Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
MP 1,1-5,1 Mac Pro A1289 Broken Chip Advice
- Thread starter mgaudio
- Start date
- Sort by reaction score
I'm officially inspired by everyone's gung-ho attitude in helping me. I've disassembled the heatsinks again, and I'm going to detach the circuitboard from the metal tray. I'm gonna document every square inch of this thing with a newly-discovered macro-mode on the camera in my new phone. Lay back until tomorrow guys . . . I need a glass of wine for now ! Have a great evening, and again many thanks !!
So I've scanned my CPU board at high resolution to see if it helps spotting the location of the broken-off chip.
If I'm gonna consult a repair tech, I can then at least hopefully give him the chip and the location.
I've made a shared G-Drive folder, and have uploaded only the underside so far at 600 dpi.
Higher resolution 1200 dpi to come soon at the same shared folder link, and also the front side scans.
(there's a slight mis-alignment of the joined scanned images, but it's tiny - and nowhere near any chips)
It might be better to download and view locally, as it seems slow to load such a large image within a browser.
Maybe this will also be useful to macrumors techheads for further reference too, so please feel free to download and share between fellow macrumors users.
Let's all play "Spot The Missing Chip !" . . .
MacPro_2009_CPU_Tray_Board - G-Drive Shared Link:
drive.google.com
If I'm gonna consult a repair tech, I can then at least hopefully give him the chip and the location.
I've made a shared G-Drive folder, and have uploaded only the underside so far at 600 dpi.
Higher resolution 1200 dpi to come soon at the same shared folder link, and also the front side scans.
(there's a slight mis-alignment of the joined scanned images, but it's tiny - and nowhere near any chips)
It might be better to download and view locally, as it seems slow to load such a large image within a browser.
Maybe this will also be useful to macrumors techheads for further reference too, so please feel free to download and share between fellow macrumors users.
Let's all play "Spot The Missing Chip !" . . .
MacPro_2009_CPU_Tray_Board - G-Drive Shared Link:
MacPro_2009_CPU_Tray_Board - Google Drive
So I've scanned my CPU board at high resolution to see if it helps spotting the location of the broken-off chip.
If I'm gonna consult a repair tech, I can then at least hopefully give him the chip and the location.
I've made a shared G-Drive folder, and have uploaded only the underside so far at 600 dpi.
Higher resolution 1200 dpi to come soon at the same shared folder link, and also the front side scans.
(there's a slight mis-alignment of the joined scanned images, but it's tiny - and nowhere near any chips)
It might be better to download and view locally, as it seems slow to load such a large image within a browser.
Maybe this will also be useful to macrumors techheads for further reference too, so please feel free to download and share between fellow macrumors users.
Let's all play "Spot The Missing Chip !" . . .
MacPro_2009_CPU_Tray_Board - G-Drive Shared Link:
MacPro_2009_CPU_Tray_Board - Google Drive
Seems to be Q6890:
OMG ! Already !! You've seriously got eagle-eyes tsialex.
I'd been staring at this for frikkin' hours and hours and could not spot "where something was missing".
You're absolutely correct. Seriously man, many MANY thanks.
I must have knocked it off when I cleaned/reseated the Northbridge heatsink, and I thought I was ultra-careful.
Obviously not.
OK, well now I need to find a guy to give me a quote to see if it's worth fixing.
Is it appropriate to post any potential names of a tech I'm considering contacting ? Or is it better via PM ?
p.s. I just posted the topside scan too, but it's irrelevant now that you've found it. I will leave them in this shared folder for any future members. I've also uploaded the 1200dpi versions now too.
It was harder to scan the topside as I could not get it flat on the scanner bed due to the RAM locks.
I might try photoshopping together several of my phone-camera photos too - mainly as a thank you to all who helped me track this down.
I'd been staring at this for frikkin' hours and hours and could not spot "where something was missing".
You're absolutely correct. Seriously man, many MANY thanks.
I must have knocked it off when I cleaned/reseated the Northbridge heatsink, and I thought I was ultra-careful.
Obviously not.
OK, well now I need to find a guy to give me a quote to see if it's worth fixing.
Is it appropriate to post any potential names of a tech I'm considering contacting ? Or is it better via PM ?
p.s. I just posted the topside scan too, but it's irrelevant now that you've found it. I will leave them in this shared folder for any future members. I've also uploaded the 1200dpi versions now too.
It was harder to scan the topside as I could not get it flat on the scanner bed due to the RAM locks.
I might try photoshopping together several of my phone-camera photos too - mainly as a thank you to all who helped me track this down.
Last edited:
I just added some ultra-high resolution scans to the folder (4800dpi) of the chip, both top and bottom - and also a 4800dpi scan of that precise section of the processor board.
Does it look like I ripped anything beyond repair ?
