• Please review our updated Terms and Rules here

Restoring Multiple PS/2 Model 95s | Part III: Three Out of Six Ain't Bad Edition

ButINeededThatName

Experienced Member
Joined
Nov 29, 2018
Messages
176
Location
Fort Wayne, IN
THEY LIVE!!!

Well, three of the six do anyways. Two are still waiting for parts and the third shudders well, the third had "tenants" at one point as someone left a filler plate out of the machine while it set and apparently a family of nice decided to call it home. More on that later though.


IMG_20190705_191907.jpg

First up we have the 8595. This system originally would throw out errors 161 and 163 (Date and Time) and wouldn't save whatever you entered in. Obviously a bad battery. Well, while attempting to replace it the metal retainer that held the battery in place and contacted the positive side snapped off when I tried to remove the old one. After buying a replacement battery holder, desoldering the old one and soldering in the new one the date and time issues were easily resolved and this system was working, though I wasn't done with it yet. I ordered a bunch of parts off of FeeBay in order to start "standardizing" my systems, which will become apparent below when I list system specs, which for this system are:

A 50MHz Type M Complex
64Mb (8x8Mb Parity) of Memory
A 400Mb IBM Hard Disk
Latest Revision Cached SCSI Adapter
An XGA-2 Card
A Madge MC32 Smart RingNode (Apparently this is capable of 40MB/s streaming! Oh, and it has attractive red plastic instead of the typical IBM blue.)


IMG_20190705_194813.jpg

This is the first of the two 9595 systems I have. I received this one in a different lot of PS/2 gear I picked up a few months after I bought the lot of four 95s/85s. Aside from being dirty this one fired right up and was the easiest to restore. It's configured with the following:

A 60MHz Type 4 Complex
64Mb (8x8Mb ECC) of Memory
A 1Gb IBM Hard Disk
Latest Revision Cached SCSI Adapter
An XGA-2 Card
A Madge MC32 Smart RingNode
Do you see the theme yet?


IMG_20190705_201746.jpg

The second of the two 9595s I have. This was my first Model 95 that I made a nine hour round trip for. It's also the most cosmeticaly pristine system I have as it was used in a bank and well cared for. At one point I swapped the hard disks around and after which was unable to run the refdisk as the floppy drive had seemingly failed. After some cleaning of the drive though everything was fine and it was back to normal. Hardware wise it has the following:

A 90MHz Type 4 Complex
64Mb (8x8Mb ECC) of Memory
A 1Gb IBM Hard Disk
Latest Revision Cached SCSI Adapter
An XGA-2 Card
A Madge MC32 Smart RingNode

And now onto the three that are left...

IMG_20190705_201942.jpg

On the left you have one of two 9585s and on the right my 9595A.

Starting with the latter, this system was throwing out all sorts of CP codes and then one day, stopped POSTing all together. After much troubleshooting I found that the planar had failed, likely bad capacitors, and nabbed a NIB replacement planar off of FleeBay for $70. Not too bad I'd say. After replacing the planar the system POSTed normally and asked for a refdisk, however I currently can't give it one as I'm out of 2.88Mb drives and need a couple more. This system also had it's RAID cages, hot-swap drive trays and RAID controller stripped from it which borderline infuriated me, as I've been wanting a set to use ever since I brought home my first 95. Obviously, I can't consider this system truly restored until I track down the cages, trays and four-drop SCSI cable that it originally had (I have a couple of cached SCSI controllers around so at least I'm good on that front). Hardware wise, I hope to have it configured with the following at some point:

A 90MHz Type 4 Complex (Have)
256Mb (8x32Mb ECC) of Memory (Halfway there)
6x 1Gb Quantum Drives in RAID 0 (Need the drives, cages, trays and SCSI cable)
Cached SCSI Controller upgraded to 64Mb of Cache (Have)
An XGA-2 Card (Have)
A Madge MC32 Smart RingNode (Need)


