Octane2, fan spins up when plugged in, power switch has no effect.

KayBee

Member
Feb 24, 2020
61
33
18
Hi All, I recently got an Octane2, as soon as I plugged it in the fan spun up, and the power button has no effect. The light bar shows no signs of life. I disassembled per the manual, the System and Graphics module look clean enough to eat off of. Not much dust inside, but I cleaned everything anyway with compressed air can. No signs of arcing pins on the PSU, it all looks very good, compression connectors look new. It appears to have low-end graphics, which is fine by me. Looks to be full of RAM though. Once reassembled, same result upon plugging it in, fan starts up immediately, power button has no effect. I tried without hard drive as well, no difference. The troubleshooting flowchart shows I think possible CPU or front-plane failure, which of course I hope it isn't.
PS, I was surprised to see an empty ROM socket on the motherboard, what's that for?

Thanks.

KB
 

Elf

Storybook
Feb 4, 2019
446
113
43
I probably have no useful advice, but in terms of troubleshooting and curiosity, does the fan spin on the power supply if you plug it in while it is outside of the system chassis?
 

KayBee

Member
Feb 24, 2020
61
33
18
What? That's crazy! Kidding, sorry Canadian humor. Interesting idea I'll carefully try that. I do have a functioning original Octane, and both have the Cherokee power supply so in theory I could borrow the known good one. I am a little paranoid to try that for some reason, like I could damage the good one.
 

KayBee

Member
Feb 24, 2020
61
33
18
Removing the PSU from the Octane2 and plugging it in, nothing happens, no fan. Thanks for the suggestion anyway.

KB
 
  • Like
Reactions: Elf

stormy

Member
Jun 23, 2019
60
23
8
Sounds like something serious as you mentioned, since you swapped the PSU and it didn't come to life. If the system had any life in it you would at least see a flashing red light with the power button working. Sounds like a faulty front plane as you said. I guess you could part it out.
 

weblacky

Member
Jan 13, 2020
79
17
8
Seattle, WA
Really dumb question, just to finish off any user error, are you trying to turn it on using the front plastic button, or did you remove the front plastic and poke the actual switch the power button tries press? I only ask because without lubrication (silicone) on the plastic button, the button tends to bind during travel and not make good contact with the switch it's supposed to press.

So just for completeness, did you try to push the actual power switch inside the metal front frame? Because much of this sounds like it never even got the power-on signal. Failing that, I'd agree with others and check the switch in the front plane and also the solder connections/plug wear on the PSU elements in the front plane.
 

KayBee

Member
Feb 24, 2020
61
33
18
No, that's a good question. Yes I had pulled the front off and then pushed the physical button, but thanks for suggesting it we all miss the forest for the trees at times. I will take a look at the switch itself and the front plane connections when I can open it up again.
 

weblacky

Member
Jan 13, 2020
79
17
8
Seattle, WA
When you say fan, I assume you mean a system fan, right? Does it just continue to spin or does it spin and then stop?

I don't think anyone has a pinout of the PSU port yet, but wires carrying the same voltage are normally grouped together on the PSU's connector. Let's say your PSU isn't the problem (still not 100% sure but okay). Remove AC cable from PSU, remove PSU from Octane, then you need to drain all the pins on the white output connector (on psu) against its metal case (one by one), use wire, paperclip, whatever. Then use a voltmeter and check every pin on the white PSU connector against it's metal chassis as well. Check that 1V or less is present (residual, less is better, 0 is best).

Now, assuming you have a working PSU, you need to record the pins have under 1 Ohm in resistance mode between the white PSU port pins and the metal chassis. This SHOULD give you all the ground pins. Record the values you get for which port positions.

Armed with that knowledge.


Sounds to me like a lower voltage rail (3.3v/5v) is shorted inside the Octane. So if you probe along the PSU connector on the mid-pane, first probe grounds (same pins you found on the PSU) to confirm grounds, then probe other pins...I'd recommend DIODE mode for this and see if a group of pins comes up like ground...but doesn't come up as ground on the recorded values you wrote down for the PSU's connector. This may indicate a valid voltage source on the PSU is being grounded out on the mid-plane. If you find this is be case (which would be great info) then you need to disassemble the Octane and perform a current controlled, low voltage, bench PSU injection into the shorted midplan PSU pins and hunt down what's heating up (using hands/finger, IPA, lighter fluid = Naptha , or a thermal imager). If you're lucky it's a shorted grounding cap that can be removed for testing. If you find the short has disappeared if you only plug the midplan and NOT the mainboard or graphics boards, then that's a clue the short is in something you've removed. Try adding back boards until it happens.

