Getting Started, Documentation, Tips & Tricks

Found an Indigo - Page 2

What do we know about the failure? It was ahead of the fuse (or the fuse failed in a spectacularly unsafe manner), and it was a fairly high-capacity leak (to trip a circuit breaker before it cleared itself).

Check that part of the PSU.

_________________
Damn the torpedoes, full speed ahead!

Systems available for remote access on request.

:Indigo: :Octane: :Indigo2: :Indigo2IMP: :Indy: :PI: :O200: :ChallengeL: :O2000R: (single-CM)
Yes, it's time for PSU design 101. I'll take out the fuses (2) today and check them - they look ok. The other clue that I have is the speaker failure (no chime at start-up). Something must have blown up though, there was a big bang from within the case when the PSU failed.

_________________
SGI machines: Image and Image
fuses (of the miniature type) don't make any sound at all when they melt. likely you had an electrolytic capacitor rupture, or another component that burned up. the force of the explosion it made could have blown it clear of the board leaving nothing behind but some solder pads. You said something about a smell?

_________________
:PI: :O2: :Indigo2IMP: :Indigo2IMP:
twix wrote:
Yes, it's time for PSU design 101. I'll take out the fuses (2) today and check them - they look ok. The other clue that I have is the speaker failure (no chime at start-up). Something must have blown up though, there was a big bang from within the case when the PSU failed.


Chances are that whatever it was cleared itself in the spectacular fashion you noted, but since it tripped your mains 15A breaker the fault is probably before the fuse - perhaps in the Y-filter? If a cap goes there's a boom and smoke, but the PSU fuse (usually under 2A) goes way before the mains breaker trips.

_________________
Damn the torpedoes, full speed ahead!

Systems available for remote access on request.

:Indigo: :Octane: :Indigo2: :Indigo2IMP: :Indy: :PI: :O200: :ChallengeL: :O2000R: (single-CM)
Those high capacitance electrolytics in power supplies are freakin' dangerous, they go off like fireworks when they short internally, and if there are others in parallel they dump themselves instantly into the failed one, thus compounding the disaster... :shock:
guardian452 wrote:
twix wrote:
Well, I took the machine apart, as shown here:

http://www.catwhisperer.co.uk/indigo/index.html

This is a new one for me. I think I'm gonna be sick :cry:

Oi, fuck me ... me, too :(

</head>

<body bgcolor="#000066" background="../furniture/wall_medi_tile_02.gif" text="#CCCCCC">
< table width="800" name="450" border="1" cellpadding="3" cellspacing="0" bordercolor="#FFD36C" bgcolor="#24003C" class="PageTitle">
<tr>
<td align="center" valign="middle"> <div align="center">
< table width="790" name="50" border="0" cellpadding="5" cellspacing="0">

<tr>
<td><p class="PageTitle">SGI Indigo<br>

What in the hell is wrong with people these days ? Something in the water ? Brain disease ? Aliens sneaking around at night performing lobotomies on the natives ? The web has been here for almost twenty years but people haven't figured it out yet ? This is discouraging :(
I'm not seeing the problem here. Maybe it's about design, maybe he wanted it aligned to the left and be exactly in that size? Ok, it maybe looks like shit, but that's in the eye of the beholder.


Anyway, just my thoughts.

_________________
Image Image Image Image Image Image

-----------------
Old polygon wrangler
___________________
http://www.edgeloop.se
sgtprobe wrote:
... maybe he wanted it aligned to the left and be exactly in that size?

Exactly ! This is the world wide web , not an A4 sheet of paper.

I haven't done this in a while and apparently the world has changed drastically but every single html book / tutorial / lesson plan / class /whatever in the past made a huge point in Chapter One of the fact that the web has a vast array of display devices. You can't decide on your own that 800 pixels is a suitable number to "make it look exactly that wide." If anything, making a web site 'exactly 800 pixels wide' will ensure that it looks totally different on different devices.

The people who wrote html went to great lengths to put the presentation under the control of the user because only the user knows how large his display is, how bad/good his eyes are, how big the text should be, whatever.

This entire wave of idiotic webdesign is terrible. It's contrary to the very essence of the web.

By making it look "exactly that wide" he really made it look like this :

Attachment:
catshit.jpg
catshit.jpg [ 212.99 KiB | Viewed 180 times ]

Meanwhile, nekochan still looks exactly the way it was intended to. Hey, someone knows how to design html ! Woo-hoo !

And other people don't have an effing clew ...
I somewhat disagree. But I do understand what you are saying, just that I don't fully agree with you. In general it should be treated as you say, but sometimes you do want something to be displayed exactly as you have designed it. Just because it is on the web doesn't mean that it can't or shouldn't be designed as a sheet of paper. It's all about design in that case, failed or not. Personally I would make the window sit in center, and have the possibility to automaticaly adjust the text and images depending on the size of your browser window, but would look at it's best if your window is larger than the "design". Some times I rather nail the text/images so it sit just where I want it to sit. If your not going to adapt to my design, well, then move on then if that's so disturbing.

