• Please review our updated Terms and Rules here

IBM RT PC experts?

Hm, I must have been mistaken about my memory re: it not being generally possible to use an IDE drive. I couldn't find a reference to the details of those reports (I could've sworn it was a FAQ item, but no) when I was originally writing that, and so going off (apparently faulty) memory. Perhaps it was something about an Adaptec ISA SCSI board.
 
What model 1g drive worked?
I tested with a Seagate ST32140A which worked.
Can you tell what kind of ISA controller it thinks the IDE drive is?
The RT PC Diagnostic prograg detects the controller as an ”IBM Fixed-Disk and Diskette Drive Adapter 1”

There have been lots of CF cards made so who knows how closely the ones you tried match what
the RT expects a supported controller to respond to.

Depending on how far you want to dig into it, you may have to trace the commands going to the
working IDE drive and see how they differ from what happens with the CF

Yes. It could be interesting. We’ll see how much time I find study this. I like to do some more research info IDE and CF in general.
Also, if you have one an ISA card that is only an IDE adapter would be interesting to try.
I tested with two different cards. Both of them Multi IO. One had the UMC UM82C862F chip the other had the GoldStar Prime 2 chip. At this point the only one I tested with the 1 gig Seagate was the UMC chip card. With the 128 Mbyte CF card I tested both and there were no difference in the behaviour. My gut feeling is that the Goldstar chip will work fine with the Seagate. But not tested, though.

I don’t have an IDE only board. Isn’t those that just have a bus buffer chip and not much more?
 
Last edited:
Did some further testing of various hard drives.

I successfully formatted, installed VRM, installed AIX and booted the following drives, ST3144A, ST3283A, ST3491A, ST31240A. On the ST3283A the first format attempt failed but the second attempt was ok. Then I also tested a ST3850A which formatted OK, installed VRM on the second attempt and almost installed the OS. At two occasions I had LED blinking C6 02 with this drive. Perhaps the drive is flakey?

I also tested a couple of more modern drives. a 40 Gigabyte Seagate ST340016A and an IBM DJNA-351520. Both hang the VRM installer boot with error code 35 on the LED.

So all in all. Old IDE drives work. Newer do not. CF card do not either. I have no idea why.

I think I figured out why I failed to run the ISA IDE controller as secondary controller. I disabled the floppy but didn't set the floppy to the secondary address. I would guess that the address decoding logic is not taking this into account really. The hard disk is using 3F6/376 port and it was most likely clashing. Will make another test tomorrow with the ESDI controller as the primary controller and the ISA Multi-IO as a secondary IDE only. Floppy and all other peripherals disabled.
 
Just to close it off for me, I have AIX running now after getting a second H310 drive. It initially didn't work but I swapped the board between the two drives and one worked in the end. Formatted, installed AIX and X-Windows and all is good.

IBM_PC_RT_running_AIX (Medium).jpg
 
Nice!

Did you install all the updates? Did you manage to install NFS and the Distributed Function packages?

Have you seen the files on fsck.technology site? A lot of RT AIX stuff there.

I am using 1 gig out of a Seagate 2 gig IDE drive. Seems to work very well.

I don’t have a 6157 so I have started working on a emulator so that I can install the updates that are bigger than a floppy in size.
 
Nice!

Did you install all the updates? Did you manage to install NFS and the Distributed Function packages?
I installed AIX 2.2.1 with updates from Bitsavers here: http://bitsavers.trailing-edge.com/bits/IBM/RT/aix/AIX_RT_2.2.1_w_Upds.zip

IBM PC RT X-Windows was on archive.org here: https://archive.org/details/ibm-pc-rt-x-windows-licensed-program_202101

I only have a token ring card right now so other than installing the TCP/IP stuff, I have not explored NFS or Distributed functions. I need to explore options for ethernet connectivity. How is your set up for this?

I don't have a 6157 either so I'm limited to floppies but I am assuming TCP/IP networking running would solve that issue?

I have the H310 so I think I have some room still.
 
I also installed the same zip package as you did. But NFS came from fsck.technology site. Unfortunately it refused to install. It claimed that the TCP/IP package or AIX base program had the wrong level. I do have TCP/IP installed though.

I don’t have TCP/IP configured yet (mostly because I have no transceiver around at this moment) but I cannot imagine that I would be unable to install NFS because of the TCP/IP is not configured. But I might be wrong. A lot of things in AIX is a bit strange…

The Distributed Function package installed halway as I recall it but failed when trying to build the kernel. This is the distrib.bz2 package here.

The fsck.technology site has a lot of updates which looks interesting. But many of them are quite big files. It seems that the only way is to install them from a tape cartridge since the updatep program only takes a device-file as an argument. Not a regular file as far as I understand. Therefore my idea of creating a qic-02 emulator. So my impression is that getting TCP/IP wouldn’t help. But I am happy to be proven wrong here. If updatep take a regular file as an argument it would be great! Anyone know for sure?

Regarding uppdates. I have been quite successful applying the package of 12 floppies of corrections in this package: http://bitsavers.trailing-edge.com/bits/IBM/RT/aix/COMPID_56010606103.zip

It took ages to apply though.
 
I gave away my RT collection years ago, but still have a 6157 tape drive in storage. If anyone is looking for one, please PM. I have no idea whether it's functional, as I was running BSD 4.4 and used an Adaptec 1542 + SCSI tape for backup.
 