This would likely be easier to check in reverse, if you find a shorted pin group on the Mid-plane connector while the system if fully assembled, remove optional boards, and recheck those pins, If the short is gone...then you know the short is in whatever you just removed. That's what may get you a better clue if it's in your mainboard, graphics/cards, or mid-plane.

This would be my approach if I knew the PSU was good. as I would assume from the lifelessness, that a voltage used by the actual system boards (3.3v or the like) is missing or being dragged down. If the system fan runs, you have 12v...because that's what the system fan uses.
 

KayBee

Member
Feb 24, 2020
61
33
18
Update on the Octane2, short version: it appears to be a problem with the Mainboard and I have decided to donate it to a friend.

Longer version: I have a couple of Octane1's and was concerned about just trying parts swapping, thinking I could do damage to a known working Octane. Weblacky and I were going to do some parts trading anyway, so he offered to help me prior to us exchanging parts. He explained a measurement plan that would reduce the risks and I ended up learning a lot. I unplugged the Cherokee PSUs from both machines and let them sit overnight to be sure they had no juice. With his guidance, I measured the resistance between the pins he specified on the Octane2 PSU to look for a short. We compared my measurements to his, there were some differences but nothing major. Then I measured my other PSU from the known good Octane1 and it was quite similar to the Octane2 PSU.

So I tried the Octane2 PSU in the Octane and it booted up just fine. No issues, great.

Then per Weblackys plan I measured the pins he specified inside the Octane2 itself, he did the same and we compared. Not much difference there either. So without significant anomalies so far I decided to take the possible risk and try the Octane1 System Module in the Octane2. It fired right up, booted normally, no issues.

Then I put the Octane2 System Module in the Octane 1 and no good, same as the Octane2, all 3 fans spin up as soon as the power cord is connected. So per his suggestion I removed and reseated all the RAM, no difference. Then I removed the Octane2 RAM and put good sticks of RAM from the Octane1 and tried again, no difference. Then I removed, used a little canned air on, and replaced the processor. No change, still the fans fire up as soon as the machine is plugged in, no picture, and the lightbar stays completely dark. Then put the known good Octane1 CPU in the Octane2 System Module and tried again. Still no good.

Oh I should also mention, the insides of this machine, and all compression connectors were spotless.

So at this point I decided that I have hit the end of what I have time and energy for, time mainly with ridiculous hours, a family, and learning C. I almost bought another mainboard from eBay, but I decided instead of rebuilding a functional machine, and then having more functional machines than I need, and no spare parts, I will keep its RAM, CPU, Cherokee, etc. as spares for my Octane1's. To each their own right? I do prefer the blue case of the Octane2, so I will put Octane1 internals in it, from an updated Octane1 with plenty of RAM.

Thanks for reading.

KB
 

weblacky

Member
Jan 13, 2020
79
17
8
Seattle, WA
I should mention, that the Octane2 CPU was tried on another Octane system's board and it worked just fine (that might not be clear from the above summary). So after a RAM swap, CPU swap, PSU swap, and entire chassis swap - only the Octane2's mainboard is the constant that has this issue following it. So it was shown the the mainboard has an issue itself (what, is unknown at this time). But the Octane2's mainboard issue didn't affect the CPU, RAM, or other peripherals attached on it. So those parts all functioned in other systems!.
 

KayBee

Member
Feb 24, 2020
61
33
18
Oh right Weblacky, I did do that swap as well, the Octane2 CPU (R12K 400) worked without issue in the Octane1 mainboard. Hey Chulo, I didn't measure the caps, so I don't know. physically (which may not be your point) everything in this machine is spotless.
 

About us

  • Silicon Graphics User Group (SGUG) is a community for users, developers, and admirers of Silicon Graphics (SGI) products. We aim to be a friendly hobbyist community for discussing all aspects of SGIs, including use, software development, the IRIX Operating System, and troubleshooting, as well as facilitating hardware exchange.

User Menu