Sure, it sux if your gonna view a design made for 1024*768 pixels on a device only capable of displaying 800*600 pixels. But in all honesty, isn't it time to enter the year 2012? Or maybe we all should think of all the mobile devices and have one design that fit's all? In that case, I have to agree, since it's a much more elegant solution than to have a specific version just for mobile devices. But then again, did I or the client really care about that when the prototype sketches where drawn in photoshop?

But I think it all depends on how you are, I really don't think on the techincal side as much as I look at the visual design, I find stuff that changes disturbing at times, since it wasn't how I wanted it to look in first place, where somebody as you might find it extremely disturbing if it isn't using autoformating windows and such. But of course, everything also has it's place. A forum like this with hardcoded dimensions would be really stupid.

I mean, who's to say what's right or wrong? From a personal viewpoint only? If that's the case, I would simply say, deal with it ;) Everyone has an opinion, and as long as it works it seems it's only down to taste and personal preferences.

Anyway, just my humble opinion as of right now.

Hmm, I think we derailed this thread a little Hamei :P Sorry for that. Back to that lovely Indigo, I really hope that you get it up and running.

_________________
Image Image Image Image Image Image

-----------------
Old polygon wrangler
___________________
http://www.edgeloop.se
Yes, time to get back to the topic. Didn't have too much time to continue, but ......

- Found out that both fuses are blown. There's nothing between the fuses and the mains except the on-off switch, so that is not surprising.
- The speaker still works - I suspected it because of the lack of a chime sound when starting the computer - but it seems to be OK (4 ohm resistance as stated on the speaker itself
- Absolutely no trace of any failed part (caps, etc), no broken bits, no molten leads. Everything looks very clean.

_________________
SGI machines: Image and Image
Check to see if the bridge rectifier is shorted. Look for shorts between the AC and DC side.
Looks like I'll have to disassemble the PSU to get to all the parts and test them (as far as is possible). I hope that I will time for that soon (Christmas holidays?).

But I'm still thinking about the "clues" and how they could me find the cause of the failure. Three things stand out:
- no sound from the speaker after power-on. The speaker is wired to the backplane connector, so maybe something wrong with the backplane?
- the harddisk was unable to spin up (the light only comes on briefly after initial power-up). Not enough power available?
- Since the system came up and was going through diagnostics, the mainboard and graphics board were being supplied with enough voltage/current/power. So the PSU did work.

Any theories are welcome.

_________________
SGI machines: Image and Image
Lying loose inside the cage were also 2 plastic parts with holes that look as if they can be screwed to a hard-disk or CD-ROM. Some kind of sled. The number is 050-8031-001 (Rev B). I can't find them on any SGI parts list.

_________________
SGI machines: Image and Image
twix wrote:
Lying loose inside the cage were also 2 plastic parts with holes that look as if they can be screwed to a hard-disk or CD-ROM. Some kind of sled. The number is 050-8031-001 (Rev B). I can't find them on any SGI parts list.


Black body, green lever, IDC50+Molex on the end of pigtails? They're disk sleds.

_________________
Damn the torpedoes, full speed ahead!

Systems available for remote access on request.

:Indigo: :Octane: :Indigo2: :Indigo2IMP: :Indy: :PI: :O200: :ChallengeL: :O2000R: (single-CM)
No, it's not a drive sled, since the Indigo came with a (spare) drive installed. Here's a picture:

Attachment:
indigo_4.jpg
indigo_4.jpg [ 83.66 KiB | Viewed 171 times ]


I assume that one of these is attached to each side of a drive, since the holes line up. My guess is that it belongs to some other machine, not an Indigo, but it is definitely a SGI part.

Anyway, I tested the bridge rectifiers this weekend. Had to do some disassembly, but I managed in the end. The bridge rectifier on the primary side seems to be gone, as I am not measuring any voltage drop over the diodes, i.e. no reading at all.

_________________
SGI machines: Image and Image
Those things belong on the LG1 and LG2 ('entry') graphics boards.
Do a google image search and you'll figure it out.

_________________
Now this is a deep dark secret, so everybody keep it quiet :)
It turns out that when reset, the WD33C93 defaults to a SCSI ID of 0, and it was simpler to leave it that way... -- Dave Olson, in comp.sys.sgi

Currently in commercial service: Image :Onyx2: (2x) :O3x02L:
In the museum : almost every MIPS/IRIX system.
Wanted : GM1 board for Professional Series GT graphics (030-0076-003, 030-0076-004)
Yes, I see it now. The stand-offs are missing from the graphics card, so these "rails" cannot be easily refitted.

_________________
SGI machines: Image and Image
Still no further with the PSU, since I'm not much of an electronics man. I don't think that I can fix that.

Decided to install the spare Conner 500 Mb drive that came with the Indigo in a PC running Ubuntu. The disk tool doesn't recognize the partitions on the disk and shows wrong information. I first tried to install the partitions as XFS, did not work. Then decided that the Indigo is from 1991, that maybe the IRIX version is 4 and the file system is EFS. Mounting the partitions as EFS works fine. It appears that the disk has been partioned, IRIX (4) installed and some settings applied (network, etc). There is a root account and a guest account. I am assuming that this is a full basic installation, but without any applications.

The disk itself also works well in the PC hanging on an Adaptec 2940. It uses about 20 watts max when starting up, then about 8-9 watts in operation. So its not the drive that took out the PSU.

_________________
SGI machines: Image and Image