Does it look like I ripped anything beyond repair ?
The board is too dirty and the contrast is not good enough to tell for sure, black solder mask PCBs are the worst for this type of investigation, but seems that the pads are still there. I have a 45x trinocular microscope, but I don't have the camera yet, I'll check my early-2009 dual CPU tray next week to see if the traces are exactly the same.
My concern is this trace:
Maybe you got lucky, but the PCB needs to be cleaned and inspected with a microscope to be 100% sure.
Last edited:
Thanks Borowski. I'm in LA, so there's gotta be a good computer tech here somewhere that can handle this task.
There's a nearby guy I think does mac tech repairs that I'm going to write to today for a quote, with some of the notes and pictures that tsialex has kindly provided to me.
Just giving some props to Macschrauber, as his initial analysis/advice of checking the backside of the board was in fact the correct one. Thanks man !
UPDATE: So it's been a week, and I've since made enquiries into getting it fixed with several mac techs, but it seems to be a tough one. So I've therefore decided that I've got nothing to lose, and I'll attempt to fix it myself. I'm gonna document the process in photos and video which I'll continue to post here for anyone interested.
I've got a soldering kit, and a digital USB microscope, and after watching some detailed chip-soldering tip/tutorials I've got some solderwick and a fluxpen on its way too. I will practice on something I can dispose of.
I'm now in the process of trying to identify the correct orientation of the chip on the board, but it seems that this chip has no dot-indentifier for Pin-1 on it. I've photoshopped an overlay of my broken-off chip onto a new closeup of the board location identified by tsialex - both in the orientation that I suspect is correct, and also rotated 180 degrees.
I didn't try to straighten any bent pin-legs yet either - as I think it helps to identify the direction in which it was torn off.
You can clearly see now the path of where I must have slipped with my tool when removing the Northbridge heatsink pins - and it looks like I hit that top corner of the chip and knocked it away from the pinhole. So anyone attempting this Northbridge cleanup process - learn from my mistake and be ultra-careful when removing those pins !
Here's a new shared folder with my microscope views of:
1. the board location, plus photoshop overlays (traces seem ok?)
2. the dislocated chip topside/underside on a 5mm grid (visible on the right)
3. some other locations on the CPU tray where this same K72WO chip is still intact.
If anyone has some advice with regard to identifying the correct orientation it would certainly be appreciated !
I thought that there was a slightly greater distance between the K72WO text and the edge of the chip on the position closer to the "V" marking on the CPU board, but I might be imagining this.
Microscope Photos - Pre-Soldering 09-05-2021 :
drive.google.com
P.S. additional question - do you guys recommend replacing the Northbridge heatsink retainer-pins anyway (even if the existing pins appear to be ok). If so, should I get metal pins with plastic protection/insulation washers, or use plastic ones for a little more "give" during heat expansion/contraction ? There seem to be conflicting opinions on this.
I've got a soldering kit, and a digital USB microscope, and after watching some detailed chip-soldering tip/tutorials I've got some solderwick and a fluxpen on its way too. I will practice on something I can dispose of.
I'm now in the process of trying to identify the correct orientation of the chip on the board, but it seems that this chip has no dot-indentifier for Pin-1 on it. I've photoshopped an overlay of my broken-off chip onto a new closeup of the board location identified by tsialex - both in the orientation that I suspect is correct, and also rotated 180 degrees.
I didn't try to straighten any bent pin-legs yet either - as I think it helps to identify the direction in which it was torn off.
You can clearly see now the path of where I must have slipped with my tool when removing the Northbridge heatsink pins - and it looks like I hit that top corner of the chip and knocked it away from the pinhole. So anyone attempting this Northbridge cleanup process - learn from my mistake and be ultra-careful when removing those pins !
Here's a new shared folder with my microscope views of:
1. the board location, plus photoshop overlays (traces seem ok?)
2. the dislocated chip topside/underside on a 5mm grid (visible on the right)
3. some other locations on the CPU tray where this same K72WO chip is still intact.
If anyone has some advice with regard to identifying the correct orientation it would certainly be appreciated !
I thought that there was a slightly greater distance between the K72WO text and the edge of the chip on the position closer to the "V" marking on the CPU board, but I might be imagining this.
Microscope Photos - Pre-Soldering 09-05-2021 :
Microscope Photos - Pre-Soldering 09-05-2021 - Google Drive
P.S. additional question - do you guys recommend replacing the Northbridge heatsink retainer-pins anyway (even if the existing pins appear to be ok). If so, should I get metal pins with plastic protection/insulation washers, or use plastic ones for a little more "give" during heat expansion/contraction ? There seem to be conflicting opinions on this.
Last edited:
The "V" marking is guiding the direction of the transistor location.
Take a look at the same component nearby (Q6380) and arrange the transistor similarly.
