Auxiliary fan low speed troubles

[Inkling]

The aux fan won't run on low speed in my 1996 318i, I've checked the main culprits and so far nothing, so I'm wondering if anyone has any advice on how to find the exact problem.

The high speed works when I jumped the fan switch connector, but not the low speed. No click from the relay.
So I tested the low speed relay itself with a multimeter: all terminals behave as expected, applying power triggers the switch just fine. Also I tried plugging in another relay of the same type and it was the same deal anyway.
Additionally, jumping the correct plugs inside the relay socket does indeed make the fan run at low speed.

Ok... so for some reason the relay isn't being powered when the switch gets triggered. The tricky part is that the low speed wire on the fan switch connector DOES have a voltage running through it when connected to ground, a bit lower than the high speed wire connected to ground (difference of maybe 1V). I looked up the electrical diagram (below) to see the reason why - the circuit runs through the ECU/DME. Removing the F16 resistor doesn't remove the voltage from the low speed wire (but does from the high speed one), confirming that the ECU is supplying power to it from elsewhere, and the circuit that actually switches the relay is separately controlled by the ECU.

So, someone please confirm if I've got this right - as per the diagram, the cause of the issue must either be the ECU, the wiring between the ECU and the relay, or the wiring between the fuse and the relay, correct? If so, what's the most practical way to go about accessing and testing these things? Thanks.

[Inkling]

Pulled out the ECU/DME out and did a bit more testing, there's continuity between the relay and pin 57 on the ECU plug (as per diagram), as well as between pin 62 and the temperature switch (which I expected given the earlier voltage test). There's also continuity between the F16 fuse and the relay.

So I guess the only thing left is the ECU? Is there any other input signal required for low fan speed? Otherwise I guess I'm gonna have to open the thing up and check all the connections/components - are there any common problems in particular I should be looking for?

So much for my hopes of an easy fix.

[KevinMullins]

For kicks... jumper pins 30 and 87 in the normal speed relay socket and see if the fan comes on (low speed).

EDIT: Nevermind. Just re-read your post and it seems you've already done that, which eliminates the fan resistor.

- - - Updated - - -

As it seems you already know, if you jumper pins 1 and 2 on the dual temperature switch the fan should come on low.
Otherwise the fan will only come if the temp is right or the AC is on, so do further testing with the AC on.
With AC on there should be a ground signal triggered from the ECU, pin 57 going to pin 85 of the relay. You can check this by doing a continuity test at pin 85 of the relay socket and any chassis ground point.
If there is no ground signal from the ECU I would suspect the ECU itself IF the AC is actually coming on as it should.

[flyfishvt]

I think you have a damaged wire somewhere between the radiator temp switch connector and the dme. I had this problem with the ground wire. Yours appears to be a problem with the low speed wire. Those three wires run from the switch back across the top of the radiator then back up to the rest of the car. The ground wire in my car had 2 breaks in the section running along the top of the radiator.

The radiator temp switch does not supply power to the relay. Neither does the dme. They supply a ground to the relay. Look at your wire diagram. Power comes directly from F16 to the relay coil. The only thing missing is a ground for the other side of the coil. You've already confirmed that the relay closes when the coil gets ground and power and you've confirmed it's not closing when you jump the switch connector because you don't hear a click. All your other tests confirm that all fuses and other parts of the system are working. The only thing left is that the relay is not getting the ground it needs.

Trace that low speed wire back and use a sharp probe or a needle to pierce the insulation. Then touch that probe to ground. I bet you'll have low speed. If that works then the easiest solution is to splice a new piece to bypass the break. In my case the breaks were inside the insulation. They were not visible. Judging from the results of your tests I think this is the only solution. I really doubt its a problem with the dme

[KevinMullins]

The high speed works when I jumped the fan switch connector, but not the low speed.

Can you be more specific on which "switch" you are referring to?
Double temperature switch? Jumpering pins 1 and 3?

High speed side activates ground directly to the high speed relay solenoid.
Low speed side activates ground which has to go through the ECU first before it gets it's the relay solenoid. Or it gets activated directly from the ECU when the AC is on. (common component in both low speed cases being the ECU)

The ground (distribution ground) should be fine if high speed works when jumpering the temp switch, and it appears he checked continuity in between each component starting at the relay socket working back towards the dual temp switch.
This would make it seem as if the "wiring" is good.
I'd like to see him stick a jumper in pins 1 and 2 of the temp switch and then check to see if there is ground at the relay socket (guessing not) and work backwards from there checking for an actual ground signal. (ie. just continuity between test points and the chassis ground is fine)
If there is ground continuity at pin 62 of the ECU, but not at pin 57 of the ECU...the only thing in between is the ECU.

If that makes any sense. I'm much better at physically doing the electrical stuff than I am explaining it in text.

[flyfishvt]

If you read his thread again you'll see that he jumped the low speed relay sockets. That is not the same as testing for continuity between the rad temp switch connector and the relay or dme. I bet if he did actually test for continuity between that rad temp switch connector and the dme he would not have continuity. I still believe its a broken wire at the radiator

[KevinMullins]

If you read his thread again you'll see that he jumped the low speed relay sockets.

