Nekochan Net

Hello,

I got a SGI O2 coming from a trash bin . Everything inside looks ok, but the PSU is missing.

As I have plenty of ATX PSUs, and I rewired some PSUs for old HP netservers and old Dell Optiplex, i think it would be possible to connect an ATX PSU to this O2 .

The only problem is that I don't know the pinout of the PSU connector, and an one can't guess everything with a ohmmeter (except the ground, +5v and +12v).

Does someone have the wiring scheme, or a photo of the connector ?

I guess this is the PSU I miss http://www.sgidepot.co.uk/sgidepot/pics ... 32-001.jpg

For the 3.3V "AUX", I can have it by regulating the ATX's +5V "SB".

Hi, the connector on the P/S is a Molex Mini-Fit BMI (Blind Mating System) 15-06-0241.

I am not sure that ATX PSU will work, this is a special supply with active PFC and not-standard voltages.

The pinout is (facing the power supply):

[ ?? ] [ ?? ] [3.3sb] [+ 12] [- 12] [Gnd] [Gnd] [Gnd] [Gnd] [3.43] [3.43] [3.43]

[ ?? ] [!POn] [5.10] [5.10] [Gnd] [Gnd] [Gnd] [Gnd] [Gnd] [3.43] [3.43] [3.43]

One of the ?? is the pwrgood signal, I am not sure which. probably the left one on the lower row. The other two ?? may be returns or some other function.

All at your own risk.

_________________

Thanks, I will try.
The non standard voltages aren't really a problem, usual PSU, including ATX (and I suppose SGI's) have a 5% to 10% tolerance.
If the O2 is not tolerant to ATX voltages, It's possible to get these by changing the ATX PSU's feedback resistors.

If you can get this going, it will be pretty interesting for the future. Right now you can still find an O2 PSU with a little searching, but who knows how easy that will be in another 5 or 10y.

_________________

Important point needs to be taken here - were the voltages measured under the load ? (then 5.10 might be showing instead of 10 etc).
Anyway, There must be a bunch VOM's somewhere in the O2 (anyone dare find them ), which should have pretty high tolerance to the input voltage, in addition to 5V going directly to: PCI bus (if that's not the 3.3V only, can't remember), CD-ROM and HDD"s. Same for 12V.
3.3V might be powering some of the logic directly, and also be converted to say 1.5V for PIMM (depends on the PIMM of course).

BUT - depending on the VOM, there is (very small though) chance, that 5.10 may be really adjusted to 5.00 by some VOM which would then have to be extremely efficient with LDO (low-dropout), but even those need slightly more voltage (10-20%) afair - that would rule this case out, but if this were true, then those voltages really have to be like it says.

In the very short a little more voltage should not be a problem - as long as it's not too far away.

_________________
[click for links to hinv] JP: [ :Fuel: :Octane2: :O2: :Indy: ] PL: [ :Fuel: ]
For Sale: 2*O200 M/B, 2*O200 PSU, 6*256MB O200 RAM, 2*O200 SCSI Backplane, 2*O200 MSC, DMediaPro DM-2 ( 030-1653-002 Rev. H , XT-DIGVID) with Octane XIO pull (Origin pull optionally available)

Hard to believe the O2 is sensitive to a 0.1 volt drop. The ATX specs say +/- 5% tolerance which puts the O2 within +5% limit. I'll be getting my hands on one soon enough to take some readings.

5.10V and 3.40V are the rated voltages on the label of the power supply. If they are much different under load that would mean quite poor load regulation. The supply uses cermet feedback trimmers, that seem to be factory adjusted very close to the spec.

The O2 uses 5V PCI signalling and most of the I/O devices run on 5V. the chipset and cpu use 3V3 and 1V8. I think you are right that LDO regulators cannot drop e.g. 5V1 to 5V, it is more likely that the supply voltages are specified slightly high because of passive losses in the design. Remote sensing may also be used, there are two unknown signals that might be involved there.

