KDA50 and Some SDI Disks

A few years ago I acquired a KDA50, which consists of 2 modules, the M7164 and the M7165, an RA72 disk, two RA70 disks, an RA60 disk drive, an operator control panel (OCP) and some big cables, with one I/O bulkhead for installing disks and controller in separate cabinets.

The controller and the disks use SDI, the Standard Disk Interconnect, which was part of the Digital Storage Architecture, and allows controllers to connect to large system disks (large for the time anyway). SDI allowed disks to be connected via long cables, so the disks could be housed in separate cabinets. SDI encoded signals serially, so it seems to be an early pre-cursor to SATA. SDI appears to date from the early 1980s.

I decided fairly quickly that I was unlikely ever to have the time and resources to get the RA60 going so I passed that on to a friend. The other stuff sat in my pile for a long time and I kept promising myself I would look at them one day. Well one day finally arrived a couple of days ago and I decided to get everything out and see what works.

To begin with I brought the disks in from the cooler storage area and allowed them to warm up to ambient temperature overnight before doing anything with them at all. I first decided I would see if the disks spin up when powered up outside the machine, by connecting them to a regular PC power supply. The disks did not spin but I saw some LED activity. I was not too disheartened because I thought that perhaps they might need to be connected to the KDA50 controller before they do anything, and that did indeed turn out to be the case. It turns out that they really need to be attached to the KDA50 controller before they will do anything.

So I needed to get the KDA50 installed in a machine. I wasn’t sure which machine to use. I decided to try the VAX 4000-500, mainly because it is handy. This turned out to be a poor choice. I think when I first got the KDA50 I may have installed it in this machine and it appeared to work, it displayed a cycling pattern on the LEDs of both modules, which indicates that it is not talking to the host yet. This turns out to be normal until you actually boot an operating system.

However, before installing the KDA50 again I decided to check the KDA50 User Guide. I found that both boards in the set have two jumpers, W2 and W3 that are supposed to be removed in a Q22/CD slot. I checked the enclosure manual for the VAX 4000-500, it is a BA440, and all the slots are Q22/CD. Having not checked for this before, I asked on the VCF forums for advice. The jumpers are present by default, presumably because the KDA50 is for an older generation of VAXen with a serpentine Q-Bus. I think I may have been lucky the first time, possibly because I had a TK70 with a dummy card in the CD slots before the KDA50. Unfortunately the jumpers are soldered in and not the type with a header, so I had to desolder them. Rather than remove them completely though I desoldered only one of the leads, lifted the link while bending it out of the way and finally putting a little bit of insulating tape over the holes left behind.

With this done I installed the boards in my VAX 4000-500, I got the cycling LEDs and the following from the firmware console:

>>>sh qbus
Scan of Qbus I/O Space
-20000130 (760460) = 0080
-20000132 (760462) = F081
-20000134 (760464) = DD18
-20000136 (760466) = 0000
-20000138 (760470) = 0000
-2000013A (760472) = 0000
-2000013C (760474) = 8000
-2000013E (760476) = 0000
-20001468 (772150) = 0000 RQDX3/KDA50/RRD50/RQC25/KFQSA-DISK
-2000146A (772152) = 0B20
-20001940 (774500) = 0000 TQK50/TQK70/TU81E/RV20/KFQSA-TAPE
-20001942 (774502) = 8276
-20001F40 (777500) = 0020 IPCR
>>>sh dev
DSSI Bus 0 Node 0 (RF72)
-DIA0 (RF72)

DSSI Bus 0 Node 1 (DISK1)
-DIA1 (RF73)

DSSI Bus 0 Node 2 (DISK4)
-DIA2 (RF72)

DSSI Bus 0 Node 7 (*)

DSSI Bus 1 Node 7 (*)

UQSSP Disk Controller 0 (772150)

UQSSP Tape Controller 0 (774500)
-MUA0 (TK70)

Ethernet Adapter
-EZA0 (08-00-2B-2B-AA-C0)

I think the “?” for the UQSSP Disk Controller is because there was no actual disk attached.

