• Updated 2023-07-12: Hello, Guest! Welcome back, and be sure to check out this follow-up post about our outage a week or so ago.

The B&W G3 that won't die... or live.

Big Ben

Well-known member
Hi!

This is an other episode from "BigBen tries to make garbage working again".
Years ago, I bought a bunch of PowerMac spare parts. I was interrested by a PSU to replace a dead one in a PowerMac G4.

Within the spare parts was a B&W G3 motherboard. Guess what I tried to boot it up.
Starting by rewiring an ATX PSU to get it working... painfully.

Power on button doesn't want to work, the button itself is fine. But it does nothing, if the motherboard is powered up, it properly shuts down the computer.
Internal speaker won't work... but I did get one bip once. So it is actually working. But can't get any sound, no chime, no system beep from the OS.
Above points meant that I actually did manage to boot the thing. By shorting the ATX gren wire you can power on the PSU. It worked, sometimes.

Then I checked many things without success and I some point I manage to find on a shelf the B&W PSU, why the heck I haven't used it before you may ask? Because I didn't know it was a B&W PSU. (Didn't have the computer case, never had a B&W). But the wiring on the ATX connector just hint me when I was looking for another PSU for cross-testing.

I realized, I made a mistake. Power OK on ATX PSU is +5V. On the G3 it's only 3.3V. Maybe I fried something.
At some point I managed to boot the computer with the right PSU, still using the green wire trick.

Running MacTest Pro several times. Complete logic board tests are successfull.

Still the damn thing won't work properly. If I try to kickstart the PSU using the green wire, CPU led stay green and it's shuting down right after. Holding the CUDA button pressed make some how the computer beging to boot, diagnostic LED going in the right state like it's booting, but it doesn't. No surprise here. Cuda button is not meant to be pressed.

Most tests have done with and without a PRAM battery, a CR2032, but it may not be enough for this motherboard?

What the heck is going on with this Yosemite motherboard (rev 1 of course)?

Is there any hope? I wouldn't mind a stealth G3 in a brand new case as another gaming/bridge machine.
 

SiliconValleyPirate

Well-known member
The refusal to start sounds like a PMU issue. It’s difficult to say what the fault could be as the board’s been out of the chassis for some time, so it could be physically damaged, could have had ESD shock etc.

I do know that B&Ws are pretty sensitive to scrambled PRAM. I’ve had several over the tears and all of them have had instances where they’ve suffered from a low battery or something that’s scuppered the content in the PRAM and go off in a massive sulk, the solution was the usual pull the PRAM battery, pull the mains and leave them overnight (min 12h) and then tgey usually recover. One thing I’ve found that did cause that was constant stop-starts and incomplete startups.

However as the board came to you stripped that may not be the issue, unless it had a half-dead battery in it when you got it.
 

Big Ben

Well-known member
No battery.

But I left it unplugged for about a whole day.
I became suddenly easier to boot, still forcing it as the power button won’t do a thing, but it exactly booted every second try, I did few boots, never had an exception.

Super strange. ESD shock is totally a possibility.
What is strange is MacTest Pro not detecting anything and the computer seems stable when booted.

Maybe the cuda chip or the reset circuit is kinda broken.

I tried to dig out a schematic for more investigation but… didn’t found a thing.

Too bad
 

SiliconValleyPirate

Well-known member
Okay, wasn’t sure as you mentioned trying to jury rig it with a CR2032. Because of the lower voltage that might have given it a headache.

It does sound like either the reset, CUDA or power on circuit is faulty, or the PMU is decidedly unhappy.

Without delving into oscilloscopes and logic analysis and deep stuff like that, or digging up the schematic, or both, it’s not an easy one to diagnose past just checking traces for continuity around the PMU chip and switches.

Unfortunate, but I guess you can’t win em all. You often don’t know what’s happened to a board, especially after it’s been pulled out of a case, or why it was pulled out. I guess maybe this one was shelved due to not working? Who knows…
 

Big Ben

Well-known member
Sadly no, just a kid who wanted to put a pc in the B&W case 😢

As for the low voltage of the battery I had my suspicions, so I did a lot of testing and cuda resets.

PRAM seems fine, passed multiple MTP runs without any trouble. But MTP is not good as AHT I suppose and no AHT is available for this machine.

Guess it’s a lost cause without schematics, rev1 no case… It would have been a neat G3 in a new case but… a B&W is 25€ on tue market nobody wants them.

Thanks for your time.
I guess I’ll still try to fix it when I’m bored
 

SiliconValleyPirate

Well-known member
Hey, I have a PC in a graphite G4 case (I bought the case already empty from eBay, so I a wasn’t responsible for said G4’s demise!) so I can’t exactly crow about how awful that is ;)

As a penalty clause I bought 3 G4s during the build at silly-low prices and used one (with broken handles thanks to poor packaging!) as a donor and rebuilt the other two to working condition xD
 

Big Ben

Well-known member
Got absolutely not problem with that.

Just that I knew from the kid that the computer was working (actually both have G3 and G3 MDD parts), he gutted parts throwing everything in a big plastic bag. (ESD and scratches here you are)

I just think there’s better ways to do it. Buying an empty case is fine I guess.
 
Top