temp sender repair+description! (long) (useful)

[Brett Dikeman <brett@pdikeman.ne.mediaone.net>]

If no one has done this before, this post would be a good addition to the
archives.  I have some snapshots that will be ready as soon as the photos
get developed and turned into a PhotoCD(along with pics of Boomer!)  Web
site will develop.  This is a fairly long post; 5k by Eudora's count.

Today, with the help of a dremel, I disassembled two sensors; the Oxygen
sensor ($60-160) and the thermo-switch($81-91.)

I'll covered the thermo-switch here(and first; I will follow with the OXS
later, stay tuned.)

The coolant thermo-switch is a 2-part assembly; the brown plastic and the
brass.  For those that haven't seen one or taken it out, a cylinder extends
from the base of the sensor into the coolant stream.  This cylinder houses
wax, which expands with heat.
  The plastic section houses a set of two contacts and a large lever.
 \   |
  \  |
   \ |
    \|
     *
     |\
     | \
 ----|~~~


The vertical piece serves as a pivot point.  The tildas are where the 2
springs are(in line, one for each switch contact.)

Kind of hard to describe, but as the lever gets pushed down, the other end
comes up, changing the position of the springs so that the force they exert
on the levers they are attached to goes more UP than down, and the lever
flips upwards and hits the other contact (BTW, the two springs have
different tensions and/or slightly different placement, so that one flips
before the other.)

Failure mode 1: sensor always reports overheat.
Funny thing, but after I took mine apart, there was a nice line of metal
fuzz from one contact to the other.  Bingo.  Fix?  I bet if you, uh,
practiced your fastball with the sensor, that fuzz would get knocked clear
out of the way.  No more overheat signal.  More complicated solution?  See
later.

OK, so the sensor also does one other thing aside from overheat and fan
stage switching.  It also drives the coolant gauge.

  Off the plastic base there is a tower that extends towards top of the
"chamber" inside.  This tower has a contact spring which presses against a
small black disk, which has silver or some other metal coating both sides.
It is held in place with a U shaped clip.
________
| ____  |
| | D | |
--|---|--
  | ~ |

D is the disk, ~ is the spring pressing up.  The outer U is the clip, and
it holds on to two indents on the little "tower" that keep it from popping
off, but the disk and clip can travel up and down over a fair range.

The disk and clip then press against the brass inside of the sensor at the
top of the "chamber" or empty cavity inside.  This thing uses the
temperature of the body of the sensor, not directly monitoring the coolant
temp.

Failure Mode 2: sensor stops sending signal to gauge, is intermittent,
bumps bring sensor back or kill it.

Kinda obvious now why; if the inside corrodes, there are a grand total of 3
connections where it could fail; that's 6 surfaces that could corrode or
rust.  One fix could be to put the brass part in a vise and turn the
plastic part with a wrench _carefully_...this would move one side of one
connection for that sensor.  One could also just completely disassemble the
whole thing(use a moto-tool with a grinder wheel to remove the "lip", pull
the plastic part out of the brass part.) and clean it all out, then put
back together and epoxy(holding tight with a vise or something so the part
is held together until the epoxy sets.)

It is now also obvious why this sensor hangs _DOWN_.  It's also now
semi-obvious why they use brass(low/no corrosion.)  If it hung the other
way, water would collect _inside_ the sensor, since it is mostly sealed
with a pressure fit between brass and plastic. I won't get into spaces
opening up because of the brass+plastic heating up, changing size and so
on.  There is a small, thin rubber seal that the plastic part is pushed up
against and I guess acts as a seal, though with time small spaces could
open up.  People will complain about the connector going bad, but OTOH
what's worse?  A $81-91 sensor or a $1-5 connector?  Connector failure
doesn't seem to be the worst problem.

So, there's the inside of the multifunction temp sender/sensor and a few
ways to repair it.  It looks like it would function fine once being taken
apart, just make sure you press it back together nice and tight or the
overheat and fan switch will not work properly.

Like I said, pictures and a page to follow; a post about the oxygen sensor
coming shortly(tonight some time.)  Pictures of the disassembled OXS are on
the same roll of film.

Anyone who wants me to take apart any other misc sensors or parts, let me
know; you send it to me, I'll hack it up :)  At some point I'll post about
window motors; they're suprisingly easy to take apart.

Brett "Stop the temp sender insanity!" Dikeman
91 200q 20v aka "Boomer"  (with new temp sender and OXS.)

"Son, you have too much time on your hands" - Paul Royal

------
Brett Dikeman
brett@pdikeman.ne.mediaone.net
~)-|
Hostes alienigeni me abduxerunt.  Qui annus est?
Te audire non possum.  Musa sapientum fixa est in aure.
Ita, scio hunc 'sig file' veterem fieri.
------