Technical Cooling fan on from engine start (188 1.2 16v)

[AdamScriv]

Hi folks,

My 1.2 16v Punto mk2b has decided it wants the rad fan to run at all times while the engine is running. I did a quick search but couldn't find anything conclusive; one person suggested a damaged cable between the relay and the ECU but I don't *think* that's my issue (see below). Car is single speed fan and no air conditioning.

Anyone else had this?

Things I've looked at so far....
- Disconnect battery overnight
- Cleaned up all the grounds under the hood (battery to body, gearbox, body to earth block behind headlight, ECU to loom)
- Checked battery voltage
- Swapped the fan relay for a different one (no change)
- Disconnected/cleaned/visual inspection of ECU plugs & connectors under engine bay fuse/relay box
- Used a meter to check for high impedance to body/gearbox earth from fuse box (basically seeing 0 ohms)
- Used MultiECUscan to check for fault codes (none present apart from the usual side light warning in the body computer)
- Used MES to test the fan with the ignition ON but engine stopped, which responds correctly. I.e. the ECU has control of the fan, so I think on the output side of the ECU the wiring/relay/fan are probably all fine.
- Looked at the live data from OBD using MES, which says the ECU has turned the fan on (speed 1)
- Checked coolant temp and air temp sensor readings are normal
- Had a look at the air-con demand parameters in MES (the two I found said "NO" plus, the car doesn't have a/c anyway)

So far as I can tell then, the cooling fan electrics themselves are all working fine, the temperature readings are normal and there's no fault codes, yet for some reason it feels like the ECU "wants" the fan on at all times while the engine is running. Beyond engine temp, air con request, bad data from the temperature sender or a CAN issue, why else would the fan run? Am I missing something stupid?

I'm starting to think maybe the ECU has actually failed, but in my experience that's pretty rare and it's normally an external influence that gives the impression of failure. I'm a bit lost with it now. Gonna give it a week then I'm drilling a hole in the dash and putting a manual override switch in

Any help/thoughts greatly appreciated!

 

[AdamScriv]

Oh wait, I'm assuming when polling the fan status with MES/OBD the readout is what the ECU has decided to try and do not a measurement of the real world state? (i.e. the ECU has, for some reason, deliberately set the relevant output pin low to energise the relay, and that "decision" is reported as Fan = ON)

But maybe if the ECU's cooling fan pin was externally grounded by bad wiring and pulled to 0v then the "ON" status I'm seeing in MES could be a reflection of that....? That would seem more clever than I'd expect.....!?

 

[AndyRKett]

Cooling fan will run constantly if the ECU feels it needs to protect the engine, for example if the coolant temperature sensor has failed. So consider this a symptom and not the illness. 'What else is wrong'

 

[AdamScriv]

Thanks @AndyRKett !

Based mostly on the fan test successfully turning the fan on & off just fine while the engine isn't running, I am indeed currently assuming the fault to be on the "input" [sensors] side of the ECU rather than a wiring fault on the "output" to the fan relay or beyond.

I guess my issue is I don't know what else to check beyond supply voltage & negative grounding being good for the ECU to work with, and the coolant/ambient temperature readings being normal. Oh and no false air-con request.

I wonder if the coolant temp sensor could fail in a way that doesn't show up in temperature readings (on MES or instrument cluster) but is unstable or noisy enough for the ECU to put the fan on, without error code?

 

[AdamScriv]

Just to report back.... (and so I can find this thread when the problem re-occurs)

Really boringly in the end, despite there being no visible bad temperature readings (ran it up from cold and graphed it in MES) and no fault code, I just unplugged the coolant temp sender which made the ECU even more upset (i.e. throw an actual error code and run the fan like it should in that scenario) and then when I re-connected it and cleared the now-present fault code, everything worked fine.

So, to summarise my experience so far: The fan running failsafe mode doesn't always show a fault code. Forcing a fault code and then clearing it may help. Leaving the ECU unplugged overnight(s) did not in my case.

Unless it was one of the corroded earths which caused a random unseen glitch, I doubt I've cured the root cause, but I guess will have to see how it goes. I really don't feel like the connection between the plug and the coolant temp sensor is/was ever the problem itself. Drove the car on 5 short trips today and so far so good.

