Well, thanks to the assistance of David Beierl <dbeierl@ibm.net>, I got the
water tank sensor working properly today!  It required modification of the
circuit that runs the three LED's in the panel.  Here's the basics of what I
(we) did - I will post full details when I get a chance to write it all up.
(This is from memory, and there may be a few points that I am incorrect
about - David will correct me, I am sure <g>)

Essentially, the red LED was not lighting when the tank was empty (it would
stay yellow).  Upon examination of the sensor wand in the tank with a volt
meter, it was determined that when the bottom reed switch engaged ("empty"),
it was not providing sufficient voltage (5v) to trigger the red LED.

Using a 1 mega ohm linear-taper potentiometer in place of the resistor on
the circuit board that controls the red LED (had to remove the resistor with
clippers), I dialed the pot until the red light lit up with the float in the
"empty" position.  Then I measured the voltage at the green sending unit
wire, making sure it was correct (David?)    Then I tested the other two
float positions to make sure the LED's all worked.

I then measured the resistance of the pot where it was set, determined the
value to be 650k ohms.  Using an assortment of resistors that I got from
Radio Shack, we combined four of them in a little "bundle" which I then
soldered together.  Since it was too bulky to fit to the front of the board,
I soldered the bundle in place on the back side of the circuit board.  I
shielded the bundle from accidentally shorting to some of the other
components and solder points by encasing it in a small piece of 3/8" heat
shrink tubing.

By changing the voltage value at which the red LED kicks in, I now have a
working water level sensor!  Many thanks to David for his electoral
engineering expertise.  :-)

Brent Christensen
'89 Syncro Westy "Klaus"
'95 Cherokee Sport