Now onto the first of my two 9585s. This system has acted very strangely since I've had it. It powers on the moment you insert a power cable and will not power back off until unplugged. It also does not POST whatsoever. I suspect it has multiple issues, though the power supply or switch is probably one of them. Aside from missing a floppy drive, this system is pretty much configured how I want it with the following:

A 66MHz Cyrix DX2
32Mb (4x8Mb Parity) of Memory
A 1Gb Quantum Lightning Hard Disk
An XGA-2 Card
A Madge MC32 Smart RingNode (Need)

Finally, the second 9585. As mentioned, this system was home to some mice at one point and as such, pretty much everything aside from the case is ruined and even then that's fairly corroded inside. If I can save the case, I'll probably just recap my bad 9595A planar, track down the miscellaneous parts (side panel fan, op. panel with display, etc.) and convert this over to a 9595 with the following:

A 66MHz Type 4 Complex
64Mb (8x8Mb ECC) of Memory
A 1Gb IBM Hard Disk
Latest Revision Cached SCSI Adapter
An XGA-2 Card
A Madge MC32 Smart RingNode

Anyways, that's it for this update and sorry for the inconsistency of them, I'm typically juggling ten different projects at once so it can be hard to find time for everything. Have some extra pictures of the 8595 which actually had a bootable install of Win-95 (and some strange backup utility I've never heard of) on it!

IMG_20190705_193236.jpg

IMG_20190705_193310.jpg
 
Last edited:
Hah, if only... I don't get why SSKs are so damn expensive.....)

About 10 years ago I was in a bad spot and sold one for about $100; not it would cost me between double and triple that to get one back. To my surprise, Unicomp which does manufacture these things, has no interest in manufacturing them despite the premium they fetch.
 
"Now onto the first of my two 9585s. This system has acted very strangely since I've had it. It powers on the moment you insert a power cable and will not power back off until unplugged. It also does not POST whatsoever. I suspect it has multiple issues, though the power supply or switch is probably one of them."

Probably the switch. Replace that and then you can play with the no post.

http://ps-2.kev009.com/ohlandl/errors/Error_Index.html

Try to get the switch model, here's one I came across when dealing with inconsistent power on or power off...
http://ps-2.kev009.com/ohlandl/9595/Op_Panel.html#Replacement
 
"Cached SCSI Controller upgraded to 64Mb of Cache (Have)"

Giggles. You are referring to the Fast / Wide RAID Adapter/A [FWR]. The "Cached SCSI Controller" could also be applied to the 512K SCSI Adapter w/Cache

Mylex manufactured the Passplay for IBM
http://ps-2.kev009.com/ohlandl/RAID/RAID_Passplay.html

-BUT- they also made the Mylex branded DAC-960M
http://ps-2.kev009.com/ohlandl/RAID/RAID_DAC960M.html

DAC960M external port is a little odd. It doesn't use a sidecard like the Cheetah, but instead has a cable attached to the front port (Channel 1) that goes to an HPDB68 port that is mounted to the slot cover.
 
Think about it real hard

Think about it real hard

"Finally, the second 9585. As mentioned, this system was home to some mice at one point and as such, pretty much everything aside from the case is ruined and even then that's fairly corroded inside. If I can save the case, I'll probably just recap my bad 9595A planar, track down the miscellaneous parts (side panel fan, op. panel with display, etc.) and convert this over to a 9595"

You seem to imply that you have a 9585-K/N, a 83MHz POD capable system that can take FPM in sizes up to and including 32MB... Industry Standard SIMMs, IIRC. 256MB of screaming FPM is no laughing matter. I have done the same with a 9585-K/N that had a electrolytic vent spooge under the F/W controller chip [ :( ] , but K/N are quite a bit less common than 95As... Just about everything the 95A had. The 9585-K/N had LogicLock, Remote Power-On jumper, PTCs on the planar [actually useable with 400w PSU], Wake on Serial, VPD, Streaming components on system board ["planars" are properly used with complexes, "system boards" are properly used with CPUs on the PCB]

I suggest that you re-assess how bad the 9585-K/N is before you drop the hammer. At the very least put the planar aside for future work.

