g5 DC 2.3Ghz One fan runs at max rpm when idle, but Hardware Monitor doesn't spot it

[kifkif]

Hi guys, me again. I'm trying to solve a fan problem, and whittled it down to this:
When idle, and right from the moment the on button is pressed:
One fan is noisy - because it's rotating at, or somewhere near, maximum rpm (circa 3200).
However, Hardware Monitor reports it as operating at minimum rpm (circa 500rpm).
Hardware Monitor reports all fans as operating around minimum rpm when computer is idle, which seems totally correct, apart from this errant lower front cpu intake fan.
I've tried a replacement fan unit - it does the same thing.
I've tried just about every conceivable obvious thing - stripping down all connections and RAM, even erasing the HD and starting with no OS installed.
I've reset the pram and the other thing.

If Hardware Monitor can't spot it, what does that mean/not mean?

I'd like to run the hardware test, but don't have the original discs that came with the computer, only an apple leopard install dvd which doesn't seem to do it.

 

[Hrududu]

Take it to an Authorized Service Provider so they can run the thermal calibration on it. More than likely it just needs to be re calibrated.

 

[drorpheus]

I was going to suggest replacing the front inlet fan (the big one with 2 fans and handle) which is Apple Part: Front Inlet Fan 922-7085, but you say you already replaced it with a new/different one and still the same result. I don't want to sound dumb but are you sure it is the right fan that's running too fast, I mean basically you can see that the fan is spinning faster than the other one, and its not possibly one of the rear exhaust fans, either way if you replaced the fans and the same result, I would suggest finding ASD 2.6.3, I believe this is the service diagnostic disc for the Late 2005 Powermac G5's, the Apple Hardware Test will not really tell you anything more than hard drive failed or ram failed, I had a Dual 2.0 G5 that would freeze up on booting, and AHT tests passed everything, though ASD confirmed the processor was cooked with TAFS out of range, I don't have ASD 2.6.3 otherwise I'd send you it, but basically its a bunch of tests for your mac to determine what's failing, it has fan tests that will see if the fan fails or not, the only thing I can think if you replaced all the fans and still get the same result is that the processor is starting to run hot, but if that was the case, I would figure your hardware monitor would show the fan running at 3200 rpm. Either way ASD will tell you what's wrong if any of the 7 fans, you can buy most of the fans on ebay for $7-$20

 

[drorpheus]

Yeah, also forgot to say that ASD 2.6.3 will let you run the thermal calibration and save you the $30-$50 service fee (if you can't find the disc), I didn't think about it but, Hrududu may have hit it right on the nail with it needing to be thermally recalibrated.

 

[kifkif]

Hi, Thanks loads for all the pointers! You saved me wasting £10 on a hardware test CD I was just about to buy.
It's definitely that fan. I tried stopping it with my finger - ouch! but then all is very very quiet.
I just googled ASD 2.6.3 and
I've been downloading it from rapidshare (am I allowed to say that on this forum?) all day - I've been calling it rapidshnail, as I'm doing it on the free, there are four files, and they make you wait two hours between downloads. If it works I'll be ever so grateful to them though.
I still haven't worked out how you put all the files together to make a CD, or what you do with them, hopefully that will become clear.
Also, I emailed the problem to Marcel Bresink, creator of Hardware Monitor, asking if he could help understand what's happening. He replied this morning:

On some particular G5 models, Mac OS X is unable to "see" the actually measured fan speed
(it is seen by the power management hardware only). In this case, Hardware Monitor shows
the speed the cooling management of Mac OS X is "sending" to the fan controller. Under
normal conditions, the sent and received speeds will always be the same, so this does
not matter. If you experience a difference between these speeds however, this would
indicate that something with the fan controller hardware is not working correctly.

I would recommend to reset the System Management Unit (SMU). Please see Apple's
instructions at http://support.apple.com/kb/HT1436 for further information. If this
does not help, you might like to run the Apple Hardware Test which came with your
system DVD and to check the fan cable (which contains the sensor line).

I hope this information is helpful to you. If I can be of further assistance, please
let me know.

Best regards,
Marcel Bresink

So there we have some words from the great man himself - fantastic!
I have already tried resetting the SMU, but will do so again.
The fan cable is worth checking if other stuff fails, but I've tried replacing the 2-fan unit with no luck, so the cable on that part must be OK.
Let's hope the thermal recalibration does the trick, hey?

Only another 52 minutes until I can download the final part of ASD 2.6.3 ...
I will post results here as soon as have any. That maybe a little while as I'm in the UK and it's getting on for bedtime now..
Thanks again so much to everyone for your help!
kk

 

[kifkif]

Well, I managed to create the ASD disc and run it on the computer.
Results:
The fan in question failed the diagnostic test with this message:
(Testing Fan - CPU B INTAKE)
ERROR - Fan IS SPINNING FASTER THAN IT SHOULD 8

All other tests on all other fans/sensors and everything else: passed

Then when running thermal calibration:

ERROR CPU0 - CPU Intake Fan speed error
Check that inner door is secure and in place then rerun calibration

ERROR CPU1 - CPU Intake Fan speed error
Check that inner door is secure and in place then rerun calibration

Needless to say the door, which seems to fit perfectly well, has been removed and replaced several times, to no avail.

Unfortunately, it doesn't say what to do if the error recurs each time you rerun thermal calibration, or what to do with a spinning too fast fan.

It's still only intake-B spinning so fast, despite intake-A next to it failing the second test. Intake-A is most likely running very slightly fast because of the fast air from the bad fan next door.

Any further ideas warmly welcomed..

 

[kifkif]

Just to tidy this up - I sent the last info to Mr. Bresink, and, based on the first error message, he said:

this would indeed confirm a hardware problem resulting in a mismatch between the actual speed and the speed "sent" by the fan controller which is the RPM value displayed by Hardware Monitor.

So - bad news.
I also asked him where this problem would physically be, and:

From the point-of-view of the software, this would be the System Management Unit (SMU)

which is basically a trade secret of Apple.

So, now I know, all those unfinished threads on fan problems probably end around here too..

Best of luck everyone else!