I guess it doesn't matter, as this is a dual-channel transistor.
If the K72 transistor failed, you can buy new chips to try. They are not expensive, anyway. The name is below. (10$ will gets you 100pcs)
2N7002DW-7-F 2N7002DW SOT363 K72 DUAL N-CHANNEL ENHANCEMENT MODE FIELD EFFECT TRANSISTOR
Datasheet in the below link
2N7002DW PDF
Part #: 2N7002DW. Description: DUAL N-CHANNEL ENHANCEMENT MODE FIELD EFFECT TRANSISTOR. File Size: 66.24 Kbytes. Manufacturer: Diodes Incorporated.
pdf1.alldatasheet.com
Yes, I'd definitely replace the pins of the NB, just to be sure the heatsink construction will be working reliably for years to come.
Opinions differ as you mentioned, IMHO it doesnt matter so much which way you choose , as long as the heatsink is mounted with proper contact to the die.
Personally I never searched for plastic push-pins but use (nylon) M3x 16mm screws, nuts and washers.
Metal screws and plastic pins work fine too.
Note that with screws I recommend to measure the distance from the head of the pin to the heatsink before replacing the pushpins to maintain the same amount of pressure on the heatsink.
Delta NB diode - NB heatsink should be 12 Celsius max for a good NB repaste with the MP running at idle.
Thanks guys . . . very helpful.
as you can tell, this is all new to me - being no electronics expert.
After reading your infotips I found this page showing a small internal schematic, which visually explains what's going on inside that chip - and why the orientation doesn't matter:
original pdf document:
screenshot:
_______________________
and also, the coding after the "K72" is all about the year & month of manufacture apparently.
mine (K72WO) deciphers to 2009 / October:
_______________________
I'm guessing it might be easier to solder in a brand new one, rather than straightening pinlegs.
After looking at Nguyen's links - I found this:
www.aliexpress.com
do you think this one would do the job ?
and yet another question . . .
if I use MX4 instead of AS5 for the Northbridge, can/should I also use it for my 3.46Ghz X5690 Processors ?
as you can tell, this is all new to me - being no electronics expert.
After reading your infotips I found this page showing a small internal schematic, which visually explains what's going on inside that chip - and why the orientation doesn't matter:
original pdf document:
screenshot:
_______________________
and also, the coding after the "K72" is all about the year & month of manufacture apparently.
mine (K72WO) deciphers to 2009 / October:
_______________________
I'm guessing it might be easier to solder in a brand new one, rather than straightening pinlegs.
After looking at Nguyen's links - I found this:
100PCS 2N7002DW-7-F 2N7002 SOT-363 DUAL N TRANSISTOR 2N7002DW MAKRING K72 702 2.2 bubao - AliExpress 502
Smarter Shopping, Better Living! Aliexpress.com
and yet another question . . .
if I use MX4 instead of AS5 for the Northbridge, can/should I also use it for my 3.46Ghz X5690 Processors ?
Last edited:
Buying a new 2N7002DW from AliExpress will take at least 3 weeks to arrive.
If this happened to me, I'd remove the solder from the pads with the solder wick, straighten the chip/remove all the old solder from the chip terminals and then solder it back. I know that it's not an easy job for a newbie, but everyone was a newbie at one point in time.
You have nothing to loose here, the board is pretty dead right now and could be dead even after a new MOSFET is back to it's place.
No need to use the same expensive thermal paste everywhere, even cheap white silicone thermal paste is enough for the northbridge, but you will need to remove and apply new thermal paste once you remove the Xeon heatsinks.
If this happened to me, I'd remove the solder from the pads with the solder wick, straighten the chip/remove all the old solder from the chip terminals and then solder it back. I know that it's not an easy job for a newbie, but everyone was a newbie at one point in time.
You have nothing to loose here, the board is pretty dead right now and could be dead even after a new MOSFET is back to it's place.
No need to use the same expensive thermal paste everywhere, even cheap white silicone thermal paste is enough for the northbridge, but you will need to remove and apply new thermal paste once you remove the Xeon heatsinks.
yeah, that makes total sense. thanks tsialex. I'll give it a go then.
from what you can tell in the new photos, do those traces seem ok in pic 1a. ?
drive.google.com
from what you can tell in the new photos, do those traces seem ok in pic 1a. ?
Microscope Photos - Pre-Soldering 09-05-2021 - Google Drive
yeah, that makes total sense. thanks tsialex. I'll give it a go then.
from what you can tell in the new photos, do those traces seem ok in pic 1a. ?
Microscope Photos - Pre-Soldering 09-05-2021 - Google Drive
The photos you took don't have contrast enough to see if the traces are not damaged, curse of the dark solder mask PCBs, you will probably need to check with a multimeter to be sure.