The next step was to connect the disks. It was at this point I realised that the BA440 enclosure isn’t really designed to be used with SDI hardware. There are no molex power cables to get power to the disk drives, there is nowhere for the OCP to go and no connection for the OCP to the CPU. The VAX 4000-500 requires a separate enclosure for RA-series disk drives, this is confirmed in the manual. I realised that I would have to switch to my MicroVAX 3400, which is housed in a BA213 enclosure. I hadn’t switched this machine on for quite a long time, so I was a little concerned as to whether it would work. I needn’t have worried, it powered up fine.

I installed the KDA50 in the MicroVAX 3400. I then had to remove the DSSI disks, the DSSI media faceplate and the DSSI OCP before I could install the RA72. The problem was how to get power to the RA72. The RA72 uses 4-pin molex power connector, and the 3400 uses DSSI power connectors “natively”. The TK70 tape drive uses a little adapter cable (part number 17-01937-01) to convert from DSSI to the ordinary 4-pin power connector, so I took the TK70 out. Sadly it meant I could only test one disk at a time because I couldn’t find a splitter cable. Finally I installed the RA OCP and connected it all up, with the disk connected to the KDA50 by Port A. Unfortunately I don’t have the RA media faceplate and I didn’t bother to remove the internal cables from the I/O bulkhead at first, so the setup looked a bit ugly but it was functional:

I am not entirely clear about all the buttons on the OCP and without a media faceplate I don’t have the labels to know for sure. The OCP is shown in Table 1-4 of the BA213 manual. The bottom button halts the CPU and puts you back to the firmware console.

When I powered it on I got the following from the firmware console:

>>>sh dev
DSSI Node 6 (*)

UQSSP Disk Controller 0 (772150)
-DUA0 (RA72)

UQSSP Tape Controller 0 (774500)

Ethernet Adapter
-ESA0 (08-00-2B-1A-94-7C)

Ethernet Adapter 0 (774440)
-XQA0 (08-00-2B-0D-AF-71)

This was very encouraging and I was able to boot VMS from a boot node and verify that the disk is OK.

I tried the two RA70 disks next. I had to swap the mounting rails round though so that the end of the disk with the SDI and power connectors faced the front of the machine. One of the disks worked and one did not. The curious thing is that the two RA70 disks showed up in the console like this:

>>>sh uqssp
UQSSP Disk Controller 0 (772150)
-DUA33 (RA70)

and like this:

>>>sh uqssp
UQSSP Disk Controller 0 (772150)
-DUA64 (RA70)

I don’t know where the names DUA33 and DUA64 came from. I know the firmware has some SET HOST/DUP commands, I have used them with my DSSI disks to change things like that, but those commands don’t work with this controller and disk combination, not even if I try SET HOST/DUP/UQSSP. I feel it must be possible but I don’t know how.

The MicroVAX 3400 is back to using the DSSI disks, but the KDA50 is now permanently installed in it too. I would like to find the RA media faceplate for the BA213 enclosure that goes over the OCP. This is part number 70-24534-01 and is also described as a ‘BA213 RA DISK OUTER PLATE ASSY’. I would also like to find a couple of 17-01937-01 adapters to convert the DSSI power connectors to standard 4-pin power connectors.

This entry was posted in Retro-Computing. Bookmark the permalink.

2 Responses to KDA50 and Some SDI Disks

  1. Prof Greg Egan says:

    Good to see a few fighting the good fight Rob. Time I dusted some of my old much more modest Z80 based CPM systems with 8” floppies being the main challenge! A better 2021 for us all. Stay safe.

    • rjarratt says:

      Thanks 🙂 I certainly hope 2021 is better. I used to use a Research Machines 380Z with 8″ floppies, but that stuff is unobtainable (or way beyond what I am prepared to pay!).

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s