7451 to 7455 iMac G4 CPU swap attempt (WIP)

[alectrona6400]

I have a 2002 iMac G4 15" (P80) that I pulled off one of my shelves to mess around with because I don't really use it a whole lot. I got curious since I had a couple 7455A CPUs on hand I bought off of eBay about a couple years ago and after years of practicing soldering and hot air work I thought I'd try swapping the CPU out. These 7455 CPUs already came with solder balls (which don't work so well, more on that later) and as such I thought it would be straightforward. It wasn't. Well, it would be... but it turns out those solder balls that come with every PowerPC processor (0.89mm) never really work half the time, and I'm not the only one struggling with this... even @dosdude1 has been struggling to get this to work and at first I thought it was my hot air station. While the CPU can make contact, it won't make full contact. This was evident as when I lightly pressed down one of the corners of the chip I could see some of the IPA I was using to clean up the board get pushed out of the area between those solder balls and pads. Seems as though reballing the chip is the way to go, and luckily I have some equipment for that work on the way! This is something I've always wanted to do for several years.

As for what I was doing, I was curious to see how well those would work, and confident that I can get it working again with the right equipment when that comes. I was running the original 7451 CPU at 900MHz and thought I could do better because I had a chip with the solder balls already present.... again that didn't work out. I was planning to run a 7455A at 1.2GHz, and even then I had some more work to do on the board. It seems like a few capacitors actually started leaking (and no, before you ask, it's not dried flux) so I went ahead and replaced those with fresh SMD caps from an assorted kit I bought off Amazon. They actually work great in other devices I've used them with! I referred to this blog post (in spanish, so you'll need to translate it) to see how exactly I could configure the computer. I have vcore set to 1.6V by removing R262 and replacing R258 with a 10kOhm resistor. https://rosysumenteinquieta.blogspot.com/2023/12/imac-g4-800-cpu-swap.html

If anyone has any other advice for me that would be great. I'm confident in my skills that I can get this working. I also have another BGA project in mind which involves swapping the CPU on an MMC-2 Intel Pentium III card to bring it up to 800MHz.

 

[herd]

Thanks for sharing your experience. The OEM chips typically used a 0.89mm Diameter 90% Pb and 10% Sn (by Weight) “High Melt” solder ball held between CBGA substrate and PCB with regular eutectic (63Sn/37Pb) solder. The idea was to have a ball that would not collapse when attached. The resulting column of metal would allow some flex to accommodate the difference in thermal expansion between the ceramic and PCB. This was later addressed by the HCTE substrate, and these chips could be soldered directly to the PCB with relatively thin fills.

So basically, to attach a 90/10 ball to a copper pad, you need to add some solder. Solder paste is the typical solution, but other methods are possible for hand assembly. It is not really feasible to directly attach the “High Melt” solder ball to the pad because extreme temperatures would be required and if the ball collapses then there would likely be other problems too.

Or you can remove the OEM balls and use smaller/shorter 63/37 balls. The long term thermal fatigue performance will be reduced somewhat, but BGA technology has proven to be more reliable than skeptics feared at its debut.

 

[alectrona6400]

I see... I'll probably try that. I did tin all the pads with fresh solder and clean it up with wick like dosdude1 does in his videos, but perhaps that isn't enough for these.

 

[dosdude1]

Yep, as mentioned, those weird solder balls that come from the factory on these PPC CPUs cannot simply be soldered to a bare pad, solder must be added to the pads before you're actually able to solder one of these CPUs with those solder balls. I've found that simply tinning the pads on the board before soldering works well enough. Only issue there is getting the chip aligned properly is a bit more difficult (as you have to "balance" it on top of the existing solder you applied to the pads), though with such large balls even that is not too difficult. I've never had an issue with them since I started doing it that way

 

[alectrona6400]

That did the trick! Unfortunately the system froze right as Geekbench was finishing up and now it starts without a chime and it won't finish loading. Might be a RAM issue or something else but I'm curious as to how far I should push a 1GHz 7455A at 1.6V. 1.3 was stable until that happened, as was 1.2... but the cooling system on this computer isn't great when it comes to actually working.