I wouldn't be too quick to assume what the margins are. O2 is one of the most sensitive machines to power irregularities, and many will suddenly not turn on anymore when power is removed in the wrong way (reminiscent of Tek 7000-series scope plugins, there is a warning label that you can only see once it is too late and you have fried the device)

_________________

hi robespierre! thanks for this info really helped me confirm my multimeter reading

btw i have questions for the following pins: (upper pins 13->24, lower pins 1->12)

PIN 15: [3.3sb], what do you mean by sb? is this the one for sense stated on ATX?
PIN 1: [ ?? ], i read 3.44V here, is this output from PSU or mainboard?
PIN 13: [ ?? ], i read 0.009V, maybe this is pwrgood? coz in ATX is also gray, but shouldn't it active high? like giving a higher voltage like 3.3V? or this is also active low?
PIN 14: [ ?? ], this pin is color BROWN on my side and i can read 3.3V, maybe this is the pwrgood? or +3.3V sense like in ATX which is brown too?

btw, i got all reading with load running, i mean O2 is fully running. So maybe i should redo the reading with no load? can i safely put [!POn] directly to GND? or should i bias something?

_________________
(Sakura) (Sasuke) (Naruto) ... lil Jesse! (O2 laptop)
“Imagination is more important than knowledge.“ – A. Einstein

i got it, it means standby but on ATX its at +5V hmmm

_________________
(Sakura) (Sasuke) (Naruto) ... lil Jesse! (O2 laptop)
“Imagination is more important than knowledge.“ – A. Einstein

This could easily be regulated using LDO VOM, there are a few models which could take 5 and produce 3v3.

_________________
[click for links to hinv] JP: [ :Fuel: :Octane2: :O2: :Indy: ] PL: [ :Fuel: ]
For Sale: 2*O200 M/B, 2*O200 PSU, 6*256MB O200 RAM, 2*O200 SCSI Backplane, 2*O200 MSC, DMediaPro DM-2 ( 030-1653-002 Rev. H , XT-DIGVID) with Octane XIO pull (Origin pull optionally available)

thanks kuba! yes i know but my intention is not to use ATX, i'm planning to build my own PSU for my O2 laptop project, so need to know which pins are PSU output and which are not. my tentative game plan is to buy an existing DC adapter like 24V maybe then 32A capable, from there will chop it to the specific voltages that the O2 board needs this is possible ayt? progress a bit slow coz too many pins to decipher one by one hehe my eyes sometimes rolls outside hehe
na ja, no pain no gain

_________________
(Sakura) (Sasuke) (Naruto) ... lil Jesse! (O2 laptop)
“Imagination is more important than knowledge.“ – A. Einstein

still designing my PSU but got this question.

there are 6 pins of 3.43V and 2 pins of 5.1V
then according to the label outside: 3.43V/31A then 5.1V/10A

does it mean each pin of the 6 pins can source 31A? or should i 31/6 = 5.17A source for each pins?
coz i'm looking for a ready made PSU then ill use DC-DC converter to produce each needed voltages

_________________
(Sakura) (Sasuke) (Naruto) ... lil Jesse! (O2 laptop)
“Imagination is more important than knowledge.“ – A. Einstein

The power supply can put out that current total across all pins, if you check the continuity they are most likely wired to the same source and shorted together. If one pin tries to draw full current it will probably melt the connector; each pin is designed to handle approximately 5A so they work in parallel.

It is also possible for a low-power supply and high-power supply for the same voltage. I.e. 12V at 2A and 12V at 10A in the same supply, but connected to different things. This is how it works if the same voltage appears twice on the label.

_________________
(Maradona) (DavidVilla) A1186 (Xavi)
A1370 (Messi) dp43tf (Puyol) A1387 (Abidal) A1408 (Guardiola)

"InfiniteReality Graphics - Power Through Complexity"

aha! so it really means 1 pin only can source 5.17A, NOT 31A right? thanks for this confirmation!

ah ok2 noted, thanks!

so ok, to decide which PSU ill buy, should i total all currents from the specs like: 31A + 10A + 3.2A + .25A + 10mA = 44.46A?
OR
i just choose the biggest value which is 31A?

but what wonders me is output Power = 170W only

_________________
(Sakura) (Sasuke) (Naruto) ... lil Jesse! (O2 laptop)
“Imagination is more important than knowledge.“ – A. Einstein