[Irinikus]

Here you go:

[indy99]

Your mention of issues starting sometime after adding the Audigy board are interesting. If you're willing to try, I'd suggest removing all of the PCI peripheral boards and trying the system in a minimal power consumption configuration to see if that has any effect on the issue.

Ran into a somewhat similar situation while testing various PCI boards in my Tezro and found certain combinations of boards and their relative slot position could cause the system not to boot.

[indy99]

When I originally ran into the PCI power issue I made a post on nekochan detailing what I found. Unfortunately couldn't turn that post up in a search of what's left of that forum, and it's been long enough I don't recall the specifics.

Fortunately all the tools I used are still easily available to any Tezro/O350 owner via the L1 port.

I'd probably start with the L1 "pwr" command. "hlp pwr" will give you the list of available pwr command queries, so "pwr vrm" will show you the status of the system vrm modules. As an example with the system not running "passed" in the "Present" column, and pwr off in the Okay column. While you're there, check the status of the non running system with the "pwr" command. My 800Mhz system may be different than your 1Ghz, but when not running the values in the "Margin" column show "normal" for 1.8V, 3.3V, 2.5V, XIO 2.5V, IP53 VCPU, IP53 SRAM and IP53 1.5V. All other values are shown as N/A.

Next I'd look at the L1 command "pci". As I recall this is how I was able to determine the PCI bus topology and resolve my Tezro's failure to power up issue. There are four PCI busses spread across seven useable slots.

I'd run the PCI command and record the results of the pci query. Record the original locations of each board then disconnect power and remove all the PCI peripheral boards, except of course the IO9. Reconnect power, cross your fingers, and try restarting the system with the L1 command "pwr up". At this point I was able to start up my Tezro, and following that added back the PCI boards one by one until I found the combinations and positions that were stable.

Currently this system has the IO9 in bus 1, an Adapter red PCI Firewire board and an OEM SGI PCI 4 Slot USB board in PCI slots 3&4 (bus 2), and a SAS PCI HBA in slot 7, bus 4.

Best of luck, and hopefully you might find some of this helpful. There's quite a bit of useful information available via the L1, but run "hlp" at the L1 prompt for a list of commands.

[weblacky]

This isn't the exact error, and I never had the fan errors.

But I had an old post on Nekochan (long gone) for a Dual 800Mhz Tezro that had 5V low ref failures that looked exactly the same (just complained of 5V instead of 3.3V). System would boot and work fine for like 5 minutes, then suddenly show up these errors (always within 10 minutes of start-up at the longest) and they'd continue until I shutdown.

I changed the VRMs (still have them all) and such and it didn't help. I started taking PCI cards out and I found when I removed the Adaptec firewire card that everyone claims works (red board), 5V errors stopped, system ran for over 30 minutes with no issues. Called it solved, powered down and haven't used it since. So the Firewire PCi card was over drawing the 5V ref circuit (causing low voltage).

PCI cards can use 3.3v or 5v. Pull the PCI cards, likely the voltage errors will go away. The fan errors are something else (check fan connectors and movement and such.

Thinking about it all recently, I think (no proof), it's a sign my power supply is getting weak and the PCI card that SHOULD work, end up causing low voltage issues when you really load cards in the system. I have an itching in the back of my mind that if I were to rebuild the PSU or get a hacked replacement, that putting the card back might work. Again, no proof, but signs of "inadvertent overload" could easily be a weakened PSU.

I suggest you run it lean, remove all non-essential cards and see what happens.

The fan is another issue.