PowerMac G5 boot chime but no video out

[wottle]

Just picked up a really clean PowerMac G5 (dual 2.5 PCI-X). Owner said the video card was bad. I had another slower machine with a working video card, so I figured I'd just swap it out and have a nice, fast G5 as my bridge machine. When I first tried booting it up with the original graphics card I got the boot chime, but no video out over DVI (which I somewhat expected).
So I swapped the graphics card with one from another G5 and initially I got video output, but there were no hard drives, so I shut it down. Added a hard drive from the other G5 and OS X started to load and then it looks like it hung.

Wanting to rule out something weird with the HD, I threw the boot CD that came with the G5 and restarted, holding down the C key. Boot chime with no video again. I've now swapped for a 3rd graphics card, put the original, zapped the PRAM, reset the SMC, restarted again. Tried throwing a PCI VGI graphics card in there... No video out on either video card. Tried removing all the RAM but one pair, used RAM from the working machine, no change.

If I leave the machine running for a few minutes, the fans increase until they are max speed and then it just stays running at full speed until I hold the power button down to shut it down. The power LED on the front of the case is solid white (no blinking) and if there's a diagnostic LED on the motherboard (I know there is on my other PM G5), it is not turning on.

Any ideas what might be causing the issue? CPU?

I have a Dual 1.8 Powermac G5, but it has a PowerPC G5 970 CPU whereas this one has the 970fx CPU. Not sure if I can try swapping the CPUs between the machines to rule out CPU failure as an issue. Also, can I run the dual CPU motherboard with a single CPU at a time to similarly test it out.

Any ideas would be appreciated, as I've seen several people with similar symptoms on the PM G5, but never seen a confirmed reason. Thanks!

 

[wottle]

So, I thought this earlier G5 2.5GHz machine was air cooled, but I now realize I was wrong. I'm going to hold off on doing anything with it until I can inspect the liquid cooling setup for leaks. I thought only the last year model had the liquid cooling. I'm guessing that alone will prevent me from trying a processor swap.

Still interested in any ideas why the video out wouldn't be working (working consistently).

 

[Certificate of Excellence]

Hmm, so the dualCPU 2.5ghz a1047 is the 97fx processor and has PCIx. The A1117 Quad core 2.5ghz has 970MP processor and has PCIe. First things first, did you try the dead card that it came with in another machine to see if it is in fact dead? Obvious things Id do is visually inspect the video card AGP port (assuming its a a1047 as your other G5 videocard fit (an a1117 would be PCIe) for debris that might be stuck in there causing intermittent failure. Hit it with canned air for good measure. Just for giggles, I'd take the G5 branded cpu shroud off as well so you can get a good look of the guts. That will tell you real quick whether you see two CPU heat sinks or a tangled nest of tubing and gaskets for a LCS.

 

[wottle]

So, I did try to "broken" AGP card in my working G5 and it worked fine. I also tried throwing an additional PCI graphics card in there (it only had VGA out) and I didn't get vide out on either the DVI from the AGP card(s) or the VGA out on the PCI card.

However, I just removed the AGP card and only had the VGA card installed and I was able to get VGA out and it booted to the desktop (and then hung). There's definitely something wrong with this machine. I can now sometimes get it to boot to OS X's desktop via the PCI card, but it hangs.

The fans are also kicking in to a high speed. I wonder if the CPUs are overheating and it's shutting down for protection?

 

[Certificate of Excellence]

So, I did try to "broken" AGP card in my working G5 and it worked fine. I also tried throwing an additional PCI graphics card in there (it only had VGA out) and I didn't get vide out on either the DVI from the AGP card(s) or the VGA out on the PCI card.

However, I just removed the AGP card and only had the VGA card installed and I was able to get VGA out and it booted to the desktop (and then hung). There's definitely something wrong with this machine. I can now sometimes get it to boot to OS X's desktop via the PCI card, but it hangs.

The fans are also kicking in to a high speed. I wonder if the CPUs are overheating and it's shutting down for protection?

You could try and boot into Apple system diagnostics ASD 2.5.7

All Apple Hardware Test (AHT) & Apple Service Diagnostic (ASD) disks - Macintosh Repository

Apple Service Diagnostic (ASD) is Apple's bootable hardware testing console that lets you run extensive technical test routines on your...

www.macintoshrepository.org

See what the diagnostic results say. Also, did you try and boot into verbose mode? Seeing where in the boot cycle the machine hangs could shed light as well. On a whim, did you reset NVRAM in Open Firmware? Here's how you do it.

Command, Option, O and F keys at start up bong. At the prompt enter:
reset-nvram
hit enter key
set-defaults
hit enter key
reset-all
hit enter key

After this your mac should restart and boot up.

