Indigo2: populating memory banks
#11
RE: Indigo2: populating memory banks
I cannot find a good picture of the underside of the Indigo2 mainboard to see where the memory slots go all the way through? Anyone got a good picture of this board to know what kind of job we'd be signing up for? I'm leaning toward helping, but I need to see what I'm up against before I'd commit.

Also, anyone got a source for SIMM slots? New parts.
(This post was last modified: 08-13-2020, 05:57 AM by weblacky.)
weblacky
I play an SGI Doctor, on daytime TV.

Trade Count: (10)
Posts: 1,716
Threads: 88
Joined: Jan 2019
Location: Seattle, WA
Find Reply
08-13-2020, 05:55 AM
#12
RE: Indigo2: populating memory banks
Here's 72-pin slots for the Amiga. Should fit here no problem:

https://amigakit.amiga.store/product_inf...cts_id=594
Ninja: https://www.mouser.com/ProductDetail/TE-...IHpg%3D%3D Found these too!


These are angled, but assuming you orient them correctly (or worst case, replace all of them) it shouldn't be a big deal. Only new part supplier I know of though.

It's through the hole on both Indy and Indigo2, but I can't find any pictures of the motherboard's underside.

I'm the system admin of this site. Private security technician, licensed locksmith, hack of a c developer and vintage computer enthusiast. 

https://contrib.irixnet.org/raion/ -- contributions and pieces that I'm working on currently. 

https://codeberg.org/SolusRaion -- Code repos I control

Technical problems should be sent my way.
(This post was last modified: 08-13-2020, 06:50 AM by Raion. Edit Reason: Added another link. )
Raion
Chief IRIX Officer

Trade Count: (9)
Posts: 4,240
Threads: 533
Joined: Nov 2017
Location: Eastern Virginia
Website Find Reply
08-13-2020, 06:48 AM
#13
RE: Indigo2: populating memory banks
I'm trying to see how small the holes are, I've seen some holes so small, even a vacuum gun won't guarantee to clear it. I have a HAKKO desoldering gun and a few tips, I have a small preheater as well. I wanted to find someone with a board they can take a close-up to see how close the pins are together and how pronounced everything is. Again, beter this is 20 year-old tech than today for this kind of thing.

With holes this small, the tactic isn't to heat gun and remove the slots...that's suicide! The tactic is to try to use a desoldering gun to remove ALL the solder from each pin (while wiggling) to clear the tiny holes while the pins are still running through them (makes for much better heat transfer through the via). Assuming you free the pin in the hole during cool-down (keep wiggling the gun), then after you've done all 100 holes or whatever, you attempt to wrench the old slot out of dry, bare, holes. Heating a pin or two that look stuck. I have a single Micro soldering tip that can reach through those holes and still require a solder sucker on the other side to clear them...it's tedious if you leave solder in before extraction in vias this size. But it beats having to hot air the board. A basic prewarm to hedge my bets, then hope my desoldering gun can clear would be approach A, approach B is the opposite: try to remove the entire slot while warm and then trying to clear the solder-filled holes (that are tiny and desoldering braid will only half-empty) using either an iron with compressed air to punch out molten solder, or the manual method I described (stick a micro-soldering tip in the hole, and use manual solder sucker on opposite side of the board while heating the hole to clear it).

No, this is job is so hard...I'd only be removing bad slots, won't remove healthy slots, it will likely take nearly an hour a slot until I find something that works (which could be over an hour). It's a big job, I'd like to believe it can be done with patience and perhaps taking breaks over several days if needed (as long as it's working).

anyone that thinks you're going to hot air this thing and pluck the slots out, then magically insert new ones with the solder staying molten...hasn't soldered a huge motherboard before. You cannot do that without risking other parts being damaged or failing off the board.

synthetix:

You say there are just two damaged slots? Are they next to one another (a touching pair)? Are they on the extreme end of the all memory banks, towards the CPU? If so, that likely increases chances of success.

If you're going to remove the mainboard anyway, could you look into taking pictures of the rear side of the mainboard memory area (all in frame with CPU region at edge of the picture frame) and look into instructions to remove the CPU module from the mainboard as well. Then take another sets of picture of the front (top) of the mainboard focusing on the SIMM slot area with the CPU removed.

That will show me how much space we have with the CPU out and any hidden components around the memory slots that could pose an issue.

If we can get good pictures, I'd be willing to make a decision based one those pictures to help you in this matter.
(This post was last modified: 08-13-2020, 07:19 AM by weblacky.)
weblacky
I play an SGI Doctor, on daytime TV.

Trade Count: (10)
Posts: 1,716
Threads: 88
Joined: Jan 2019
Location: Seattle, WA
Find Reply
08-13-2020, 07:09 AM
#14
RE: Indigo2: populating memory banks
Thanks for the reply! I don't have the board out of the case yet, but I attached a photo of (I think) an IP22 R4X00 Indigo2 motherboard.
The slots that are damaged in mine are in bank C, near the CPU. One of the metal clips broke off one side of one slot, and the little plastic guides that go through the holes in the RAM stick are broken on another one (see attached photo).


Attached Files Image(s)
       

O2 Indigo2 R10000/IMPACT
(This post was last modified: 08-13-2020, 08:12 AM by synthetix.)
synthetix
gcc -O3

Trade Count: (0)
Posts: 45
Threads: 11
Joined: Aug 2020
Location: 🇺🇸
Find Reply
08-13-2020, 08:05 AM
#15
RE: Indigo2: populating memory banks
Hmmm, when you do remove the board (if you decide to try anything) that underside memory region has me concerned. Does your board really have those SMD component clusters between every slot on the underside?

That kind of tight neighboring is going to present problems.

If you can get a good close-up picture, then I can see better. The real issue is, if a vacuum desoldering gun cannot easily clear those vias, then you’d try to use another method but any other method would put those small neighboring components at risk.

I was hopeful for more clearance in that area, less chance of collateral damage.

Guess the big question is, with this much uncertainty, what’s your level of risk feeling like? Would you rather stay where you are or risk the board to remove those damaged slots?

Things might go good, or they might be an instant fight from the word go.
weblacky
I play an SGI Doctor, on daytime TV.

Trade Count: (10)
Posts: 1,716
Threads: 88
Joined: Jan 2019
Location: Seattle, WA
Find Reply
08-13-2020, 06:21 PM
#16
RE: Indigo2: populating memory banks
Okay, new development.
I removed all the memory from the machine, and installed it back in banks B and A only. It works! This way, I can completely avoid bank C altogether. So it seems the problem is solved for now.
I think the reason SGI instruct you to install memory in bank C first is for the simple fact that you can't physically install memory the other way around. If you start with bank A, the next module won't go in because it can't get angled enough to go into the socket.
I haven't been able to use the machine because the hard drive is no good, but once I get a replacement and reinstall IRIX we'll see how it runs.

O2 Indigo2 R10000/IMPACT
synthetix
gcc -O3

Trade Count: (0)
Posts: 45
Threads: 11
Joined: Aug 2020
Location: 🇺🇸
Find Reply
08-13-2020, 08:59 PM


Forum Jump:


Users browsing this thread: 1 Guest(s)