in the rain to take it (91) out, and put the 90 back in.
s