Spent some time trying to understand the IBM RT keyboard protocol. My old original RT rubber dome keyboard had became sticky and was impossible to type on.

I reverse engineered the original firmware and translated it into C code which is running on a STM32 blue pill. The new firmware is able to configure a standard model M into scan code set 3 and at the same time respond correctly to the RT host.

The result is that I can use my standard model M on the RT!

https://github.com/MattisLind/RTKBD
 
4o62GKl.jpg


After some browsing on bitsavers I found the Gold CD. It was indeed gold! http://bitsavers.org/bits/IBM/RT/gold-cd.zip

It contains a much more curated set of AIX images than found elsewhere. There is even a X11 package and two updates which brings it to X11R3 in there. The X package that is found elsewhere is X10R4. On the CD there is also a Motif Window Manager package. It installs fine, but I cannot figure out how to run it. Normally I would edit the .xinitrc file so that it starts the window manager, but it seems that the IBM packaged xinit doesn't care much about the .xinitrc file. It always starts the aixwm and one xterm. Nothing else. The standard trick which I practised 35 years ago to kill the current wm and start a new one doesn't work.

On the CD there are precompiled sets of X11R4 and X11R5 for AIX, but I cannot get the server to run. Many of the clients run under the native IBM X11-server though. At least those that doesn't make use of any fonts.

I also downloaded the X11R3 sources and tried to compile the xlogo client which actually did compile after som fiddling. There are some small differences between the X11R3 dist and the IBM X11R3 package. But it compiles and run.
 
Picking up this old thread:

I've acquired two IBM 6150 RT PC systems. Both of them are in working condition, but one of the two is missing the drives and controller. The other one has two E70 drives and boots AIX. 140 MB are barely enough to install the operating system and one application, and I'd love to run UniPress Emacs and maybe some stuff from the Gold CD. Trying AOS or BSD would also be kind of cool.

To follow the lead of @MattisLind, I have collected two IDE controllers, one of which is based on the UM82C862F:

1749233103004.png

I've also collected a couple of IDE drives:

1749233148474.png

Unfortunately, none of the drives are detected by VRM (02.01) or the diagnostics (2.2.3). Did anyone else got their 6150 to run with IDE drives? Any hints would be greatly appreciated.

Thanks,
Hans
 
This is the card I am using. Seems to be very similar to your board.

IMG_4591.jpg

Did you configure the board correctly? Have you verified that it works in a standard PC?

I had luck with the ST3491A so try with that one first.

I found the jumper settings for my board:


Code:
BioStar tech support was kind enough to provide the jumper settings,
which are (in case anyone else wants to know):

The IODE-3290U jumper settings are:
JP1 : HDD LED
JP2 : FDD port Enable/Disable select
1-2, closed to enable FDD port
2-3, closed to disable FDD poet
JP3 : Print port Enable/Disable select
1-2, closed to enable print port
2-3, closed to disable print port
JP4 : COM1 (first serial port) Enable/Disable select
1-2, closed to enable COM1 port
2-3, closed to disable COM1 port
JP5 : COM1 (second serial port) Enable/Disable select
1-2, closed to enable COM2 port
2-3, closed to disable COM2 port
JP6 : Game port Enable/Disable select
1-2, closed to enable game port
2-3, closed to disable game port
JP7 : HDD (IDE HDD port) Enable/Disable select
1-2, closed to enable HDD port
2-3, closed to disable HDD port
I/O Port Address Select
JP8 : FDD port address select
1-2, closed to select address 3F1-3F7
2-3, closed to select address 371-377
JP9 : Print port address select
1-2, closed to select address 378-37F
2-3, closed to select address 278-27F
JP10 : COM1 port address select
1-2, closed to select address 3F8-3FF(COM1)
2-3, closed to select address 3E8-3EF(COM3)
JP11 : COM2 port address select
1-2, closed to select address 2F8-2FF(COM2)
2-3, closed to select address 2E8-2EF(COM4)
JP12 : HDD port address select
1-2, closed to select address 1F0-1F7
2-3, closed to select address 170-177
Interrupt Select
JP15 : Interrupt select for print port
1-2, closed to select "IRQ7"
2-3, closed to select "IRQ5"
JP16 : Interrupt select for COM1 port
1-2, closed to select "IRQ4"
2-3, closed to select "IRQ3"
JP17 : Interrupt select for COM2 port
1-2, closed to select "IRQ3"
2-3, closed to select "IRQ4"
 
Last edited:
@MattisLind Thank you for encouraging me to test the Controller/Disk combinations in a PC - I did, and the one with the UM82C862F chip is toast. The other one does not work, but I'll now go hunt for more controllers to test.
 
I've been working on emulation of the RT PC in MAME and struggling with a similar issue. I now believe that the RT firmware and VRM disks won't "detect" a drive unless it has a valid configuration block written on the second sector of the disk. I believe this config block will be created by the low level format program available from the VRM maintenance facility (which allows you to manually input drive parameters, and will format a disk that is not "detected"). Hope this helps.
 
I have a good assortment of floppies and now also collected a couple different IDE controllers that I'm going to try. Is the VRM maintenance facility part of the standard VRM install disks or is that something else?
 
I'm puzzled why a non bus mastering IDE adapter would make any difference at all
assuming all of the other ports on the card are disabled. It is just bus buffers.
I suspect the magic sauce is the IDE behaviour of the drive.
 
Back
Top