• Please review our updated Terms and Rules here

DB37 Pinout for External IBM Floppy

...Since I've tested this drive in my PS/2, and it did not work, it could be either the cable or my MCA floppy controller. When I booted my PS/2 last, I believe it did recognize the controller because when it started up it said it was mounting the floppy. Not 100% sure on this...

There is a DOS driver required for your microchannel adapter. Do you have the earlier version which has a switchblock, or the adapter without (both have the same 15F7996 FRU)? I haven't determined what the switches did (a physical switch on a microchannel adapter is always intriguing, because the architecture defines that configurations should be done in software).

I did a DOS utility years ago that checked for presence of the adapter (the purpose was actually to determine which main FDC chip was present for the different PS/2 systems), but need to get back into that research...
 
In my experience good quality cables are usually the least likely problem, which is why I asked about the adapter (and any related drivers etc. of course).

But it's been interesting; BTW miniature 4PDT switches are readily available for that interface switch (see KISS principle ;-) )
 
But it's been interesting; BTW miniature 4PDT switches are readily available for that interface switch (see KISS principle ;-) )

Consider that alternate wires on the floppy interface are grounded--you need only a DPDT swtich (common as dirt) to change the signals.
 
...But it's been interesting; BTW miniature 4PDT switches are readily available for that interface switch (see KISS principle ;-) )

Even through they aren't "double-throw", I've wondered what the four line switchblock was on the initial 15F7996 microchannel adapter (my notes said that factory settings seem to be all switches "ON"). IBM had the drive select signals mapped so that the internal drive (if present) is primary, making the single external drive secondary.

That adapter version was covered in SOIC chips compared to the later released 15F7996 (without the 4 position switchblock). Having the same IBM FRU is supposed to mean the same functionality, but I wonder if the switches changed the drive select mappings between the internal and external connections. The I/O port of the FDC would have been changed by other means (in the ADF), and I've not heard of any odd adaptations that IBM would have done to connect a PS/2 to an older floppy unit.
 
Have you tried a different power supply with the drive? The "spins slowly" behavior has me wondering...

No I haven't, but the drive did work when connected via the internal cable while using the external power supply, so it's definitely working.
 
There is a DOS driver required for your microchannel adapter. Do you have the earlier version which has a switchblock, or the adapter without (both have the same 15F7996 FRU)? I haven't determined what the switches did (a physical switch on a microchannel adapter is always intriguing, because the architecture defines that configurations should be done in software).

I did a DOS utility years ago that checked for presence of the adapter (the purpose was actually to determine which main FDC chip was present for the different PS/2 systems), but need to get back into that research...

I believe I have the later version of the MCA adapter. The thing is, this floppy was previously working on this PS/2 and ceased to function after the move.
 
In my experience good quality cables are usually the least likely problem, which is why I asked about the adapter (and any related drivers etc. of course).

But it's been interesting; BTW miniature 4PDT switches are readily available for that interface switch (see KISS principle ;-) )

Cool. Again, something else I've never heard of!

I just looked up the "4PDT switch" and it looks like it would take care of it in one fell swoop. And it sounds like a DPDT would work, too. It's nowhere near as "fun", though! :)OK. Fine, I guess I can't really build my fancy circuit switch with a solution as simple as this out there.

So now I know how I'm going to build the switch, but the cable mod didn't work so there's no point building the switch until I figure that one out.

Does anyone have any thoughts about the "drive constantly on" phenomenon after ChuckG's mod? Again, the power supply is working. Also, the drive didn't respond to any commands or boot - it just spins if turned on and connected to the IBM 5150's floppy adapter via external cable. The PC doesn't even have to be on! If the drive is on and connected to the DC37 port, it spins slow and constantly.

Doesn't all this make one think it's a cable problem?
 
Last edited:
OK, I got my 4869 working on my IBM 5150!

First of all, the "slow spin" behavior is because the card-edge connector is installed UPSIDE-DOWN. IBM in their infinite wisdom did not use plugs with the "non-reversible tab". It's my fault, though. I should have checked this first.

Also, my config.sys line was wrong. Here's what it should have been:

device=driver.sys /d:3 /f:0

I had previously used "/d:2". I was about to throw the towel in when I just tried "/d:3" for fun, and it ended up working!

So ChuckG's 4869 cable modification worked great.

Thanks everyone!
 