From my swap-out, IIRC, you will have to find one more black screw for the 95A planar. Though I'm pretty sure you can make-do until you can source one.

95A Bezel
95A Op Panel
95 Wall fan
95 wall fan power terminals

Planar Fast/Wide SCSI [Corvette] is nice, you don't burn a slot...
 
Isdn

Isdn

"A Madge MC32 Smart RingNode"

I will be happy to create a page just for you when you reveal how to port FDX Token Ring to Fast Ethernet. Token Ring was superior to 10Mbit Ethernet as far as transfer speed -AND- operation in a congested LAN. Switches enabled Ethernet to surpass Token Ring, finally. IBM gave up on Micro Channel before they came out with their PCI HSTR adapters :( ...

So, for you to achieve alpha-geek bragging rights, figure out how to shotgun a Dual LANStreamer on a switch and then uplink fast ethernet.

Better yet, shotgun TWO Dual LANstreamers on a switch and the uplink to Fast Ethernet... A single LANstreamer is notionally capable of 32Mbit/s in FDX by itself, so the hardware of four LS32s could reach 120Mbit/s, but then we have overhead, and lack of drivers, and whatever.
 
Last edited:
"Cached SCSI Controller upgraded to 64Mb of Cache (Have)"

Giggles. You are referring to the Fast / Wide RAID Adapter/A [FWR]. The "Cached SCSI Controller" could also be applied to the 512K SCSI Adapter w/Cache

Mylex manufactured the Passplay for IBM
http://ps-2.kev009.com/ohlandl/RAID/RAID_Passplay.html

-BUT- they also made the Mylex branded DAC-960M
http://ps-2.kev009.com/ohlandl/RAID/RAID_DAC960M.html

DAC960M external port is a little odd. It doesn't use a sidecard like the Cheetah, but instead has a cable attached to the front port (Channel 1) that goes to an HPDB68 port that is mounted to the slot cover.

It's the DAC960M. I realized that I had forgot to specify that in my configuration list for the 95A this morning and that it may be a little confusing to some. Oh well, can't edit it now. :lol:

"Finally, the second 9585. As mentioned, this system was home to some mice at one point and as such, pretty much everything aside from the case is ruined and even then that's fairly corroded inside. If I can save the case, I'll probably just recap my bad 9595A planar, track down the miscellaneous parts (side panel fan, op. panel with display, etc.) and convert this over to a 9595"

You seem to imply that you have a 9585-K/N, a 83MHz POD capable system that can take FPM in sizes up to and including 32MB... Industry Standard SIMMs, IIRC. 256MB of screaming FPM is no laughing matter. I have done the same with a 9585-K/N that had a electrolytic vent spooge under the F/W controller chip [ :( ] , but K/N are quite a bit less common than 95As... Just about everything the 95A had. The 9585-K/N had LogicLock, Remote Power-On jumper, PTCs on the planar [actually useable with 400w PSU], Wake on Serial, VPD, Streaming components on system board ["planars" are properly used with complexes, "system boards" are properly used with CPUs on the PCB]

I suggest that you re-assess how bad the 9585-K/N is before you drop the hammer. At the very least put the planar aside for future work.

From my swap-out, IIRC, you will have to find one more black screw for the 95A planar. Though I'm pretty sure you can make-do until you can source one.

95A Bezel
95A Op Panel
95 Wall fan
95 wall fan power terminals

Planar Fast/Wide SCSI [Corvette] is nice, you don't burn a slot...

From what I recall seeing when opening this particular 9585 up; There was a lot of "rodent grime" on the system board (this machine was on it's side when it was inhabited) and the expansion slot contacts as well as some board traces looked like they were corroded. Both of the 9585s I received are the same type though, so I do have one 9585 to play around with.

