Nekochan Net

I think I'm going to puke

No, all warning messages always were about the +5V. Other voltages were within tolerances. Good point - maybe this can narrow the search...

Actually I've used quite a few Sparkle (Fortron) PSU's here for PC kit, they're really pretty good quality for the price and seem to OEM a lot of the more 'popular' units. Once stripped an Antec PSU apart, lo, it was a Sparkle inside...

_________________
<> x2+ x3 x2 x2+ x2 x3 x2+

Once you step up to the big iron, you learn all about physics, electrical standards, and first aid - usually all in the same day

I made the mistake of going to their website. Fortron isTaiwanian. In Kunshan I live surounded by Taiwanians. They make me puke. I used to think that Milpitas was the armpit of the Universe ...

It isn't.
The Fuel uses standard ATX-EPS connectors, but with different color codes and a completely different pin-out.
Installing a non-SGI PSU looks like a disaster waiting to happen...

Not that big a problem if you are aware of it - Dell is also notorious for changing the PSU pinout (but they don't have reliably nice PSUs on all equipment). Sit down with a VOM and a dental pick, trace out the voltages on the SGI and then, using a ATX pinout (see http://www.hardwarebook.net) pull apart the connector using the dental pick and reassemble. Done the rearrange a few times with various computer plugs.

Hamei sent me some pics of the Fuel motherboard, and it looks like environment monitoring is done with a couple of Dallas DS1780 chips which provide monitoring for 10 voltages per chip. I'd guess that the chip beside the L1 serial connector is likely related to PIMM voltages, and the one next to the power supply (right next to the "Timekeeper Snaphat" chip in the corner) monitors power supply voltages. Since the chips 'read' OK, one can assume that either the analog input to the chip on the offending line is bad (implying a chip problem), the voltage reference for the device is bad, or the origin for the derived voltage is bad.

Anyone know for certain if only the 1.5V line has the voltage problem?

Looks like replacement chips are available. If someone with a *new* motherboard can take a picture of the board around the DS1780 near the Timekeeper chip that would be useful.

_________________
<> x2+ x3 x2 x2+ x2 x3 x2+

Once you step up to the big iron, you learn all about physics, electrical standards, and first aid - usually all in the same day

Did you ever consider the possibility that the chips themselves might indeed be ok but the checking /algorithm/ is faulty?

Shouldnt than sgi can fixed that with a firmware upgrade? According to the post sgi replaced the boards.

regards
Joerg

_________________
http://www.irixworld.net

I have been reading this thread up to here with great concern, since all my full-equipped Indigo2 R10kq195MHz (Phobos G160, 1GB etc.) were recently replaced by R14k@600MHz Fuels...

About quality of PSUs: even the R10k "Impact" PSUs tended to fade (fail) some time (two of mine did).

There was a nice recipe for repair (replacement of dried-out capacitors in the PSU) later.

Make up some ATX PSU as a Fuel PSU replacement is not difficult? Good to know. Fingers crossed (mee, too) - here there´s no problem so far.

In case it is an issue of firmware on the mobo maybe SGI has fixed this already? You know, patches 5737 and 5404 AFAIK were about IP35 PROM - swmgr alerted for a PROM flash when istalling these on 6.5.22...

Ar am I heavily wrong here?

Walther

_________________
Valueing life is not weakness; disregarding it is not strength. -Mirage-

Exactly, that's what happened in my case - I simply got a new motherboard.

If it was software-only, as Joerg points out, SGI would have just reflashed the boards and saved the time and cost of replacing them. There has to be a hardware funky on the board that causes the monitoring to die over time.

Again, if someone with a 'new' Fuel mobo can take a picture around the 'Timekeeping Snaphat' chip (should be next to power supply stuff and the Dallas chip) I can do a comparison with the 'old' boards. Maybe it's just a faulty run of Dallas chips, who knows?

_________________
<> x2+ x3 x2 x2+ x2 x3 x2+

Once you step up to the big iron, you learn all about physics, electrical standards, and first aid - usually all in the same day

It seems that there are two seperate issues here with Fuel - (1) the environment monitoring can go south and (2) the PSU can fail with overvoltage. In Obelix's case, it seems to be the PSU since he has tested it with a VOM. Perhaps this could go somewhere in a Fuel FAQ or a SGI PSU FAQ (since they seem to have problems...)
along with the enviromental monitoring outcome (definitely recommend the use of a good DMM to verify PSU accuracy before disregarding EnvMon warnings).

I don't have a Fuel available, but it seems like a good thing for hobbyists to have would be a pinout of the Fuel power connector so it can be replaced by generally-available sources. (provided the ratings are correct). Anyone with a Fuel, VOM, and bit of time want to oblige?

In my case, the local SGI support couldn't say for sure whether it was hardware or firmware related - so I just got a new board.

Thanks to anyone who gave me informations and advice!

Just a final note to let you people know the end of the story of my broken Fuel PSU. Remember that I had the +5V line gone up to +6V, causing the machine to auto-shut down. Other voltages were right.
Eventually, it was impossible to adapt a normal ATX power supply to the Fuel, not because of the differences in the connector pinout nor because of the insane current requirement for the +3.3V line, but because the environmental monitor *needs* a feedback signal from the PSU fan, a thing that no ATX power supply ever had. No signal means fan failure, which means auto shut off.
So I did what I should have done from day one – I had the PSU serviced by a competent technician.
It did cost me *a lot* of money, yet just a fraction of the 550$ that a well-known supplier of SGI spares charges for a new PSU. The culprit was a single IC that with its different sections individually controls each of the four voltages.
The technician went lyrical about the excellent design and the build quality of the power supply – and with good reason. Once repaired, all voltages remained constant within a few millivolts, even under extremely unbalanced loads (25 Ampere from one line, zero from the others).
I put the PSU back at work and closely monitored its behaviour for the first three hours, and all voltages remained rock-steady up to the second decimal digit. The only changing figures were the speed of the fans and the temperature of the chips. Apparently, nothing was damaged by the previous over-voltage.
And (hopefully) that’s all, folks!