PDA

View Full Version : Diskette drive faulty in IBM PS/2 Model 30-286



ajcc
July 12th, 2014, 09:36 AM
Hi guys!

I got myself an IBM PS/2 Model 30 with a 10 MHz 286er in it. I've been having quite a fight with it recently and now I'm completely stuck.

The ALPS diskette drive that came with the computer had its capacitors leak and ruin the circuit, so it doesn't work. I thought it would just have a standard 3.5" diskette drive inside, but there was no power connector?

Trying to find an answer I found these two threads: Reborn IBM PS2 30 286 (http://www.vintage-computer.com/vcforum/showthread.php?36883-Reborn-IBM-PS2-30-286) and IBM PS/2 Model 30 (8530-002) (http://www.vintage-computer.com/vcforum/showthread.php?25201-IBM-PS-2-Model-30-%288530-002%29/page3). Sadly I was discouraged by the comments and decided to put my voltmeter to the board and see if they were sound. Pin 4 isn't Ground so I wasn't sure... if it had been you could have just put the four wires into the power connector neatly, not tearing out single strands of wire.

Anyway, there doesn't seem to be a Model 30-286 service manual out there on the Internet. But I found out the the P70 had the same ALPS diskette drive in it (searching the IBM part number: 72X6068 ), so I finally found the correct pin-out of the diskette drive! (Note that there's a 40-pin connector on the motherboard, which seem to have extra pins for power, not sure).

I found the PS/2 Model P70 Technical Reference (http://bitsavers.trailing-edge.com/pdf/ibm/pc/ps2/PS2_Model_P70_Technical_Reference_Apr89.pdf) on the bitsavers ftp (http://bitsavers.trailing-edge.com/pdf/ibm/pc/ps2/).

On pages 41:
19440
And page 42:
19441

There is the pin out.

Then after fighting and destroying a couple of IDC connectors, I had an adapter. It didn't work! So I decided to test every diskette drive I had, still didn't work... I checked the wiring, checked the voltages, everything was fine, the 5V line didn't drop in voltage during read... maybe all my diskette drives are broken?

Sad and ready to just give up, I realized I didn't try a 720kB diskette. That did work! Trying to run format A: crashed the computer though, might be a broken DOS, I don't know. I could even write a file onto the diskette.

So, now I wonder. How do I get it to see my 1.44MB diskette drive?, the /Density Low or High doesn't seem to be used on any of my drives. It's always 5V on the PS/2 as well, so maybe it isn't used? Now the problem is that the Advanced Diagnostics diskette is a 1.44MB image, and I can't boot 1.44MB diskettes.

I'm stuck :(

ajcc
July 12th, 2014, 09:40 AM
19442

ajcc
July 13th, 2014, 02:25 AM
Been working a bit more on this strange problem.

I've copied the content of the Startup Diskette for the Model 30-286 onto the hard drive by splitting it up onto 720kB diskettes. And I managed to run SC.EXE from the directory, it seems to find everything and configure the system correctly. The System Configuration was set to Not Installed for Both A: and B:, it did auto detect the 1.44MB drive.

This wasn't it, or not all of it. No progress, it didn't make a difference. Neither MS-DOS nor the BIOS can read of 1.44MB diskettes, I can boot of 720kB just fine but 1.44MB diskettes aren't read at all. I did notice when I probed the voltages that the /HEAD select pin was always 5V. If I'm not wrong, during the read of a 1.44MB diskette, it should alternate between head 1 and 2 (or 0 and 1, don't know the numbering).

Maybe I should build a logic probe?, even if it's a 10Mohm resistor in the multimeter... :/

So I'm still stuck :(,
still working on it.

SpidersWeb
July 13th, 2014, 12:17 PM
If I remember correctly the PS/2 range use "Drive ID" pins to identify drives. It's possible that it just thinks you've plugged in a 720KB.
Both 720KB and 1.44Mb disks use both heads - the difference is data rate (as far as the computer cares that is). Head 0 will always read first.

