Mac Pro 4,1 to 5,1 RAM slot 1 not recognized

[Bob-K]

Just did a 4,1 to 5,1 upgrade on my 2009 Mac Pro running 10.11.6 El Capitan. All is well EXCEPT that RAM slot 1 isn't recognized.

Before: 2009 Mac Pro 4,1 2 x 2.26GHz quad-core

After: 2009 Mac Pro 5,1 2 x 3.33GHx (x5680) hex-core

Firmware update went fine. Both CPUs are recognized in System Report.

RAM slot 1 is not recognized, and red LED is lit on the logic board, between the heat sinks, next to CPU B.

Tried: resetting NVRAM, resetting SMC, re-seating RAM, swapping RAM with slot 2 (the RAM is fine).

Last week when I repaired my Northbridge heat sink, a diagnostic LED on the logic board lit up, and I got it to go away by tightening the heat sink bolts. I didn't know what the LED was indicating, but it was in the same general location between the heat sinks. (I think the CPU LEDs are on the sides.)

Tonight I tried tightening the heat sink bolts a little, but I'm hesitant to go further. Apple Service Guide says "finger tight plus 1/4 turn". I think I'm there, or even a little past.

Suggestions?

Thanks!

 

[flowrider]

When I updated my 5,1 with X5677s I had the same issue. I switched CPUs and the problem followed the CPU. I know switching would be a lot harder on a 4,1 than a 5,1 cMP. Anyway I exchanged the offending CPU, and the problem went away with the replacement.

Lou

 

[Bob-K]

When I updated my 5,1 with X5677s I had the same issue. I switched CPUs and the problem followed the CPU. I know switching would be a lot harder on a 4,1 than a 5,1 cMP. Anyway I exchanged the offending CPU, and the problem went away with the replacement.

Lou

Thanks, Lou!

After posting this, I found the other thread where you posted.

I swapped CPUs and the problem followed the CPU - instead of slot 1 being not recognized, it was now slot 4. I just received the CPUs today, so I contacted the seller.

In your case, was there any visible damage to the defective CPU? And were your processors lidded or de-lidded?

 

[flowrider]

^^^^None that I could see. In my case, when I switched out the CPUs, the offending slot remained the same. Since I have a 5,1 cMP, my CPUs were stock with the lids.

Lou

 

[DNComputers]

Thanks, Lou!

After posting this, I found the other thread where you posted.

I swapped CPUs and the problem followed the CPU - instead of slot 1 being not recognized, it was now slot 4. I just received the CPUs today, so I contacted the seller.

In your case, was there any visible damage to the defective CPU? And were your processors lidded or de-lidded?

If it's following the CPU, then I think it's safe to assume no pins are bent/damaged, and nothing is stuck to the pins, so it's most likely something on the CPU, a bad CPU, or you're just not tightening it enough. You could try cleaning the bottom of the CPU with all the pads with an alcohol wipe, get rid of any oil, or thermal paste that may have made its way there.

If this is delidded (I assume so since you're reading from the Apple guide), tightening it further may solve the issue, unless you have amazing strength, it is basically impossible to overtighten, from your description, I think you may not be tightening it enough. I find a lot of people get too scared and even with the delidded processors they undertighten. With delidded processors it will stop when it's all the way.

I'm sure it can happen, but in thousands of these processors, I've never seen one that didn't recognize a RAM slot, and not physically damaged (missing pad etc). At the same time, you don't want to keep removing and reinstalling the CPUs, each time you do that, you risk damaging your expensive CPU board, if you can wait, the safest thing might be to just swap the CPU out with the seller, and try a different one.

 

[Bob-K]

If it's following the CPU, then I think it's safe to assume no pins are bent/damaged, and nothing is stuck to the pins, so it's most likely something on the CPU, a bad CPU, or you're just not tightening it enough. You could try cleaning the bottom of the CPU with all the pads with an alcohol wipe, get rid of any oil, or thermal paste that may have made its way there.

I checked with a magnifying glass and didn't see any visible debris on the bottom of the CPU. Still, I cleaned it with alcohol anyway, while doing the swap.

unless you have amazing strength, it is basically impossible to overtighten...With delidded processors it will stop when it's all the way.

Good to know - thanks for this info!

