Mysterious temp gauge

[lcheng@watson.ibm.com]

I received a lot of mail concerning this problem.
This seems to be a more popular topic than I thought!

Some checkpoints:

- As stated in my earlier note, the temp gauge shows the correct reading
  as the engine warms up, and as soon it reaches a stable operating
  temperature, it dies to zero.  The gauge at random times chooses
  to indicate the correct temperature.

- My '88 90 Quattro has an analog instrument panel.

- Engine warm, ignition off: checked resistance across the coolant
  sensor.  It was roughly 800 ohms, and climbing by the second as the
  temperature decreased.  The Haynes manual for the 5000 (a former car)
  says that the resistance should be between 1000 and 400 ohms, assuming
  that it's the same kind of sensor (probably a fair guess).
  Forgot to do a continuity check on the wiring to the gauge.

- Engine warm, ignition on: jumped the connector's contacts with a 700 ohm
  resistor, which I'm guessing should be roughly halfway along the temp
  gauge.  Said gauge doesn't budge, still dead at zero.  Shorted the
  contacts together, without resistance - still no effect.  Measured
  4 V DC (shouldn't it be about 12 V?).  Gauge still works today, still
  intermittent.

- Haynes indicates that the sensor is wired directly to the gauge (with
  both the temp and fuel gauges connected to a voltage regulator - makes
  sense).

So I was hoping to pull the instrument panel and take a look, making
sure that there aren't any flaky contacts, and so on.  Any more guesses
as to what's wrong are certainly welcome.

By the way, the original question still stands: how do you remove the
instrument panel?

- Lawrence Cheng
  lcheng@watson.ibm.com