MP 1,1-5,1 Likely consequences of broken PRAM battery holder

[JedNZ]

I worked on a cMP 4,1 dual CPU yesterday. POSTs and power light stays on but nothing displays on screen.

Upon opening it up I notice the PRAM battery holder has been previously damaged, and a paperclip and tape have been used to fix it. Despite the fix, the battery is not held securely against the backplane and the whole setup is loose.


what is the likely consequence of a PRAM battery that doesn’t make constant contact? That is, what damage might it do if it makes then breaks contact, then makes contact again etc while running?

While poking around I found one of the Northbridge rivets was missing. Yikes! I cleaned it all up (thermal paste was like hard clay!), applied new thermal paste, and did the same with both CPUs. The CPUs have not been delidded and there are no spaces or washes to compensate for the height of the IHS still being on the CPU. Whoever upgraded the original CPUs to the two X5680’s was a butcher.

Can’t believe the machine even worked originally with all these issues.

I got it to POST and get almost to the login screen with just CPU A installed. KP indicated ram issues, and the power light was flashing on and off which also suggests that. Removed CPU tray, tightened heatsink screws, removed all but one ram DIMM and got to login screen but KP again. Couldn’t get it to POST after that. Installed CPU B just to see if that might change anything. No POST. Then discovered the boot SSD (some junk brand - manufactured in 2013) had severe issues - would mount on my own cMP but wouldn’t display its content. Wouldn’t show in Disk Utility or DriveDX even though the icon showed on the desktop. Beach balled Finder.

So a lot of problems. I’m thinking replace the backplane as the next step. But need to know the effects of a broken/botched PRAM battery holder and how this might have played a part in the whole issues this sad cMP is experiencing.

 

[Snow Tiger]

I worked on a cMP 4,1 dual CPU yesterday. POSTs and power light stays on but nothing displays on screen.

Upon opening it up I notice the PRAM battery holder has been previously damaged, and a paperclip and tape have been used to fix it. Despite the fix, the battery is not held securely against the backplane and the whole setup is loose.


what is the likely consequence of a PRAM battery that doesn’t make constant contact? That is, what damage might it do if it makes then breaks contact, then makes contact again etc while running?

While poking around I found one of the Northbridge rivets was missing. Yikes! I cleaned it all up (thermal paste was like hard clay!), applied new thermal paste, and did the same with both CPUs. The CPUs have not been delidded and there are no spaces or washes to compensate for the height of the IHS still being on the CPU. Whoever upgraded the original CPUs to the two X5680’s was a butcher.

Can’t believe the machine even worked originally with all these issues.

I got it to POST and get almost to the login screen with just CPU A installed. KP indicated ram issues, and the power light was flashing on and off which also suggests that. Removed CPU tray, tightened heatsink screws, removed all but one ram DIMM and got to login screen but KP again. Couldn’t get it to POST after that. Installed CPU B just to see if that might change anything. No POST. Then discovered the boot SSD (some junk brand - manufactured in 2013) had severe issues - would mount on my own cMP but wouldn’t display its content. Wouldn’t show in Disk Utility or DriveDX even though the icon showed on the desktop. Beach balled Finder.

So a lot of problems. I’m thinking replace the backplane as the next step. But need to know the effects of a broken/botched PRAM battery holder and how this might have played a part in the whole issues this sad cMP is experiencing.

Gosh , if your Mac was a horse it would've gotten a mercy shot by now ...

The thing I worry about your Mac the most is you tightening those heatsinks with the upgraded retail silicon without any washers on the standoffs . You're playing with fire and can easily destroy the socket pin array if those sinks lower down a little too much .

Find some 2 mm ( actually 2.3 mm ) nylon washers now and place one on each of the eight CPU standoffs .

The thing I would do now is verify you didn't already bend some socket pins .

You also need a properly functioning NB controller heatsink or your Mac will simply not boot or will shut down under load if you manage to get her into an OS session .

I think the battery only affects the system clock and the time since boot report . These are Intel Macs . I have run cMPs without batteries , you just have to manually set the time each session . But apps need to know the actual time .

But I wouldn't proceed further with this project until both CPU sockets have been checked for bent pins .

 

[JedNZ]

LOL Luckily it’s not mine. I sold them one of my single CPU 5,1 (original) with upgraded GPU, NVMe PCIe boot and upgraded to 6-core. I was just doing some troubleshooting for them because they use the cMP in their sound studio and have a big job on this week.

NB clean + rivet has been sorted, so that’s one thing taken out of the equation. Thanks for the info about the PRAM battery. I’ll recommend the nylon washers. We did find one CPU pin bent upwards which no doubt was a result of not having the nylon washer.

 

[Snow Tiger]

LOL Luckily it’s not mine. I sold them one of my single CPU 5,1 (original) with upgraded GPU, NVMe PCIe boot and upgraded to 6-core. I was just doing some troubleshooting for them because they use the cMP in their sound studio and have a big job on this week.

NB clean + rivet has been sorted, so that’s one thing taken out of the equation. Thanks for the info about the PRAM battery. I’ll recommend the nylon washers. We did find one CPU pin bent upwards which no doubt was a result of not having the nylon washer.

here are the washers I use . They are made of Nylon 66 , which is a pretty safe plastic around electronics .

OD = 7 mm
ID = 5.2 mm
Length = 2mm

The seller no longer stocks them .