Coolant Temperature Gauge Misbehaving - Looking for Input / Thoughts

[christenkonrad]

Hello everyone,

I recently picked up a 313k mile UrS6 Avant that was non-running for the last 8 months to a year. I've been doing my best to bring back several of the broken things inside the vehicle.

One thing that I've been struggling with is the coolant temperature sensor / gauge on the cluster. So I'm going to walk through what I've tried, what weird behaviors I'm having, and say what I think is going on in hopes that someone can validate or give me a direction to move into.

So first; let me explain what I understand.

MFTS sensor is the one that communicates with the gauge cluster.
Grounding the pin (Pin "T", which is a brown/yellow wire below the unpopulated pin on the 3 pin MFTS), should cause the gauge to climb to maximum.
If the gauge doesn't climb to maximum the problem is in the wiring or the cluster.

What I've tried:
-Pulled off the connector from the sensor
-Found corrosion
-Cleaned corrosion
-applied electric grease to pins
-tested grounding the pin, had no success for the gauge pinning itself to hot, but had the needle rise from "nothing" to the first starting dot.
-removed and disassembled cluster
-reflowed solder joints on all connectors across the cluster
-reinstalled in vehicle
-tested grounding sensor again, had intermittent results, had some occasions where the needle would rise to the "starting point" but go no further, had some occasions where the needle moved UP to the point where I said "yep it's working", like halfway up the gauge. But couldn't replicate.
-thought that maybe I was having an issue with the contacts not contacting the wire I was using, tried different wires, tried removing the pin from the plastic connector and cleaning the heck out of it. No positive or consistent results.
-I checked continuity / resistance between the pin at the connector and the pin at the connector at the cluster (pin 3 & 16, both had good continuity, less than an ohm or two)
-removed the cluster again, tried specifically reflowing pins 3 and 16 (I know 16 is the right pin but I figured what the heck) and made sure to introduce some additional solder and ensured that I didn't short anything.
-reinstalled the cluster
-tested again this evening.
-in this test, I got positive needle movement, the needle was climbing and I was standing there watching it, all of a sudden the needle starts falling, and falls down below the "starting" point and basically doesn't move again for the rest of my troubleshooting.
-I have a video of this test, nothing is moving. Not the harness, not the car, engine is off, cluster isn't being shaken or touched. Nothing is moving.

So what do I think based on this?

1) I think the harness connection between the MFTS is OK
2) I think there is no intermittent connectivity in the harness
3) I think it's possible the MFTS is failed, but that's not contributing to my problem
4) I don't think any of the solder joints are broken between the cluster and the gauge. Nothing I could see anyways from an uneducated look of how the signal might flow across the board.
5) I think that the power to the gauge might be going out or the gauge itself might be dying.
6) I have not tested trying to ground the pin on the back of the connector to the cluster while the cluster is plugged in, but I have anxiety about possibly breaking the connector or something trying this and I think based on my evidence that isn't my issue.



Here are my questions:
1)
Can someone confirm if the needle has a "starting point" when the gauge receives power and enters "normal operation"? What I mean by this is, experimentally it looks like there's a "I have no power" state for the needle, and there's a "I have power but the temperature is too low to raise the needle" state. I want to confirm if this is a real state or not.

2)
Has anyone experienced a failure where the gauge itself is malfunctioning? If so, can you share your troubleshooting steps for that? Or point to the article that undoubtedly exists somewhere in the internet?

3)
Assuming my gauge is actually the culprit, is there a part number that I can hunt for for a new gauge? I really really need to have a functional coolant temperature gauge in the car and I am thinking that finding an operable cluster to buy isn't going to be cheap and is going to be hard to find.

I've been trying hard to use the multitude of resources that are out there for these wonderful wagons but I feel like I've hit a wall. So I apologize if there's a forum post somewhere covering this exact issue. It's really killing my vibe not to be able to roll around in it. (The radiator mounted sensor that turns on the fans is also failed, and I've been waiting until I could diagnose this issue to not have to order parts multiple times and pay shipping multiple times).

Thanks so much for any and all input.

[newsh]

There's a voltage regulator for the temp and fuel gauges, on the S2 it's plugged/bolted to the back corner of the cluster, corrosion on connections loose bolt or bad reg will cause similar symptoms, if it's that then the fuel gauge would be playing up too.

[JayCab]

I had much the same symptoms as you here...

WARNING WARNING: You mention as a side note your rad fan switch has failed. That's a far more serious problem than the temp gauge and your engine will overheat without fixing this. I would do this and confirm a working fan before even looking at the gauge and MFTS.