Edit: nevermind, realised you just said it auto detected a 1.44

ajcc
July 13th, 2014, 12:34 PM
If I remember correctly the PS/2 range use "Drive ID" pins to identify drives. It's possible that it just thinks you've plugged in a 720KB.
Both 720KB and 1.44Mb disks use both heads - the difference is data rate (as far as the computer cares that is). Head 0 will always read first.

Edit: nevermind, realised you just said it auto detected a 1.44

I'll still check it over, could just as well be the System Config defaulting to the standard 1.44MB drive. There was a pin sitting at around 0.3V so I'll redo the adapter anyway. This was one of the reserved pins, on my modern PC diskette drives these are Open (not connected), so this could also be a floating gate. If only it was easier to work with those IDC connectors...

ajcc
July 26th, 2014, 06:36 AM
Hi guys!, it's been a little while. Been busy at work as usual.

I'm still stuck. But I have been poking at the Model 30-286 again. There's something odd going on with the way the floppy drive is wired up. So I've been focusing on the reserved pins.

On the ALPS (the original drive) PIN 4 is connected to the gate or base of a transistor, can't tell which type it is because one of the legs is badly corroded. It shows a diode property between Ground and the collector or emitter. The other side of the transistor is hooked up to the base of another transistor. Which sits between the IC and +5V. So it must be getting a signal from the controller on the IBMs floppy drive controller. This pin is NOT connected on modern PC drives. However it sits at 5V at all times.

The other reserved pins: 10, 14 and 34 are not connected on the ALPS.

So the reserved pins might have been a red herring :(

Now, I did find something strange. /High Density Select, PIN 2, is alternating between 0 and 5V, resting at 0V. I can't really tell with my multimeter's slow update, but it feels like it's alternating at a speed twice that of the head stepping. As if it was actually the Head Select signal! However... if it was correct, it would be resting at 5V. I have tried wiring it up to the /Head Select 1 pin on the PC drive, but that makes it unable to read any diskette.

Edit: I was wrong about it always sitting at 0V. Until there's a DD diskette in the drive, it sits at 0V and when trying to read a HD diskette it will alternate between 0 and 5V, resting at 0V when failing. When reading a DD-diskette, it will stay at 5V and go back and stay at 0V when no diskette is in the drive. So the logic levels might be correct after all.

So the reason I can read 720kB diskettes without a problem is that /Head Select 1 is always tied to 5V, never selecting the second head. When I try reading 1.44MB diskettes DOS complains about not being able to read the sectors, which alternates between the tracks on both sides.

Still confused and stuck though :/ But maybe I have found something to poke at.

Edit: Maybe I'm best of trying to find some sort of service manual or checking the signals on a known working PC?

Chuck(G)
July 26th, 2014, 08:21 AM
A 720K floppy is traditionally a lower data-rate (fewer sectors per track) double-sided beast and not a single-sided one. If you were to format half of a 1.44M floppy, the other half would show up as 720K of bad sectors. As far as I am aware, there never has been a standard single-sided 720K high-density format for MS-DOS.

ajcc
July 26th, 2014, 08:37 AM
A 720K floppy is traditionally a lower data-rate (fewer sectors per track) double-sided beast and not a single-sided one. If you were to format half of a 1.44M floppy, the other half would show up as 720K of bad sectors. As far as I am aware, there never has been a standard single-sided 720K high-density format for MS-DOS.

I see... then I really don't get what the signals are doing. If it's all true what you guys are saying. There should be no reason for if being able to read 720kB diskettes and not 1.44MB diskettes. Clearly the controller is alternating between DD and HD in an attempt to read the diskette, so it at least tries to read the 1.44MB diskette.

ajcc
July 26th, 2014, 09:02 AM
The PS/1 Type 2011 seem to use the same 72X6068 Alps drive.

The Reserved PIN 4 is Drive Type ID 1.
19602

Chuck(G)
July 26th, 2014, 09:09 AM
How are you monitoring the "Head 1" pin? A DMM will likely not be fast enough--a logic probe or scope will catch line wiggles. The normal condition for inactive lines on a floppy interface is pulled high (+5V), so there may be nothing wrong. I'll have to go back to some of my old PS/2 documents, but I don't think that the Model 30-286 cares about the density aperture--I seem to remember that you could format 720K disks to 1.44M--and vice-versa. So perhaps your adaptation isn't relaying the proper drive information to the system.