 

[wottle]

So, I've been distracted with other things, but I've been unable to get it to boot successfully from the AHT / ASD disk while also putting out video.

I'll try getting into Open Firmware if I can get video out again. And also try booting in verbose mode. It's inconsistent in how far into the boot it gets when it freezes,, which is what makes me suspect there may be a cooling issue and the CPUs are overheating (that and the fans ramping up to full speed even when I can't get the video to output.

 

[Certificate of Excellence]

Its not bad to take those apart, clean, repaste and reassemble. I'd say a 20-30 minute job between the two with the right tools. It's a big PITA if you dont have drivers long enough though. I wonder if you turn the case on its side so the weight of the heat sinks (and gravity) are pushing down on the cpus - if you could get consistent booting.

 

[wottle]

Yeah, I've done the air cooled ones and it's not bad. Just never the liquid cooled ones. I have a Torx driver that worked to get past the heatsinks on the air cooled ones. Hopefully it will work to get the liquid cooling system out. While I do that, I'm probably going to redo the liquid cooling because of their known leakage issues.

Also, I did boot into Open Firmware and reset the nvram. No improvement. A pattern I'm noticing is I only get one boot with video. After resetting the nvram, it reboot and I got no video. So I shut it down. Waiting until next morning and it booted with video (but hung on a gray screen while I was trying to use the C key to boot from the diagnostics CD). This is (probably falsely) making me suspect it could be something heat related (the time between starting up with video requires the component to cool below some threshold before using again?).

 

[Alf Torp]

I could not see it mentioned, but then again I am notoriously bad at thorough reading... pull all the memory out. Put in known good sticks if you have any, or memory not from the first slot. Disconnect everything else the computer does not need to power on. If you still don't get changes my bet is either the main board or one of the processors. I don't think the computer will boot with just one processor, but you might get some change in symptom by attempting to start with just one. I experienced something similar with a G5 years back. Faulty memory...

 

[wottle]

I could not see it mentioned, but then again I am notoriously bad at thorough reading... pull all the memory out. Put in known good sticks if you have any, or memory not from the first slot. Disconnect everything else the computer does not need to power on. If you still don't get changes my bet is either the main board or one of the processors. I don't think the computer will boot with just one processor, but you might get some change in symptom by attempting to start with just one. I experienced something similar with a G5 years back. Faulty memory...

Yeah, I tried removing everything not necessary when I first started troubleshooting. No change.

So, I finally decided to avoid the inevitable and took the CPU / water cooling assembly off and confirmed what I suspected (but was hoping I was wrong). The cooling system was leaking, but not enough to reach the PSU. However, it did leak and was corroding a good bit of the CPU daughterboard (is this CPU assembly considered a daughterboard?).

As I was trying to use a toothbrush and isopropyl alcohol to see now bad the corrosion was, caps were just falling off the board. When I was done trying to remove the corrosion, I think about 10 of the capacitors (I think they're capacitors) had fallen off.

I think I know the answer, but is there any way to restore this, solder on some new caps and recover these CPUs? If not, can I use air cooled CPUs from a 1.8GH PCI-X model I have that seems to still work (but I don't use because it smells like cigarette smoke when I run it ; tried everything to de-oderize it, but I think it's just ingrained in the power supply).

Seeing the state of the corrosion, I'm actually surprised it even booted. When I turned the machine on, I heard something rattle away, and I'm assuming it was a cap from the bottom of the board, already fallen off.

It amazes me that they decided to use a corrosive, conductive liquid for the LCS on these machines!

 

[Alf Torp]

Yeah, I tried removing everything not necessary when I first started troubleshooting. No change.

So, I finally decided to avoid the inevitable and took the CPU / water cooling assembly off and confirmed what I suspected (but was hoping I was wrong). The cooling system was leaking, but not enough to reach the PSU. However, it did leak and was corroding a good bit of the CPU daughterboard (is this CPU assembly considered a daughterboard?).

As I was trying to use a toothbrush and isopropyl alcohol to see now bad the corrosion was, caps were just falling off the board. When I was done trying to remove the corrosion, I think about 10 of the capacitors (I think they're capacitors) had fallen off.

I think I know the answer, but is there any way to restore this, solder on some new caps and recover these CPUs? If not, can I use air cooled CPUs from a 1.8GH PCI-X model I have that seems to still work (but I don't use because it smells like cigarette smoke when I run it ; tried everything to de-oderize it, but I think it's just ingrained in the power supply).

Seeing the state of the corrosion, I'm actually surprised it even booted. When I turned the machine on, I heard something rattle away, and I'm assuming it was a cap from the bottom of the board, already fallen off.

It amazes me that they decided to use a corrosive, conductive liquid for the LCS on these machines!

Easily fixable, the other processor will be the blueprint @Branchus can probably give you some soldering tips