Faulty Temperature Guage

[TeamSid]

I've just fitted two 100ohm resistors in parallel (52ohms) which put the needle onto the white line, the top edge of the needle just touching the highest edge of the line. I then fitted the resistors to the connector for the sensor which gave the same results. I'll get some more resistors at work on Friday, but as the calibration appears to be OK I'm starting to wonder if the engine has excessive heat in the thermostat area. The van had a genuine VW replacement engine about 37K miles ago and also had the metal water pipe which runs alongside the engine replaced. The van itself has covered 198K miles with a full service history. Would the exchange engine come with a water pump fitted, because if not then the pump will be the original as I can see no evidence in the history of it having a new one at any time. Could the pump be the problem maybe?

[kevtherev]

Generally water pumps fail at the bearing and leak.
Recon engines generally do not come with pumps fitted

cooled coolant enters the thermostat housing once the stat opens so it would be the coolest point of the cooling system, the big pipe contains this cooled coolant which is distributed by the water pump.
The hottest point is where the coolant leaves the engine.

[TeamSid]

I've been out tonight and had a look around the cooling system again. I removed the serpentine belt and gave the water pump a spin which appeared to be free to turn with no signs of leaks on the shaft or sideways movement. I did notice that the two lengths of hose which connect the 12mm(ish) metal pipe which runs from the top of the pump to the underside to be swollen out and difficult to press between finger and thumb. I'm not sure if this has any relevance, or what purpose they serve, but I got impression that they could be under pressure? When the engine is running, the hose to the junction unit gets very hot but seems to be not so hot immediately from the junction towards the radiator. As Kevtherev says, it's much cooler at the thermostat once returned from the radiator.

[TeamSid]

Following a few days of clearing my head I've had another look at the beast. I changed the water pump (which didn't need doing afterall), bled the system and everything on the mechanical side is tiptop. I checked the calibration of the guage again which appears to be good, then connected a good spare temperature sensor to the van's harness and dipped it in a kettle. I boiled it up and the needle rose as the temperature increased. By the time the water boiled, the needle had reached the first white dot towards the max position on the guage. I'm assuming this is too high and it should be closer to the mid position, but please tell me if I'm wrong. I then disconnected the van's harness from the sensor and made myself another harness and wired it directly to the clocks. This gave the same result which proves that the van's wiring is good. My opinion now is that something is dodgy on the circuit board which is hidden behind the speedo. I'll have a look at it tomorrow, and maybe see if another old VW at my local scrapyard shares the same board. I'll keep you posted, but in the meantime, feel free to give any other ideas.

[AngeloEvs]

I doubt it the PCB. The test you did with the sender seems to give the wrong result. Rather than connect the guage, bring the pan to boiling point and suspend the sender (so that it is in the centre of the water and not in contact with the pans structure) but use a multimeter and measure the resistance at boiling point - it should be around 100 ohms or slightly less. Your test seems to suggest the resistance is lower than it should be. I have a spare one somewhere, if i can find it I will do the same test and give you my reading - its been a while since I performed this test and I cannot remeber my exact readings but the value I obtained would not have the gauge needle far right.

[TeamSid]

I got myself another PCB thanks to my local scrapyard but as you say, it made no difference. I'll do the boiling water test again tomorrow using an ohmeter and report back. Thank.

[TeamSid]

Well, I've now carried out the boiling water test using a more controlled method and the resistance while the water was boiling averaged 97 ohms. This ties in with the guage sitting in the mid position when a 100 ohm resistor is used.
How would I go about rigging up a system where I could remove the guages from the van and using a 12Volt battery, my sensor, homemade harness and a pan of boiling water to check the guage and circuitry? If I could rule them out for sure I could move on to another possibility as I seem to be chasing my tail a bit on this one! Thanks.

