[blacksmith]

Several weeks ago, I decided to turn on my SGIs, to see if everything is OK with them. I am aware that someday they will not start anymore :-( . In my hardware check-up, I started with the older ones, and everything went OK until last week, when I tried the Fuel. It doesn't start anymore. Initially I thought that is a faulty power supply. Today, though, it was a rainy day in GTA and I had the time connect to L1. Attached are the L1 Print-Screens. Any suggestion?
Thanks.
Blacksmith            

[Raion]

I think your ENV monitoring chip has failed. That looks like how my Fuel behaved.

The power supply overtime goes out of regulation and takes out the env chips.

[weblacky]

I don't remember where this pic comes from (not me) with all the Fuel ENV chip locations.

   

I think if one IC goes shorted, it seems to drag the entire i2c bus down and all communication goes for all i2c devices on that bus just stop.  Sometimes if theses don't go short you just get 0 or erroneous values for voltage and fan RPM speed.

[dancingfools]

Hi, I think I have the same problem too, but in addition to these messages in the picture it does not give me. Keep sending this endlessly.

[Mark_G]

Little bit off-topic, but I've 2 fuels which didn't run for 3 years.
They have still the original PSU, and I want to avoid the env disaster.
Instead of power them on, should I preventive replace the PSU with an ATX PSU and the FUEL ATX PSU Adapter?

Mark

[Raion]

Yes. And I will repeat it again and again until people stop killing their graphics cards: if your environmental monitoring chips fail you must get them replaced. Do not turn off your environmental monitoring and run the system normally because the cooling system on fuels is heavily flawed.

It doesn't take a college degree in engineering or even just understanding how air dynamics work to understand that with air flow expected to travel at right angles through tiny fans can produce fan stall. It is evident in the higher-end system, the tezro, that this was a flawed design as they corrected their mistake by forcing all of the air through front to back instead of through right angles and such.

The only reason why I dislike fuels is because of this design flaw that I have seen people literally burn up the supply of v-pro models that fit the Chimera systems

[blacksmith]

Let's assume the followings:
- the defective components are the environment controllers.
- the power supply might be out of specs too.
- I will find an electronics repair shop that is able to replace the 3 DS1780E
- I will replace the power supply with a new one plus ATX adapter from ebay
After this replacements all hardware components are in good order.
My main fear is the L1 controller, or wherever the info about the machine is stored for stat-up, might be lost or corrupted. 

It is possible to reset the info in this components in such a way that the machine starts?

Thanks.

[Raion]

The L1 itself isn't a major concern here. It's base stuff is stored in a nonvolatile storage. It's not gonna get corrupted easily. Should that be a concern, you can flash it, but it's not fun. Lemme warn you there.

[weblacky]

L1 has two firmware blocks to help prevent bricking, we've never seen one "fail". As Raion is hinting, there is an upgrade order to L1 (due to the fashing utility firmware format).  Older firmware is flashed with older version of flash utility, new with newer, newest with newest (stupid I know).  So based on the version you are "at" you have to run an OS version sequence to flash up to the latest...once at latest you can run ANY Irix version supported by the system, (they don't care if you run new firmware on older OS releases).

I'm being selfish here...but if you HAVE ENV disease, please hold off for a moment as I'm trying to spin up a side hustle doing this work myself.  A shop is going to charge you a HUGE amount of money...do it poorly, (melt surrounding items) have no respect for what it is, and basically cook your board with hot air (which I don't use) trying to remove the chip like they'd remove a Nintendo Switch charging port.  As a community, I'm hoping we can offer these kinds of services fairly soon (be patient).  But they'd likely give it to the  "new guy" as easy...he'd likely blacken the board trying to remove the chip.  I have the tools to remove the chip without hot air and I don't use hot air at all on most things (not needed, too risky) that have legs.  I have used hot air to "finish" the small soldering job to pull in solder between legs...but that's lower than removal temps/duration.

These kind of concerns are being addressed, it just takes some time.

I'd say if you're worried, and want to run the system(s) now and not wait, buy the ATX PSU adapter and a compatible ATX PSU) and run the systems.  If you find you have ENV disease, it needs to isolated which component regions it is (3 ICs on mainboard, 1 on CPU PIMM, 1 on Graphics card...I don't know the location of the graphics card ENV IC at this time).  Normally you get a hint by the numbers or errors on which of the regions it is.

In terms of the original FUEL PSUs, I'm hoping to have rebuilt models in less than 1 year.  I've done SGI Indy and have done one Tezro (and have two more Tezro PSUs to do first), then I'm heading to FUEL (Fuel has 4 possible PSUs, but I only have one version, the others would require extended time to research and rebuild).

Things are moving but a new PSU isn't going to cause ENV disease, so I'd just go that way.  If ENV disease does happen...don't panic...it's fixable!  


We're here and this is known stuff...your not at the forefront of SGI failure on these :-)

[Shiunbird]

Is there any advantage in terms of functionality and/or performance to staying with the original PSUs and not migrating to ATX+adapter?