Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

Mister Speaker

macrumors member
Original poster
Oct 14, 2016
32
0
I'm not holding my breath but things seem to be working again. I say that because basically all I've done is take the CPU tray out and give it a hard look (and maybe some duster) as I've done every day this week. But when I reset the SMC and PRAM and fired it up, no fan noise. SMCFanControl shows 47C and 1113RPM on both BOOST fans - does that sound a little high?

Also, I'm not holding my breath because this happened a few days ago, after a full takedown and reinstallation of the chips I fired the machine up to no fan noise, nominal temperature readings and accurate system specs. Then I shut it down, and restarted - and the fans went full-tilt again. I'm worried it will happen again and so I'm reluctant to shut down the computer.

Lastly, a friend told me to run XBench to test the system specs and they seem a little low... Overall score of 77.11, CPU Test at 225.01, Memory Test at 308.85, Quartz Graphics Test at 173.11, OpenGL Graphics Test at 210.35, UI Test at 17.35, Disk Test at 514.09 (full results in plaintext attached). Is it just my graphics card and UI that are killing my benchmark or are those abnormally low numbers for the CPUs as well? I did run some Ableton sessions and was surprised to hear some popping and clicking even at higher buffer sizes.

EDIT: ****, I knew it was too good to be true. Put the computer to sleep, woke it again and the fans started going crazy, temp/fanspeed all out of whack in SMCFanControl. :( What the hell is going on?
 

Attachments

  • TARStest1-Oct 22nd 2016.txt
    1.4 KB · Views: 203
Last edited:

h9826790

macrumors P6
Apr 3, 2014
16,656
8,587
Hong Kong
I'm not holding my breath but things seem to be working again. I say that because basically all I've done is take the CPU tray out and give it a hard look (and maybe some duster) as I've done every day this week. But when I reset the SMC and PRAM and fired it up, no fan noise. SMCFanControl shows 47C and 1113RPM on both BOOST fans - does that sound a little high?

Also, I'm not holding my breath because this happened a few days ago, after a full takedown and reinstallation of the chips I fired the machine up to no fan noise, nominal temperature readings and accurate system specs. Then I shut it down, and restarted - and the fans went full-tilt again. I'm worried it will happen again and so I'm reluctant to shut down the computer.

Lastly, a friend told me to run XBench to test the system specs and they seem a little low... Overall score of 77.11, CPU Test at 225.01, Memory Test at 308.85, Quartz Graphics Test at 173.11, OpenGL Graphics Test at 210.35, UI Test at 17.35, Disk Test at 514.09 (full results in plaintext attached). Is it just my graphics card and UI that are killing my benchmark or are those abnormally low numbers for the CPUs as well? I did run some Ableton sessions and was surprised to hear some popping and clicking even at higher buffer sizes.

EDIT: ****, I knew it was too good to be true. Put the computer to sleep, woke it again and the fans started going crazy, temp/fanspeed all out of whack in SMCFanControl. :( What the hell is going on?

XBench is very inaccurate now, especially the GPU test. The number is quite meaningless.
 

carpsafari

Suspended
Sep 13, 2015
277
57
the Netherlands
My ¢2...you either bent some pins, or there is a problem with Processor board.

Don't you know somebody with a 2010 MP, you could change the Processor board to see if the other one is working?
 

ActionableMango

macrumors G3
Sep 21, 2010
9,613
6,909
I doubt the fan/sensor connector failed in both heat sinks simultaneously without any visible damage. They are fairly robust connectors and I feel any damage to them would probably be visible.

Have you tried disconnecting/reseating everything? (Cables, memory, GPU) Remove all non-essential hardware like USB3 cards, etc.

Are there any diagnostic LEDs lit on the boards when the problem is occurring? If so, troubleshoot them with the service manual.

If none of that works, I'm starting to lean toward non-visible board damage like perhaps a static discharge to a critical component while handling the board/CPUs without being grounded.
 
  • Like
Reactions: MacProRestoreGuy

Mister Speaker

macrumors member
Original poster
Oct 14, 2016
32
0
Is there a utility available that will show me the individual temperatures of each chip? SMCFanControl just shows me one temperature reading.
 

Mister Speaker

macrumors member
Original poster
Oct 14, 2016
32
0
Is there ANY way this fan speed issue could be related to peripherals, or improper grounding? The past two days I have turned the machine on to no fan noise and nominal readings in SMCFanControl. I put the computer to sleep, and when I wake it, the fans immediately kick up. The only thing I can think of that I changed, was turning on my audio interfaces and synths.

Additionally, is there any chance that the SMC firmware could be at fault, and if so would a factory reset fix this?

EDIT: I downloaded iStatMenus and it's telling me the same as SMCFanControl. CPU temps nominal, but there's one thing - both BOOSTA and BOOSTB fans seem to sit right at 1113RPM, is this a little higher than expected? I can't hear them. Also there's a temperature difference of about 12C at 'Relative to ProcHot', is this something to worry about?
 
Last edited:

h9826790

macrumors P6
Apr 3, 2014
16,656
8,587
Hong Kong
Is there ANY way this fan speed issue could be related to peripherals, or improper grounding? The past two days I have turned the machine on to no fan noise and nominal readings in SMCFanControl. I put the computer to sleep, and when I wake it, the fans immediately kick up. The only thing I can think of that I changed, was turning on my audio interfaces and synths.

Additionally, is there any chance that the SMC firmware could be at fault, and if so would a factory reset fix this?