For your question 1 on start position:
--> No. The gauges are totally analogue. A working, fully functional gauge will not move one micron when engine is first started. It will not "jump to zero" on Ign and then proceed to read temperature like a new car (i.e VAG Mk4 era forwards). It will not read "90C" whatever the real temp is to provide false confidence. It will go up/down and read real coolant temp between 90 and 105. Over 105 is bad. It will start to move when the thermistor in the MFTS heats into the range. Stone cold coolant is below the bottom of the gauge, hence no movement on start.

You mention you have reflowed all joints to the gauge. Have a look, VERY closely at the ring contacts on the PCB the gauge posts themselves screw against, and their continuity through the board to the tracks on the other side (test at next connection point, don't pierce the track with the probe).

--> There should be split spring washers on the gauge posts between the nuts and the PCB track ring contacts. (Low torque, no craziness, spring washers should do the contact work). Do not be tempted to over tighten or even "nip up". This is <5Nm territory.
--> I had to reflow the actual ring clamps to fix mine, with similar symptoms as yours having tested wiring and changed MFTS.

Wildcard:
You mention you have used "electric grease". Are you sure you have not used "dielectric grease"? This would act as an insulator and increase resistance if applied to actual contacts.

A wackier troubleshooting approach for the gauges is to swap the fuel and temp gauges around (they are the same units with different fascias and calibrations)..
--> If the problem follows the gauge, you have a faulty gauge.
--> If the problem sticks with the PCB gauge position, gauge is fine and it's certainly electrical.
--> Watch out for misdiagnosis here. The act of swapping the gauges can by itself trigger an intermittent contact connection problem to good or bad. Ask me how I know ;-)

This problem is very annoying. I must must have stripped my cluster 5 times.

Hope you make some progress...


p.s: Tip for brittle plastic self tappers that are old: Drive screw backwards first until you hear a "click" before screwing forwards. This seats the screw in the same thread cut before, so it won't cut a new one and (probably) crack the plastic. Snug up only just enough to hold parts or board in place. Apply feathertouch torque.

for your

[christenkonrad]

There's a voltage regulator for the temp and fuel gauges, on the S2 it's plugged/bolted to the back corner of the cluster, corrosion on connections loose bolt or bad reg will cause similar symptoms, if it's that then the fuel gauge would be playing up too.

Thank you for this piece of info! It's excellent to know they are run from the same voltage source. The fuel gauge fortunately came back on the first round of soldering and is not acting up. So I think I can safely say it isn't an issue with the regulator!

[christenkonrad]

I had much the same symptoms as you here...

WARNING WARNING: You mention as a side note your rad fan switch has failed. That's a far more serious problem than the temp gauge and your engine will overheat without fixing this. I would do this and confirm a working fan before even looking at the gauge and MFTS.

For your question 1 on start position:
--> No. The gauges are totally analogue. A working, fully functional gauge will not move one micron when engine is first started. It will not "jump to zero" on Ign and then proceed to read temperature like a new car (i.e VAG Mk4 era forwards). It will not read "90C" whatever the real temp is to provide false confidence. It will go up/down and read real coolant temp between 90 and 105. Over 105 is bad. It will start to move when the thermistor in the MFTS heats into the range. Stone cold coolant is below the bottom of the gauge, hence no movement on start.

You mention you have reflowed all joints to the gauge. Have a look, VERY closely at the ring contacts on the PCB the gauge posts themselves screw against, and their continuity through the board to the tracks on the other side (test at next connection point, don't pierce the track with the probe).

--> There should be split spring washers on the gauge posts between the nuts and the PCB track ring contacts. (Low torque, no craziness, spring washers should do the contact work). Do not be tempted to over tighten or even "nip up". This is <5Nm territory.
--> I had to reflow the actual ring clamps to fix mine, with similar symptoms as yours having tested wiring and changed MFTS.

Wildcard:
You mention you have used "electric grease". Are you sure you have not used "dielectric grease"? This would act as an insulator and increase resistance if applied to actual contacts.

A wackier troubleshooting approach for the gauges is to swap the fuel and temp gauges around (they are the same units with different fascias and calibrations)..
--> If the problem follows the gauge, you have a faulty gauge.
--> If the problem sticks with the PCB gauge position, gauge is fine and it's certainly electrical.
--> Watch out for misdiagnosis here. The act of swapping the gauges can by itself trigger an intermittent contact connection problem to good or bad. Ask me how I know ;-)

This problem is very annoying. I must must have stripped my cluster 5 times.