Perhaps someone with a machine can refresh my memory.

There were variations. Here's the diskette drive reference that explains the signals. (http://ps-2.kev009.com/ohlandl/books/S42G219400.pdf) Note that the drive can tell the system what media it can support--and in some cases, what media is inserted.

PeterNC
July 26th, 2014, 10:32 AM
IBM PS/2 Model 55 FDDs are interchangeable with the Model 30 80286. Those are more common and usually pretty cheap. Good luck. :)

ajcc
July 26th, 2014, 10:49 AM
How are you monitoring the "Head 1" pin? A DMM will likely not be fast enough--a logic probe or scope will catch line wiggles. The normal condition for inactive lines on a floppy interface is pulled high (+5V), so there may be nothing wrong. I'll have to go back to some of my old PS/2 documents, but I don't think that the Model 30-286 cares about the density aperture--I seem to remember that you could format 720K disks to 1.44M--and vice-versa. So perhaps your adaptation isn't relaying the proper drive information to the system.

Perhaps someone with a machine can refresh my memory.

There were variations. Here's the diskette drive reference that explains the signals. (http://ps-2.kev009.com/ohlandl/books/S42G219400.pdf) Note that the drive can tell the system what media it can support--and in some cases, what media is inserted.

I tried pulling Drive Type 1 high and low. Makes no difference. The controller seems to ignore this signal :(
Now I'm just out of luck. All odd numbers except 3 are ground so none of the extended signals are used.

My adapter is only taking the +5V, +12V and two ground wires and putting them into the molex. All other lines are left untouched.

ajcc
July 26th, 2014, 11:02 AM
IBM PS/2 Model 55 FDDs are interchangeable with the Model 30 80286. Those are more common and usually pretty cheap. Good luck. :)

Might be the best option.

Chuck(G)
July 26th, 2014, 11:22 AM
Well, there shouldn't be a problem with a "dumb" drive. What are you using for your replacement unit?

If you want a real replacement, call Cindy at Electronicsplus.com. She says she's got 'em.

ajcc
July 26th, 2014, 09:07 PM
Well, there shouldn't be a problem with a "dumb" drive. What are you using for your replacement unit?

If you want a real replacement, call Cindy at Electronicsplus.com. She says she's got 'em.

Hm, I wonder if I have a dumb drive. I have some you can setup with solder jumpers, which use switches to detect DD or HD.

ajcc
November 26th, 2015, 03:00 PM
I've spent the day poking around at the PS/2, tinned the ends of the two ribbon cables connecting the Floppy drive to the motherboard and hooked it into a bread board. If there's a signal somewhere, it should be available there!

This didn't get me any further than the last time. I can hear the drive switching between the DD/HD, but the Drive Type signal from the controller is High, saying it is expecting a 720kB drive (if I understand this correctly) Pin 4. On the ALPS drive there's a Diode Pin4 -I<- GND, the Pin4 also goes into a transistor, which sits between a pin on the main IC and GND. I suspect this is telling the controller it is operating as a 720kB or 1.44MB drive.

28010 28011

The DS1287 was completely flat and it wouldn't keep time settings even when there was power. I had a DS12887 in my junk box so I replaced it with this, which amazingly still got a good battery!, with this I could set the time and date! :) However, Diagnostics would still crash the computer completely. So I had a look at the RAM, there where two regular parity SIMM's, which can't be used with the PS/2, and then there was these two weird looking sticks. With some aid from IRC (sorry, I don't know your name on vcforums, nashi) the regular sticks where found to be the culprits straight away.

28012
Two halfs of a Megabyte Evil Proprietary IBM PS/2 FPM DRAM Parity 30-pin SIMM memory.
Edit: After some sleep I realize it's two 512kB 4-bit DRAM chips + 1-Mbit parity, so it really is 1024kB per stick. That isn't the issue, unless they are simply just broken. However I still get 164, memory size error during POST.

28013
Signing off for today, cheers!