 

[judderbar]

interesting fault. My take on this is since the ECU knows what the sensor should look like, if there was a fault in the sensor or wiring it would produce the circuit malfunction code.

 

[AdamScriv]

Thanks @judderbar , agreed, that sounds sensible.

That being so, the reason it decided to jam the fan on in a failsafe-like state but without throwing a fault code, nor the real time temperature data looking bad, is still a bit of a mystery to me.

I imagine one day it'll become clear, probably in the pouring rain, 300 miles from home

 

[judderbar]

Thanks @judderbar , agreed, that sounds sensible.

That being so, the reason it decided to jam the fan on in a failsafe-like state but without throwing a fault code, nor the real time temperature data looking bad, is still a bit of a mystery to me.

I imagine one day it'll become clear, probably in the pouring rain, 300 miles from home

did you check you got a code with the sensor unplugged and you turn on the car with the usual fan running when unplugged?

The sensor can be tested

-20C 15971 ohm
0C 5975
20C 2502
40C 1152
60C 576

On my car the gauge begins moving at 50C

 

[AndyRKett]

This may not be the issue or it may be a coincidence and is the issue, but sometimes a high current relay like you get on a fan circuit will arc as it is cycling on and off and weld itself (literally) in the closed position. Messing about with the car pushing power on and off through the relay combined with any mucking and knocking things about under the bonnet could have just unstuck it.

I relay that is stuck on will not show any error codes as often the ECU won't see that side of the circuit, well not on a fiat anyway.

 

[AdamScriv]

Thanks for the info!

So, the fan has now come on again, only twice, but both times cleared itself after switching off and restarting the engine (unlike when I posted originally, where it was stuck on for at least 10-20 restarts before I starting hacking at it...). So problem not properly solved yet.

I'm not ruling anything out, but I don't *think* it's a power relay issue as (a) I have swapped relays around and (b) with the engine stopped, MES can always start and stop the fan OK. I'd have thought if the relay jammed it would be unresponsive to the ECU regardless of engine idling or stopped state, but I maybe I need to do more testing.

Just measured the sensor between very roughly 40 and 60 degrees, and it tallies up with the resistances @judderbar gave. Maybe I need to measure it over a bigger range, from cold. Or maybe I could jam a fixed resistor of around 800 Ohms in the plug and see if the car will run well enough to drive it around, and see if the problem re-occurs. I guess it would run lean on cold start and rich after that, but I suppose this would prove whether it's simply an unstable sensor, or some more nuanced/complex issue I need to look for.

 

[AdamScriv]

did you check you got a code with the sensor unplugged and you turn on the car with the usual fan running when unplugged?

Sorry forgot to answer this bit: Yep, with the sensor unplugged I get an error code and the fan runs.

Unlike when the sensor is plugged in, where the fan runs continuously now occasionally jams on from cold until the next restart, but there's no error code* or bad looking temperature data on the OBD.

*edit: Actually that's a lie, I did see a P1135 code at one point when the fan was on, which is new to me since mucking about with the temp sender, and the car has sometimes shown one of the common data bus error ones (U????) for the last 10 years or so. But P1135 was a new development and I suspect a red herring as it has just shown up once, weeks after the fan fault developing. Since clearing it it's not come back, so unless a correlation appears, I'm putting that down to me disconnecting and reconnecting earths, sensors, battery etc.

 

[AndyRKett]

But P1135 was a new development and I suspect a red herring as it has just shown up once, weeks after the fan fault developing.

Fuel air mixture fault, unlikely to have any effect on what the fan is doing, the fact it comes and goes, shows that it isn't a major issue.

 

[AdamScriv]

So, just to say I've not forgotten to report back, just, not had any luck.

In desperation I swapped the seemingly OK coolant sensor (for a 7 quid ebay job + a coupla quids worth of coolant I thought it was worth a punt as I don't know what else to do) and for a few days it was OK; but I think it's just random.

Then the exhaust hanger fell off the back and the car was off the road for a week while I ordered the hanger and found time to get under it and sort that. But when I started it up the exhaust was blowing from the back box, and worse, the cooling fan came straight on. Soooo, no idea what to try next. Scrap yard maybe.