At 10:31 AM 3/2/2009, Jim Felder wrote:
>Replace the temp "relay" and the temp needle calmed down for good, it
>always operates properly now--which is strange, because I wam told
>that the relay controls the light, not the needle.

Hey Jim -- tsk tsk.  You didn't read what I told you about the
controller -- it tells the gauge that the engine is overheating, and
if you're using the old controller the needle goes to the overheat
zone.  The new controller only sends the signal in pulses, so the
needle stays where it belongs roughly.  But it's all accomplished by
simulating an overheat, whether continuously or in pulses.

Also if you'd said initially that the gauge was pegged...the light
will *always* flash when the gauge is pegged, that's what it's
for.  So if you have an old controller and a pegged gauge, the
problem is practically certain to be the controller.

As to the blinking light changing rate -- the circuit has three
inputs: +10v from the panel voltage regulator, ground (third terminal
on the back of the gauge) from the panel ground, and whatever is
coming from the temp sender/level controller line.  Blinking is
initiated by powering on the system; this is part of the power-on
panel test.  It's also initiated by an overheat condition.  These are
normal.  Any time the gauge starts to blink, it should continue for
at least three seconds.

I know that if the voltage from the panel regulator is the tiniest
bit below 9.5 volts, the gauge will blink continuously.  However, I
*don't* think that a poor panel ground would do this, because the
voltage regulator takes its reference from panel ground and would
thus raise its output voltage within its limits to accommodate.  I
would expect the limits to be up to about 1.5 volts below the panel
supply voltage.

The blinking rate of the light is set internally by the circuit.  It
is not supposed to vary, and it should not be affected by anything
that happens to the temp sender line.  The only *external* thing that
could affect it would be bad voltage -- too much or too little --
between the plus and ground pins on the back of the gauge.  I don't
know that either of these would, but they are the only
possibilities.  The other possibility is an internal circuit failure
inside the gauge.  There is undoubtedly a capacitor circuit inside
that controls blinking rate.  If that capacitor malfunctioned, the
blink rate would certainly change, and I can imagine a situation
where the cap had a progressive failure based on power-on time of the circuit.

The blink duration is also set internally, nominally three seconds,
by the large capacitor on the board.  If the light fails to stop
blinking when it should, and the supply voltage is the correct +10v,
then that cap is almost surely the culprit.  You can replace the cap,
but the board has to come out of the gauge.  The cycle-the-key trick
and the works-normally-after-warmup are both indicative of this condition.

Hope this helps.

  --
David Beierl - Providence RI USA -- http://pws.prserv.net/synergy/Vanagon/
'89 Po' White Star "Scamp"