Hope you make some progress...


p.s: Tip for brittle plastic self tappers that are old: Drive screw backwards first until you hear a "click" before screwing forwards. This seats the screw in the same thread cut before, so it won't cut a new one and (probably) crack the plastic. Snug up only just enough to hold parts or board in place. Apply feathertouch torque.

for your

First of all, thank you for the detailed and incredibly useful information!

I did discover that I was overheating and that the fault was with the radiator temp sensor, I should've mentioned that I'm not driving the car right now. I bought it as a project and I was getting stuff back to life one step at a time which did involve driving it around but then I got an over temperature warning on the cluster and realized the radiator was quite warm and it began my investigation into what causes that. Finding the Rad fan temp sensor was dead. (Relay is operational, power is at the relay, and the fans can be triggered when jumpering the harness from the rad fan sensor input, both the low speed and the high speed that can be triggered from this sensor are functional so I'm currently awaiting the arrival of a replacement sensor to install. I have plans to test via hot water before installation).

I will absolutely investigate the ring contacts on the gauge posts. I hadn't removed the fuel or coolant temperature gauges since they were bolted on and I wasn't trying to mess with what wasn't broken (although it is). I was shooting for a simple reflow job of the connector pins. So I'll give that a look and test the continuity as you have described.

In regards to dielectric grease, yes you are correct. I had initially thought "single point failure, I found corrosion, now that the corrosion is gone I'll try and prevent future corrosion since this is a common issue". But after the gauge not functioning I went through a deeper cleaning process with electric contact cleaner and more scraping to get all the corrosion off (thinking it was still a corrosion issue). So it's all gone now.

I will absolutely consider swapping the gauges and see if that brings operation back, but I think first I will try simply reflowing the one because that will leave my currently operational gauge intact.

I haven't troubleshot my oil temperature gauge yet but it's also inoperable so that will be in my future. (I did find a clearly broken solder joint on the pin that the connector attaches to so I fixed that but it didn't come back to life. However it is attached in a similar way so it's possible it's doing the same thing as the coolant gauge is.

So again thank you very much, I'm super happy to have a troubleshooting direction to move in.

[varia]

https://forums.quattroworld.com/s4s6/msgs/140449.phtml

[christenkonrad]

I had much the same symptoms as you here...

A wackier troubleshooting approach for the gauges is to swap the fuel and temp gauges around (they are the same units with different fascias and calibrations)..
--> If the problem follows the gauge, you have a faulty gauge.
--> If the problem sticks with the PCB gauge position, gauge is fine and it's certainly electrical.
--> Watch out for misdiagnosis here. The act of swapping the gauges can by itself trigger an intermittent contact connection problem to good or bad. Ask me how I know ;-)

for your

Hello, I've returned with new news.

I have troubleshot by reflowing the large bolt connections on the coolant temperature gauge side to no improvement, but I got a real good look at the "needle is climbing, then needle stops and falls" after doing this.

Then I disassembled again and swapped the two gauges (coolant temp and fuel level). Coolant temp gauge (now in fuel level location) climbed up just fine. And Fuel level gauge (now in coolant temp location) performed a similar action of climbing up a little and then falling.

Did you (or anyone) have experience playing "follow the traces" or have any insight into what the operation of the physical gauge is? Is it measuring voltage across a resistor in the cluster and a variable resistance in the sensor? Or is it simply drawing a small current from the voltage source and the motor rotates more based on an increased current? I am looking to play "follow the trace" tomorrow night if time allows but I'd love any input on what I can check beyond that.

I'm wondering if I can tag onto some pins and provide power to the cluster regulator / gauge itself and check the operation somehow. I have access to various forms of power supplies, oscilloscopes, function generators etc.

[JayCab]

I have troubleshot by reflowing the large bolt connections

--> You are looking at the connection between the two brass rings (through which the mounting+conductor bolts of the gauge pass through and the spring nuts press against) and the track on the PCB itself. To see this connection, you need to remove the PCB from the cluster unit and turn it over.

--> I would carefully reflow those, then sand the ring on the other side of the PCB.

--> I would sand the spring clip, and run the nuts up and down the threads of both posts, pressuring in each direction, with the posts soaked in contact cleaner.

--> I would sand and contact clean every link in this circuit that you have touched with dielectric grease.

Then I would reassemble the whole thing.

You have swapped the gauges and the behaviour stays with the location, so you certainly have an electrical problem applying to the coolant side only, with the fuel side being fine and the voltage regulator being okay.

This circuit is literally a series resistance circuit from voltage regulator through the gauge and MFTS thermistor.
--> The gauge needle moves in response to current flow proportional to the resistance of thermistor, which changes with the temperature of the coolant.

[christenkonrad]

Understood I will investigate thanks!