I definitely went tighter than Apple's recommendation, and tighter than any other one I've done.

Last week I removed the [original] CPUs to replace the Northbridge heat sink rivets, one of which was missing. After replacing the CPUs, I did get a diagnostic LED on the logic board lighting up, and resolved it by tightening further. From that experience, I was aware that it's possible to under-tighten, so I was a bit less gentle this time around.

If I hadn't already mailed back the CPU in question, I would have tried tightening more based on your recommendation.

I'm sure it can happen, but in thousands of these processors, I've never seen one that didn't recognize a RAM slot, and not physically damaged (missing pad etc).

What's a pad?

At the same time, you don't want to keep removing and reinstalling the CPUs, each time you do that, you risk damaging your expensive CPU board, if you can wait, the safest thing might be to just swap the CPU out with the seller, and try a different one.

Yes, I figured the more times I install and remove CPUs, the greater chance of something going wrong. Then again, I'm getting more skilled at it each round!

I figured it was likely a CPU problem after seeing at least two reports here of the same issue. But if I understand you correctly, you're saying that it's likely they just weren't tightening the bolts enough?

 

[DNComputers]

I checked with a magnifying glass and didn't see any visible debris on the bottom of the CPU. Still, I cleaned it with alcohol anyway, while doing the swap.



Good to know - thanks for this info!

I definitely went tighter than Apple's recommendation, and tighter than any other one I've done.

Last week I removed the [original] CPUs to replace the Northbridge heat sink rivets, one of which was missing. After replacing the CPUs, I did get a diagnostic LED on the logic board lighting up, and resolved it by tightening further. From that experience, I was aware that it's possible to under-tighten, so I was a bit less gentle this time around.

If I hadn't already mailed back the CPU in question, I would have tried tightening more based on your recommendation.



What's a pad?



Yes, I figured the more times I install and remove CPUs, the greater chance of something going wrong. Then again, I'm getting more skilled at it each round!

I figured it was likely a CPU problem after seeing at least two reports here of the same issue. But if I understand you correctly, you're saying that it's likely they just weren't tightening the bolts enough?

The pads are the gold circles on the bottom of the processors.

Not tightening it all the way would be my guess based on my experience. After what you've tried, I think trying a different CPU is a good idea. If the replacement does the same thing, just try tightening it down a little tighter than you did, and I think you'll be fine.

 

[ActionableMango]

Still critical information missing. New procs are lidded or delidded? If lidded, did you use the washers?

 

[Bob-K]

De-lidded.

 

[zephonic]

Weird, when I upgraded my 4,1 quad (W3520) to a 5,1 hex (W3680) I lost memory slot 1 as well.

The consensus on this forum was that I inadvertently damaged something, and I moved a DIMM from slot 1 to slot 4 (I have 3x8GB), and carried on.

This has me thinking it may not be damage after all...then again, when I put the quad back in, the problem persisted...


I posted about it here: https://forums.macrumors.com/thread...-new-cpu-to-use.1122551/page-35#post-24329423

 

[DNComputers]

Weird, when I upgraded my 4,1 quad (W3520) to a 5,1 hex (W3680) I lost memory slot 1 as well.

The consensus on this forum was that I inadvertently damaged something, and I moved a DIMM from slot 1 to slot 4 (I have 3x8GB), and carried on.

This has me thinking it may not be damage after all...then again, when I put the quad back in, the problem persisted...


I posted about it here: https://forums.macrumors.com/thread...-new-cpu-to-use.1122551/page-35#post-24329423

Your problem is different, your problem wasn't solved by changing the CPU indicating a problem with the board. To me, that sounds like you may have damaged the pins slightly during the install. The socket in yours is different, it actually holds the cpu down when you push the lever down, this helps make sure it is down all the way, the 4,1 dual cpus model like the poster has does not have this, making undertightening a common issue with those. (I don't have as much experience with the lidded CPUs like you have in your single cpu model)

I don't see any obviously bent or missing pins in your picture, which is good, but if some were pressed down just a little bit, or facing a little bit in the wrong direction, I've seen these not make contact, you couldn't see that in the picture.

