At 05:56 AM 10/16/2014, Frank Condelli wrote:
>The number one reason the coolant light flashes "slowly" when that
>reservoir is full is because of a faulty coolant level sensor. They
>are cheap, I sell them for $9.95 and

Frank, this is a mistake.  The flash rate is built into the LED
inside the temp gauge -- it's a special flashing LED that operates on
five volts.  Inside the gauge it's in series with a 5.1 volt zener
diode that drops the ten volt gauge supply to what the LED can
handle.  You can replace the gauge LED with a brighter one by
replacing the zener diode with a load resistor; but if you want it to
flash you have to use a self-flashing LED (which may or may not work
as a drop-in to the original).

The coolant level controller when triggered loads down the gauge
sender line with </= 35 ohms to simulate an overheat condition.  The
gauge has no "idea" it's being tricked.  The old-type controllers did
this continuously, so the gauge needle would show overheat; but the
newer cube-shaped ones only do it in brief pulses, long enough to
trigger the flash but not long enough to disturb the gauge
appreciably.  Any time the flasher is triggered it runs for a couple
seconds, so the controller only has to pulse every second or two to
keep the flashing going.

Long ago I analyzed the circuit to component level of the older
controller and I understand exactly how it works in detail (or did at
the time, anyway.  It's quite a clever circuit).  The newer ones have
been collapsed into a single integrated circuit but behave
essentially the same way.

Yours,
David
ps - there are two or three different p/ns for the newer controller;
they all work the same and there's no preference among them.  All are
hugely preferable to the old one as they don't peg the gauge on a
low-coolant alarm.