And the fan came on (low). So that means the fan is good.

That is not the same as testing for continuity between the rad temp switch connector and the relay or dme.

Pulled out the ECU/DME out and did a bit more testing, there's continuity between the relay and pin 57 on the ECU plug (as per diagram), as well as between pin 62 and the temperature switch (which I expected given the earlier voltage test). There's also continuity between the F16 fuse and the relay.

All of this checks is continuity of the wiring.
Now if he were to go all the way from the relay socket and the temp switch, then yes, I would agree the ground would be lost.... but "where" is it lost?

I'm saying test continuity between all of the associated test points and a known chassis ground.
This will narrow down where the lost ground is.

One probe on a chassis ground ...solid jumper pins 1 and 2 on the temp switch...with the other probe go from pin 2 on the temp switch, then pin 62 on the ECU, then pin 57 on the ECU, then the relay socket (which we already know won't be there).

A known chassis ground would be the same thing as the ground point on pin #1 of the dual temp switch. We know that particular ground point is good because the fan comes on high when jumpering pins 1 and 3. So jumper pins 1 and 2 (low speed side) and probe until the ground "disappears" while following the low speed circuit all the way back to the relay socket.

[Inkling]

I did as you said, except I don't know how to test pin 57 (with the ECU connected) without pulling apart the wiring bundle and stripping a wire, but I can tell you that there is a voltage coming from that pin with the key on.



So it does not seem to be a short from the fan temp switch to the ECU, and voltage is definitely getting to the ECU via the F16 fuse.

As an aside, for good measure I tested the fan switch itself in hot water and it switched properly at about the right temps, but that's not really relevant.

[KevinMullins]

Dont know what happened to your last post, but I got it via email...but if you can find the red/white wire coming from pin 57 off the ECU you can use a simple sewing pushpin or the likes to pierce the wire and use the metal pin as a probe point so you dont have to strip any of the wire casing.
Try with and without the relay in place to see if ground is coming from the ECU. I wouldn't think you would be reading power from F16 through the relay to the ECU on that line since there is a diode in the relay.
Are you sure you tried a second compatible relay earlier on? (Same type and diagram on it and all that)

- - - Updated - - -

Oh, and good on ya for double checking (boiling) the temp switch. I've done that a few times myself. Just be sure your coolant is properly bled after removing it otherwise the switch won't get the proper coolant level in order to work.

[Inkling]

Ok so a few more tests later and it must be something to do with the ECU. The input signal from the temperature switch is definitely reaching it, and jumping pin 57 to ground (with key on) activates the fan on low speed, so that information along with all the other various tests I've done confirm that the ECU just isn't providing that pin with a ground for whatever reason.

So I opened up the ECU, did a visual check for anything obvious (charring etc.), tested for dry solder joints on the appropriate pins, tested that the pins were making proper contact with the connector (just to be certain I tested continuity from the relay and the temp switch to the circuit board itself) - nothing out of the ordinary.

Oh and to answer the earlier question, no the air conditioning does not activate the fan.

As I see it these are the possibilities:

1) There's a problem somewhere along whatever ground wire is connected to the ECU and meant to be grounding the relay. I'm not sure which pin this would be on the ECU, though the electrical manual narrows down the options at least. This obviously assumes that the intended ground is not the pin 62 temperature switch wire itself, which I think is a fair assumption because the AC has to be able to trigger the same relay and turn on the fan independent of the temp switch.

2) The ECU requires some sort of additional signal/input/power/whatever on another pin, but it's not getting it for whatever reason.

3) The ECU has a faulty component.

Any suggestions on how to proceed?

[flyfishvt]

To make this system work you need a working relay, power to one side of the relay coil and ground to the other side of the relay coil. You have power and you have a working relay so now you're missing the ground.

Gowing backwards from the relay coil the ground goes from Pin 85 of the relay to pin 57 of the dme. Test for continuity between those 2 points.
Proceeding backwards it goes from Pin 62 of the dme connector to the low speed wire at the temp switch. Not the ground wire but the low speed wire. It should be black/green. Test for continuity between those 2 points.

Disregard the voltage you see at pin 57. That is voltage that is comming from the coil. Its normal.

If you have continuity between 57 and the relay and if you also have continuity between pin 62 and the temp switch connector then the problem is inside the dme.





Check for continuity between pin 62 of the dme connector and the ground wire at the temp switch connector.

[Inkling]

I feel like such an idiot - it turns out that the low speed on the fan works fine (when grounding the wire or turning the AC on)... when the engine is running.

Part of the reason I didn't realize this sooner is because I had the radiator removed and cooling system drained (due to an unrelated issue) until recently. Even so, I never really thought to test it with the engine running because everything I read said it should work with the key in the 'on' position (yes the key was turned as far as it could go without actually attempting to start the engine, so it wasn't something dumb like not turning it far enough).

So why the discrepancy? Obviously the ECU requires something extra that only occurs when the engine is running, but why is my car apparently so special compared to other people's?