I recently purchased an IBM 4869 drive where are some idiot cut the cable off. I own several of these drives, so I simply used a DMM and reverse engineer, the cable pin out. Here it is in the photo pin to pin with the connector pin configuration at the bottom. You’ll note DSUB37 pins 1, 2, and 20 are no connects.
 

Attachments

  • 70517014877__30D0045C-6C43-46DF-8BAA-5C3F839DCF91.jpeg
    70517014877__30D0045C-6C43-46DF-8BAA-5C3F839DCF91.jpeg
    1.7 MB · Views: 19
Had you wondered, I could have furnished the pinout. I still have a few NOS 4869 boxes. As a matter of fact, I've got a blog entry here on modifying the cable.

Oh well...
 
Had you wondered, I could have furnished the pinout. I still have a few NOS 4869 boxes. As a matter of fact, I've got a blog entry here on modifying the cable.

Oh well...
I just thought it would be nice to have a picture of the pin out on the thread that says DB37 pin out. That way people find it without having to explicitly ask or search for it. 👍🏻
 
I recently purchased an IBM 4869 drive where are some idiot cut the cable off. I own several of these drives, so I simply used a DMM and reverse engineer, the cable pin out. Here it is in the photo pin to pin with the connector pin configuration at the bottom. You’ll note DSUB37 pins 1, 2, and 20 are no connects.
I just repaired my cable using this pinout and I noticed I missed EDGE CONN. (34P) 34 to DSUB (37P) 19. Other than that, everything is good. I just tested the drive and it works great.
 
Basically, it's the old formula of "crimp a 34 pin ribbon cable to an edge connector, then take the other end and position it in the DC37 IDC such that the first three pins (1,20, 2) are unoccupied.

Got to be careful with that, however. I have one system that has its external floppy connector as a DC37, but the unused pins are 18, 37 and 19. It predates the IBM PC. I think that the Xerox 820 external floppy drives used the latter, but I don't recall offhand.
 
Note that the IBM 4865 uses the unused pins to power the drive, the pinout is the same otherwise
 
Note that the IBM 4865 uses the unused pins to power the drive, the pinout is the same otherwise
That's good to know, because I actually have one of these 4865 drives. I bought it months ago, haven't got around to building a power cable for it yet. As it did not come with the power splitter cable.
 
I found the pinout on vogons.org forum stated as follows: "+12V to pin 1, +5V to pin 4, ground to 2&3"


This would be pins 1 - 4 on the DB37, not the 34 pin edge connector, just to clarify. He states there are two models of 4865, one with the power pig tail built in for the PC/XT, and one without for the AT. The one I have is without the pigtail. I will wire this up and test the drive. I will post my results once I have them.
 
Finally got time to build my cable and confirm pinout. I bought a DB37 straight through male/female cable to make the modifications on to preserve the original 4865 cable.. I hacked into it and located wires 1 - 4. Had to make an educated guess on which of these wires were 1 - 4. They used the same scheme as resistor color coding. Brown being 1 and red being 2. The wires were twisted pair and so the brown pair was wires 1/2 and the red pair were wires 3/4. I spliced in a 4 pin Molex connector to allow connecting +5/12VDC supply the wires with the previously mentioned pinout.

After getting this cable made I was able to power/connect the drive to my IBM XT. I already had this PC configured to use a 360K 4869 drive with "DEVICE=DRIVE.SYS /D:3 /F:0" in my config.sys. I simply modified to "/F:2" for a 720K drive. This did not work though so I next tried changing the drive to /D:2 where it then worked. This being said, it seems you have to have "DEVICE=DRIVER.SYS /D:2 /F:2" or it won't work. Conclusion is the (1 = +12VDC, 2/3 = GND, 4 = +5VDC) pinout on the DB37 for a 4865 drive is correct. Attached is a picture of the power adapter cable I made.

Here are links to the parts I used.

4-pin Molex cable, this one is cheaply made but works. You get 5 for less than $10.

100 - 240VAC to +5/12VDC Power Adapter, Very nice in my opinion for only $15.

1 foot DB37 shielded straight through male/female cable
Wire Numbering -> Solid Brown = Pin 1, Striped Brown = Pin 2, Solid Red = Pin 3, Striped Red = Pin 4
 

Attachments

  • 4865_PowerCable.jpg
    4865_PowerCable.jpg
    2.1 MB · Views: 14
Back
Top