Checking Coolant Temperature
 

The coolant temperature gauge in my 1981 300TD seems to read high. I've checked it with an infrared temperature sensor gun at various points on the hoses and it supports my theory. I replaced the sender and it made no difference. I would like to sort this out. One idea I have is to put a secondary sensor and gauge on, but that would mean sacrificing a hose and would cost me about $60-$70.

What are some other ideas for determining if my temperature gauge reads high?

W123 gauges are known to become erratic because of a poor ground of the instrument cluster. Check the back of your instrument cluster for the grounding point to ensure that the solder joint is not cracked. If it appears OK, run a new ground wire (brown in color) from the grounding point to the steering column support where the other ground wires are attached. Make sure the new ground wire is long enough to ease the future removal of your instrument cluster. Since you have already replaced the sending unit, besides a poor ground, you may have a defective gauge.

That makes a ton of sense. It's exactly the first thing we check when we have an electrical problem on our 1960s Italian motorscooters. Check for a bad earth.

Look in the manual for a temperature to sensor resistance chart.

1) To test the sensor for accuracy:

Get a single pole / double throw ( SPDT ) switch ( A 3 way household switch would work ), some wire , terminals and an Ohm meter.

Assuming a single wire sensor, connect the sensor to the common switch terminal ( black on the household switch ) . Connect the + lead of the ohm meter to one pole of the switch then the previously removed car side sensor wire to the other. Attach the Ohm meter - lead to ground.

You can now switch between using the cars temp gauge or the Ohm meter as a gauge ( with the assistance of the temp to resistance chart ) .

2) Conversely, if you got a resistance substitution " decade " box you can test the gauge for accuracy.

In this case, wire the car side of the sensor harness to common , the sender to one pole then one wire of the decade box to the other. Be sure to ground the other wire from the decade box.

Even if you got 4 single resistors and manually put them in circuit to test range, that would be helpful.

I'd try to drive the gauge to the first graduation, something at operating temp , something half scale and something at last graduation.

That sound like fun in fact.

Yep, and the same setup can be used to test an electric oil pressure gauge as well as a fuel gauge.

On mine the Thermostat Housing has a place to drill through and put a sensor or something else. See attached pic.

I wondered what that cast hole was for. Makes perfect sense. Wonder if any vehicles which used that T-stat housing had such a sensor.

I shoot at the copper base of the temperature sensor. As I recall the fwd sensor in the head (driver's side) is for the dash gage and the aft one is for the EGR system. Keep in mind that the IR gun has a viewing cone larger than the laser dot. You want to completely fill its view with the surface to measure, otherwise it will average in lower temperatures such as from a wire or tube obstructing part of the view. Get it as close to the surface as you can. In theory, the surface material can matter. It is most accurate on flat black surfaces, though I seem to get the same readings on dull aluminum or copper nearby. My gun is the cheap Harbor Freight one.

But, another approach is to go by your T-stat. As the temperature rises, it should hold for a while at (or near) the T-stat setting. If over-heating, once the T-stat is full-open it will no longer regulate. If you trust your T-stat and you see it steady for a while at say 90 C on the dash gage, but your T-stat is an 82 C model then you know that is really 82 C. To trust the T-stat, place in hot water with a mercury thermometer (suspended in the water, not touching the pan bottom). I also shot the T-stat thru the water w/ IR Gun and it closely matched the thermometer. The T-stat setpoint is not exactly one value. That value probably means when it first opens. To open fully requires a slightly higher temperature. Control engineers term that offset "proportional droop".

I had a bad T-stat in my 1984 300D. The dash gage showed it settling at 60 C. Tested in water, it opened a bit sooner than a new one (and a Chrysler one I had, all same set-point) and opened more sluggishly. Replaced it and the car then ran at a perfect 82 C (slightly warmer in summer due to offset above). Had I just done the hot water test without precise measurements, I would have thought "works OK". I recall I posted the resistance values for the various temperature sensors on the engine a few years ago (search).

Coolant Temperature
 

Remember too that the gauge may be accurate and showing 100* C, that's not considered to be too hot although the average American likes to see it at or below 80* C, that's not realistic in summer nor with the AC running ~ Diesels are compression ignition and as such love heat and plenty of it .

I too use the Harbor Freight ($30) infra red temperature gauge, it's very handy and easily spots clogged radiator tubes and dragging brakes, etc., etc.......

*DO* check or add a ground wire to the dash binnacle, it makes the gauges more accurate and the lights brighter .

Use a BROWN wire for grounds, always .