VSS goes nuts when warm!

[KY Metro]

I need some expert opinions.....

I installed my MPGuino in my 1995 Geo Metro. I am very happy with it, have needed no calibration since the default settings were for the Metro. Usually everything is fine. However, I notice that the speed reading is sometimes wildly inaccurate, usually ten to twenty MPH faster than I am actually going, but also going to zero and very low speeds on occasion. This affects the distance and the MPG reading as well, showing false mileage gains. Very frustrating.

I pondered this a while, and eventually noticed that it only happens when the heat has been on for a while - the instrument cluster is actually getting quite warm, and then begins to fritz on the speed. I have even managed to get it to "calm down" by directing cold air onto the cluster.

I know that on Metros there is a reed switch in the speedo head, according to Coyote X. He was after the bounce in his VSS reading in this thread, which is NOT the issue here - when mine works, it is spot on and pretty steady.

Is this the first time anyone has heard of this? What piece has failed, and what can I do to fix this?

Thanks,

KY Metro

[McTimson]

Mine has had the VSS signal drop to zero when it's very hot in the car, and a few times when it was very cold. I'm chalking it up to a bad connection that I made behind the speedometer where I have the VSS connected, because hitting my speedo a few times usually fixes the problem.

I don't think that's the problem in your case, though it couldn't hurt to check the connections.

[KY Metro]

McTimson,

I soldered the wires inline, up near the ECM connector. From there the wires go to a RJ11 wall jack on the ECM, connected with screw terminals which haven't given me any trouble.

I am thinking the problem is at the source in the speedo head. I still get signal - it's just wildly inaccurate when hot. As I understand reed switches, they are just two contacts in a vacuum tube. When the magnet comes past, the contacts inside the glass are attracted, touch for a second,and create a signal. Then the magnet is gone again, and they pull away and break the signal. At higher RPMs, maybe they bounce and jitter, but this hasn't given false readings when cold. They are vibrating pretty fast - at 45MPH they open and close 6156 times per minute!

I wish I had an oscilloscope to see that wave pattern. Would be interesting.

[KY Metro]

Was looking around today, here is a good picture of the reed switch inside the speedo head (pic from a Suzuki Sidekick).




Apparently reed switches are the cheapie mechanical version of the usual solid-state Hall effect switch, which never fails. I don't think I can swap them, as the voltages probably look very different to the ECU.