If you want to try, you could take it back apart, clean the CPU, make sure there is nothing on the pins, (don't touch them) and that none of them are bent, or broken off, they should all be pointing the same direction as the ones next to it, and appear to be the same height, then reapply paste, and make sure it's good and tight, see if that solves your issues. I think just to try to get slot one working, if you're okay running it the way it is, it may not be worth the additional risk, that you may damage it further trying to repair any damaged pins.

 

[martyr]

Holy flying efff, guys. My Slot 1 gave out after I re-applied thermal paste to the CPUs. I assumed it was bad memory so I tried all of the ones I had (6x16GB) and all caused the same white blinking power light, which indicates not enough memory to boot up.

I read this thread and decided to put a little bit more elbow grease into tightening up CPU A's heatsink, to the point where turning it would actually rotate the processor tray on my desk, and voila! Powered up just fine.

I had some issues with Slot 7 weeks ago that I ended up thinking the stick of RAM was hosed so I ordered another one from OWC but still encountered similar issues. Gonna do some testing with tightening and see if that was the actual culprit - might have an extra 16GB RAM stick on my hands!
[doublepost=1539220982][/doublepost]Okay, after putting the same 3 sticks I had from Slots 1, 2, and 3 into Slots 5, 6, and 7, I now get a read RAM error light on Slot 6, and macOS doesn't recognize anything in that slot.

I've tightened CPU B's heatsink to the same point as CPU A and tried switching the RAM sticks, but Slot 6 still shows this issue.

Anything else that you can think of that I can try?

Thanks!

 

[martyr]

As a quick follow-up to this, my issue with Slot 6 has not gone away, even with tightening the CPU B heatsink. However, I discovered that one of the gold platings on CPU B has come off and fallen into the pins of the CPU B base, where it has likely bent some pins. Take a look at the attached images...



I removed the piece that was stuck in the pins and tried to bend the pins back into place but the Slot 6 still had issues. My guess is that the pins that are wonky are responsible for interfacing with Slot 6. Just a guess.

For now, it was booting up fine and with 16GBs in Slots 1, 2, 3, and 5, it was humming at about 23000 in Geekbench 4. So I'm going to close it up for now, eventually buy a sewing needle and a good magnifying glass, and will try to bend the pins back later to see if I can get Slot 6 working again.

 

[lie2me]

Marty:
I just purchased a used Mac Pro 2010 and it has an issue with Slot 5. Memory DIMM in slot 5 is not recognized.
I swapped memory DIMMs in the hopes it was a DIMM issue but that did not work.
I will take another stab at it hoping it is a DIMM issue with a different DIMM memory before investigating the CPU.

The gold plating that came off your CPU is likely to be the pad for a data connection to the memory DIMM if the memory controller is in the CPU. Otherwise it is likely to be a connection to the memory controller on the CPU board that interfaces with the DIMMs. Without that pad on the CPU, you are likely never going to get that CPU to interface with your DIMM slot. You are likely going to need a replacement CPU, even if you can repair (straighten) the pins in the CPU socket on the motherboard with a needle. I don't know if you can resolver that pad to the CPU substrate, probably unlikely. Good luck.

 

[Brian M]

I'm having the same issue with dim slot 5! I'm guessing it is also with my CPU. I didn't see any damage to the connector on my cpu board. I placed the old CPUs back in and still got an error. Uggg maybe it is my connector on my board? When I remove memory from slot 5 I'm fine though, no red light.

 

[MacProRestoreGuy]

I'm having the same issue with dim slot 5! I'm guessing it is also with my CPU. I didn't see any damage to the connector on my cpu board. I placed the old CPUs back in and still got an error. Uggg maybe it is my connector on my board? When I remove memory from slot 5 I'm fine though, no red light.

Same issue, slot 1. I think I damaged the board by not grounding it during repair. On another dual CPU tray , BOOSTA fan is showing 0 RPM in Macs Fan Control. The seller I got the trays from said I can exchange them so not a big deal.

 

[Trungha]

The same problem as you! How to solve it! Thanks!

Same issue, slot 1. I think I damaged the board by not grounding it during repair. On another dual CPU tray , BOOSTA fan is showing 0 RPM in Macs Fan Control. The seller I got the trays from said I can exchange them so not a big deal.

The same problem as you! How to solve it! Thanks!