 

[Nixontheknight]

That did the trick! Unfortunately the system froze right as Geekbench was finishing up and now it starts without a chime and it won't finish loading. Might be a RAM issue or something else but I'm curious as to how far I should push a 1GHz 7455A at 1.6V. 1.3 was stable until that happened, as was 1.2... but the cooling system on this computer isn't great when it comes to actually working.

I would try swapping RAM or testing the mac with one stick of RAM at a time until you find the bad stick

 

[alectrona6400]

Already tried that, did not fix anything.

 

[Nixontheknight]

I also wonder if there's a way to test cache

 

[alectrona6400]

So uh, I did get my reball jig in but it's simply not fit for actualy using it... so I bought direct heat stencils which will hopefully make this easier. How big of solder balls can you use on the PowerPC chips? My best guess is 0.76mm but I'm not entirely sure.

 

[LightBulbFun]

That did the trick! Unfortunately the system froze right as Geekbench was finishing up and now it starts without a chime and it won't finish loading. Might be a RAM issue or something else but I'm curious as to how far I should push a 1GHz 7455A at 1.6V. 1.3 was stable until that happened, as was 1.2... but the cooling system on this computer isn't great when it comes to actually working.

interesting that yours also freezes! I also did this swap a few years back

uncooking the sunflower seeds, first generation iMac G4 repair/CPU replacement

some of the long timers here may remember my post here https://forums.macrumors.com/threads/show-us-your-chips.2127981/?post=26273568#post-26273568 about an iMac G4 I had with a very dead CPU, now while this would normally be the end of the line for such a machine or its logic board, I did...

forums.macrumors.com

and mine also freezes pretty soon after booting, but I always chalked that up to the fact I never did bother to adjust the vCore voltage, but its interesting to hear yours is freezing despite you having adjusted the vCore voltage...

I trust it that you did apply fresh thermal paste to the contact pad between the CPU head pipe and the heatsink dome itself?

 

[alectrona6400]

yes. of course i put fresh thermal paste.
i wonder if this is firmware related?

 

[alectrona6400]

I've actually taken on another project. I got some direct heat BGA stencils, and a 7447B chip from a dead iBook G4. The target machine is a 2003 eMac, and I had installed the giga designs firmware patch on it. I had swapped the 7445 for the 7447B that I reballed with 0.76mm 63/37 solder balls, and set the vcore to 1.325V which is the same as the later USB2 eMacs. Unfortunately the system does not start at all... and I may know why.. I forgot to sand down the balls on the bottom to where it would be level with the board. :/

 

[indibil]

Thanks for sharing your experience. The OEM chips typically used a 0.89mm Diameter 90% Pb and 10% Sn (by Weight) “High Melt” solder ball held between CBGA substrate and PCB with regular eutectic (63Sn/37Pb) solder. The idea was to have a ball that would not collapse when attached. The resulting column of metal would allow some flex to accommodate the difference in thermal expansion between the ceramic and PCB. This was later addressed by the HCTE substrate, and these chips could be soldered directly to the PCB with relatively thin fills.

So basically, to attach a 90/10 ball to a copper pad, you need to add some solder. Solder paste is the typical solution, but other methods are possible for hand assembly. It is not really feasible to directly attach the “High Melt” solder ball to the pad because extreme temperatures would be required and if the ball collapses then there would likely be other problems too.

Or you can remove the OEM balls and use smaller/shorter 63/37 balls. The long term thermal fatigue performance will be reduced somewhat, but BGA technology has proven to be more reliable than skeptics feared at its debut.

Your comment puts me at ease. I started using 0.76 balls with 66/37 and currently I feel more comfortable working with 0.65, and so far it works well for me. I do have to say that the CPU stays closer to the PCB and the balls tend to "collapse" a little, they stay flatter.

Could this be a problem? Or maybe I apply too much heat?

 

[indibil]

instead of 66/37 I meant 63% tin - 37% lead (63/37)