If the system board isn't as bad as I think it is then I'll definitely try to clean it and set it aside for future work, but as it currently stands, I would like to have the "full range" of Model 95s in terms of different complexes and such, with the T4 P66 being the only other one I "need" (though a T0 386 8595 would be nice too...). I'll definitely set all of the 9585 specific parts that I can save aside though in case I ever want to restore this back to "original".
 
Last edited:
Quick update on the 9585s:

So after some probing with a multi-meter I have found that the power supply for the non-rodent infested 85 (which is a K/N 85 by the way!) is in fact faulty. Guess it's time to crack it open and see just what went wrong with it and if it's repairable.

As for the rodent-infested 85, it's a K/N model as well and looking over it again I highly doubt the planar, let alone the PSU, is salvageable. The chassis / case is a definitive maybe, however.
 
Seeing as how I'm waiting on parts to finish my 95A and the non-infested 85, I decided to at least start stripping down the rodent infested one and examining the damage. All I can really say right now is that this is just disheartening.

IMG_20190710_184217.jpg

IMG_20190710_184034.jpg

IMG_20190710_184200.jpg

At least I was wrong about the system board. It's probably going to be the only salvageable piece electronics wise, so at least I'll have a spare one for my other 85 (they're both 9585-0NGs by the by). Hopefully the diskette drive is salvageable as well and that the corrosion is just on the shell, but I'm not too sure about that. Other than that, I hope to be able to save the chassis, as pretty much every other part can be sourced to build a new system.

Anyone have any tips for removing rodent grime and corrosion from the chassis?
 
Quick update on the 9585s:

So after some probing with a multi-meter I have found that the power supply for the non-rodent infested 85 (which is a K/N 85 by the way!) is in fact faulty.

Alien probing?

Was the PSU plugged into the planar, or did you drop it and then test?

If there is a problem with crud in a SIMM socket or mca slot, or with IC pins, you might never get power good.

Might want to unscrew the ground strap at the front, then pull the psu foreward and off the pivot pins. Turn the psu upside down and listen for bits. Open the psu, visually inspect for little spikey bits. Clean up as needed, then plug psu in, short the correct planar power plug pins and power it up, then check the 4 pin drive power taps for voltage.

Fan runs, bad voltages, oww. Fan runs, good volts, yay! Fan does not run, good volts, probably bad fan.
 
DASD support structure is plastic, should clean right up.

Metal drive structure is sorta phosphated, if you use evapo-rust, it will take it down to shiney metal. Might clean off the rust a bit, then use a chemical to convert the remaining rust to inert coating.

Damn foam is shot, wish I knew of a good replacement.
 
There are places that sell foam rubber in various types and sizes. Shouldn't be too hard to make up something suitable.
 
Alien probing?

Was the PSU plugged into the planar, or did you drop it and then test?

If there is a problem with crud in a SIMM socket or mca slot, or with IC pins, you might never get power good.

Might want to unscrew the ground strap at the front, then pull the psu foreward and off the pivot pins. Turn the psu upside down and listen for bits. Open the psu, visually inspect for little spikey bits. Clean up as needed, then plug psu in, short the correct planar power plug pins and power it up, then check the 4 pin drive power taps for voltage.

Fan runs, bad voltages, oww. Fan runs, good volts, yay! Fan does not run, good volts, probably bad fan.

All of the SIMM sockets and MCA slots were clean. I removed the supply, powered it on (fan runs) and probed the main supply pins as well as the drive connectors. There was effectively no power to the 5v (0.0 something) and the 12v was "jumping" around from around ~2v to ~6v.

DASD support structure is plastic, should clean right up.

Metal drive structure is sorta phosphated, if you use evapo-rust, it will take it down to shiney metal. Might clean off the rust a bit, then use a chemical to convert the remaining rust to inert coating.

Damn foam is shot, wish I knew of a good replacement.

Thanks for the tip! I did spray everything down with bleach and hosed it off with hot water, so at least the grime is gone and it should be safe to touch with bare hands now :lol:

Is there such a thing as good foam in these systems? It all seems to deteriorate into "black goo".
 
Back
Top