EDIT: I downloaded iStatMenus and it's telling me the same as SMCFanControl. CPU temps nominal, but there's one thing - both BOOSTA and BOOSTB fans seem to sit right at 1113RPM, is this a little higher than expected? I can't hear them. Also there's a temperature difference of about 12C at 'Relative to ProcHot', is this something to worry about?

It's very easy to test if that's peripherals related. You can simply remove them, then you will know the answer.

Buggy SMC, possible, but very low, never heard a single case due to faulty SMC firmware, almost always able to fix by SMC reset.

The iStat reading is OK, nothing to worry about.
 

Mister Speaker

macrumors member
Original poster
Oct 14, 2016
32
0
Last post before I'm out of ideas: Is it possible that the heatsinks are on too tight? I was under the impression that this particular generation of 5,1 CPU tray and heatsinks had spring-tensioned screws which don't allow for overtightening - I've had many conversations with friends in the know and they assured me that no spacers, washers, or precision torquing of screws is necessary with this model.

BUT, after getting off the phone with a reputable local retailer/service center, I'm not so sure. The technician on the phone basically said that even with the spring-tensioned heatsink screws, I'd have to get the torque setting just right... This doesn't seem to make sense to me; if I had crushed the CPUs wouldn't I know immediately upon starting up the machine? His solution was basically just to "play around with different torque settings" to see if it has an effect. So I'm going to try loosening the heatsink screws slightly. Does this make any sense at all?

I'm officially on my last-ditch efforts with this machine before I pull the hard drives and use the case for target practice.
 

Philocetes

macrumors regular
Sep 23, 2016
106
36
If this is a 5,1 hardware wise, the cpus will be held down by the socket on the motherboard with its little arm that swings down to tighten the cpu the proper amount. I don't see how over-tightening the heatsink screws will have impact on that, as it would be designed for the heatsink to be fully screwed down.

If its a 4,1, all the washers or delidding issues come into play.

see this starting at 1:20. This is not a mac, but the cpu socket is the same type as on my mac 5,1 computers. If it doesn't look like that, then perhaps you can post a photo of what it does look like.

 

ActionableMango

macrumors G3
Sep 21, 2010
9,613
6,909
Last post before I'm out of ideas: Is it possible that the heatsinks are on too tight? I was under the impression that this particular generation of 5,1 CPU tray and heatsinks had spring-tensioned screws which don't allow for overtightening - I've had many conversations with friends in the know and they assured me that no spacers, washers, or precision torquing of screws is necessary with this model.

I've done the CPU upgrade twice in a 2010 5,1 and in both cases I simply turned the 3mm bolts until they stopped. There was a very obvious stop for me, so there was no question of whether to tighten more or not.

To try to put into words how the Mac Pro bolts felt:
Easy, easy, easy, easy, easy, STOP.​

Vs typical bolts:
Easy, easy, medium, medium, medium, tight, tight, tighter... this is probably good enough.​

So no, I don't think precision torquing is necessary on the 5,1. However, if you are already considering the computer a near total loss, I suppose it is worth a try to back out the bolts very slightly and evenly.
 

Philocetes

macrumors regular
Sep 23, 2016
106
36
When I upgraded my 4,1 with non-delidded and washers, I had the high fan which I cured by tightening down the screws a bit more--my interpretation was that the cpu wasn't seated properly. Perhaps over-tightening can cause the problem--I have no idea.

It would not be a total loss even if the tray is busted, as one can get other ones on ebay (agressive asking prices though). Maybe one of the cpu is not working--its a lot of work, but perhaps the old ones can be put back in to ensure the tray itself has not been broken.
 

Mister Speaker

macrumors member
Original poster
Oct 14, 2016
32
0
Maybe one of the cpu is not working--its a lot of work, but perhaps the old ones can be put back in to ensure the tray itself has not been broken.

I've tried putting the old chips back in, no effect. And I've asked this before but if one of the chips wasn't working it simply wouldn't show up in my system profile, right? I'm having no issues with performance - both chips show up fine and the machine is markedly faster. It just can't read thermal information from the chips so the SMC kicks the fans up and I can't figure out why.

If this is a 5,1 hardware wise, the cpus will be held down by the socket on the motherboard with its little arm that swings down to tighten the cpu the proper amount. I don't see how over-tightening the heatsink screws will have impact on that, as it would be designed for the heatsink to be fully screwed down.

I've done the CPU upgrade twice in a 2010 5,1 and in both cases I simply turned the 3mm bolts until they stopped. There was a very obvious stop for me, so there was no question of whether to tighten more or not.

Right, this is the same generation of 5,1. It's got the clamp system that Philocetes mentioned, and the spring-tensioned screws feel exactly as you describe. I figured the heatsink tension wasn't an issue (given what I said above about switching out the old chips to no effect). I don't know what to do anymore. I'm still leaning towards it being an issue with the sensors but that's an expensive die to roll as well.
 

Philocetes

macrumors regular
Sep 23, 2016
106
36
Now that you talk about heat sensing, I do recall that there is kind of a heat sink foam tape or strip and I recall someone saying it was for thermal sensing. It bridged the gap between the heat sink and a bunch of capacitors on the mobo. For the 4,1, using washers, for instance, I had to buy a thicker pad.

Maybe that heat conducting foam strip isn't seating properly. Beyond that I don't have any other ideas--maybe there are some messages in the mac logs...
 

MacProRestoreGuy

macrumors newbie
Mar 2, 2020
20
0
Same issue here. Red light coming on the CPU tray. SMC and NVRAM reset do nothing. I will try reapplying the thermal paste, I guess.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.