Coolant Temperature Gauge Reads High

[jpanhalt]

Late Summer, my compact tractor (Case DX40) coolant temperature gauge started to read high. The Case DX40 is almost identical to the New Holland TC40DA. I posted on a tractor forum and got some very useful information shown below. The gauge is a Stewart-Warner and is apparently what it calls a "bi-torque" gauge.

The gauge read well into the red zone after only about 10 minutes of running. Replaced thermostat and sensor. Both items tested OK on the bench afterwards. Gauge pegs high, as expected, when sensor wire is grounded. Now, the advice is to replace the gauge, but I am wondering whether there might be other causes. I hate swapping parts blindly and throwing good money after bad until the problem is resolved. That can get expensive. The gauge is a little over $100.

Here is the circuit for the gauge (from TractorByNet):

And here is what the inside of the gauge looks like (same source):


All of the resistance measurements on my gauge are within specifications, but apparently, a bad gauge can act like that. The S-W troubleshooting chart lists a defective voltage regulator and improper grounding as additional causes to investigate. I have checked the grounding at the alternator. Next project will be to test the alternator output at the battery (NB: battery seems to charge normally and has not lost electrolyte).

I am after suggestions for what to do, if the alternator voltage is within limits.

1) What would you accept as the absolute highest acceptable voltage at the battery? 13.8V, 14V, or ???

2) Would you advise any other tests, like with a scope to look for bad diodes?

3) Does anyone know whether additional voltage regulation is usually built into the instrument cluster? Getting to the PCB for the cluster is doable, but will be a pain. If the alternator voltage is OK, then my next steps will be to built a 12-V powered circuit on the bench to test the gauge.

4) If the alternator voltage is high, what is the likelihood the regulator be repaired at home? I believe the regulator is separate from the alternator on my tractor. Both are quite expensive.

Thanks in advance.

John

 

[picbits]

Check voltage between gauge ground and sensor ground.

The first 50-60 degrees measurement on the sensor is quite a wide range of resistance where the last 20 degrees is only a few ohms (on normal car sensors). Also don't rule out the actual vehicle overheating .....

I'd guess the battery voltage to be between 13 and 14 volts while running.

 

[jpanhalt]

I measured the engine block temp (IR), radiator temp (IR), and sensor resistance* (put a decent ohmmeter on it) and went mowing (90" flail mower) the last hot day. All the temperatures stayed normal and resistance through the sensor barely budged once it was up to normal operating. Backed up the IR measurements with my calibrated fingers. Still, I don't like running an engine without that gauge actually working.

I will check voltage between the cluster ground and alternator ground.

John

*Ran a calibration curve of R vs. T using a stirred water bath. Data from the tractor manual were not the same, but they were in the same ball park.

 

[alec_t]

Can you monitor the voltage between terminals 1 and 4 to check if the 'fixed' resistor lives up to its name?

 

[jpanhalt]

An interesting point. Will do it this weekend. We are expecting 12°C this weekend. Had our first snow earlier in the week.

John

 

[alec_t]

In the second pic there seems to be some discolouration of the plastic at the near end of the resistor. Presumably the resistor runs hot? Dry joint possible, causing higher resistance in the 'cold coil' path to disturb the gauge reading?

 

[debe]

Altenator voltage is generaly at 14.2V. If the gauge does actualy use an external voltage regulator, it will usualy be 8V. Most modern inst panel voltage regulators are 3 terminal 7808. As previously mentioned I would be checking for poor conections & wiring, as it doesn't take mutch resistance to upset the readings.

 

[jpanhalt]

Those pictures are not of my gauge. To get the back f the gauge off, you have to drill out some rivets. This would not be the first time I have seen poor solder on automotive gauges, if that is the problem. The problem, however, is more likely due to higher current than expected. That is why I want to rule out an over-voltage situation first.

Some auto restorers put in a small zener diode and resistor to regulate the voltage. I want to avoid doing something like that until I am sure of the cause. This is a 2003 model year tractor with less about 550 hours. So, it is relatively low time.

John

 

[alec_t]

The problem, however, is more likely due to higher current than expected.

As I understand the operation of a cross-coil gauge it is the ratio of the currents in the 'hot'-side and 'cold'-side coils, rather than the absolute values of those currents, which determines the pointer position. So if it were not for self-heating of the sender by virtue of the hot-side current the gauge would be fairly immune to supply voltage changes. A higher-than-normal supply voltage would however result in the sender resistance dropping excessively through self-heating, giving a high gauge reading. Equally, a poor connection or dodgy (high resistance) resistor in the 'cold'-side branch of the gauge would give a high gauge reading.

Edit:
Is the black cylindrical object in the right-half of the linked pic likely to be a v-reg?
https://upload.wikimedia.org/wikipedia/commons/6/6f/Cross-coil-gauge_car-electronics.JPG

 

[jpanhalt]

That is an excellent explanation. I was suffering from the "eyes wide closed" syndrome.

I had discovered last week that the insulation and some of the wire strands of one of the wires from the instrument harness to the connector on the display module had been sliced partially, apparently in an attempt by a service technician (using that term loosely) to make a connection. That wire (HM67) happens to be the ground for that portion of the instrument cluster (tach, fuel, and temperature) and goes to the starter mounting bolt. The remaining strands in the wire looked a little frail. It was on my list of things to check out today. Why do such bozos always make the nick close to the connector so it is much harder to splice in a repair? IDC connections are nice and have a place in life, but not when they are done with a hatchet and then left open to corrosion and flexing.

I now have a mission, once it warms up. Will give an update later.

John

 

[alec_t]

Why do such bozos always make the nick close to the connector so it is much harder to splice in a repair?

Probably for the same reason plumbers put connectors/valves in inaccessible places where there's no room to swing a spanner or pipe-cutter.

 

[jpanhalt]

Well, after much testing, including setting up a test bench with a power supply and a stirred, heated water bath, I could not isolate the problem. The meter read high both when it was plugged into the display and when it was isolated by itself. All resistances were about the same as reported Tractor By Net. Changing the supplied voltage to it from 10V to 15V made no difference in reading. So, I broke down and bought a new meter ($100). It seems to read OK, but it is a bit cold here, and the tractor was not under load. The needle stays just a bit below the center mark.

Here is a picture of the innards. The fixed resistor is labeled 56 Ω, and that is what it tested to be. All solder joints were shiny. One note, when I drilled out the rivets (1/16") to open the case, the needle just fell off. My thought now is that the real problem may just have been a loose needle that slipped about 1/2" (arc length) from where it was supposed to be.



Maybe the best solution would have been to just reset the needle and watch temperatures carefully with an independent gauge for the next hour.

Thanks all for the suggestions and help.

John