[TeamSid]

This morning I ran the engine up to temperature and once the needle reached almost the maximum position I disconnected the temp sensor, fitted my trusty homemade harness and an ohmeter which gave me a reading of late 80s early 90 ohms. I continued to run the engine for another 5 minutes, or so and the meter sat at this reading. My van's currently sorned so I'm assuming if I gave it a run the resistance would possibly increase a little as the water flowed round the cooling system. Anyway, I can now assume that the temp sensor and the cooling system are working correctly and that the wiring to the dash is OK as I've already ran temporary wires. I'm fairly sure that guage setup is faulty somewhere but as I've already tried another temp guage, thermal transistor and pcb, and checked the resistors I'm not sure what to check next. Ideally I could do with a complete set of clocks to try on the van but they appear to be difficult to get a hold of. Anyone got a set spare? Cheers.

[AngeloEvs]

I always test/repair Dashpods on a bench and you can do this without too much trouble using a Bench DC supply or a 12V battery. The best way to do this is to get the edge connector from a scrap post '86 T25 with approx 6 inches of the loom. However, with care you can use small croc clips to connect directly to the appropriate copper edge connectors always taking care that the croc clips are only connected to the relevant connector and that the POD is well supported and secure.

From the tests you have carried out, it does seem as though the problem is with the cooling system and not an electrical problem but a bench test is always advisable and then followed by an in situ test.


You have verified:

a) it is not the PCB

b) it is not the sender (the obtained readings are correct at around 100 ohms at Boiling Point)

c) it is not the gauge (you have tested its calibration using a 100 ohm resistor and two 100 ohms in parallel to give a 50 ohm load)

d) You have removed the Module 43 (to eliminate that this is not loading the Temp Gauge)

e) you have checked that the regulator is giving 10V output to the fuel/temp gauges.

f) the condition of the wiring (as you are using a fabricated loom to monitor sender resistance in situ)




My next step would be to remove the thermostat - re - bleed the system - check that the Antifreeze coolant ratio is correct and measure the sender resistance directly whilst it is connected to the gauge. To do that - remove the module 43 and place the probes in the vacant module 43 sockets as described in the wiki.

If the reading obtained starts to drop below 100 ohms with the stat removed its time to carefully examine mechanical/timing/coolant possibilities.

[TeamSid]

Just to verify, when I connected my ohmeter to the the temp sensor when the engine reached running temperature, I achieved resistance readings of late 80s/early 90s. I take it that this is too low? I'll take the thermostat out over the weekend out and see what readings I end up with.

[AngeloEvs]

If the sender resistance is around 100 ohms in boiling water at 100 degrees and you are getting a lower figure with the sender in the housing either the coolant temp is higher than 100 deg or there is some sort of additional resistance loading in parallel with the sender that hasn't been identified by any of the tests you carried out. You seem to have covered all of them though.

The thermostat is probably an 88 deg type and the coolant system is pressurised to raise the boiling point. Your reading of upper 80's to low 90's with the sender in situ and engine running isn't way off but that reading isn't low enough to have the needle far right and triggering the coolant LED. It has to drop to around 50 ohms to trigger the coolant led. Unfortunately, it is very diificult to test the linearity of the senders resistance for temperatures above 100 deg C but at the end of the day the stat should be opening at 88 deg and coolant should be circulating through the system.

Does the Rad fan kick in? The fan is independent and relies on its own temp sender in the rad switch. If the switch is working correctly and the Rad has coolant it should trigger the Rad if overheating but this also relies on the coolant circulating through the system. Water is a poor heat conductor and any loss of circulation could result in a cool Rad and excessively hot thermostat housing. My specialism is the electrical side so may be someone with an in depth knowledge of the cooling system could add their opinions or thoughts to your problem, especially since the T25 cooling system is more complex than your typical water cooled engine regards pipework, correct flow direction, air locks, blockages, etc.

[TeamSid]

Yes, the fan does work as it should, as it kicked in a few times last summer whilst in slow moving traffic after a long motorway run. I've not really done much to it today due to the miserable weather, but I did notice that the radiator heats downwards from the top which makes me think that the hoses wre fitted the wrong way round when the radiator was changed way back in 2001. I'll put a queary about that in the mechanical section.