Home MiniComputers Peripherals The RL02 disk drive and the RL11 (M7762) controller
Installing and testing the drive on the PDP11/44 w/Unibone
About

Installing and testing the drive on the PDP11/44 w/Unibone

To test I need to run xxdp on the PDP11/44. This is done the easiest using the Unibone as I have no other working hardware at the moment.

It seems important to place the Unibone after the RL11 controller, because this makes the NPR and grant lines go directly to the controller without the Unibone in between which would introduce a delay. The disk tests are often quite time-critical.

The Unibone sitting behind the RL11 in the backplane, so the NPR and grant lines reach the controller without passing through it first.
The Unibone sitting behind the RL11 in the backplane, so the NPR and grant lines reach the controller without passing through it first.

Controller 2

The first test was done with controller 2.

The xxdp2.2 Unibone script emulates a set of RL02 drives by itself on the first RL11 controller. This means that we will have to set the controller we want to test to be the second controller in the system, or we need to boot from some other (emulated) device on the Unibone.

Setting the controller as the second one

The first RL11 controller has the following settings:

  • Base address 774400
  • Vector address 160

The RL11 takes 10 (oct) bytes of I/O window (4 registers of 2 bytes each). The vector address increments by 4. So the second controller should be set to:

  • Base address 774410
  • Vector address 164

For this we need to change the dip switches. First the vector address which is encoded as follows:

The vector address scheme from the manual. Address 160 wants W3, W4 and W5 in with W1, W2 and W6 out, so moving to 164 means putting W2 in as well.
The vector address scheme from the manual. Address 160 wants W3, W4 and W5 in with W1, W2 and W6 out, so moving to 164 means putting W2 in as well.

This shows that we need to switch W2 to jumpers IN. These are not DIP switches but actually 0ohm resistors soldered to the board:

The vector jumpers are not DIP switches but zero-ohm resistors soldered to the board, the dark red ones here. The arrow marks the position that has to change.
The vector jumpers are not DIP switches but zero-ohm resistors soldered to the board, the dark red ones here. The arrow marks the position that has to change.

The tech drawings show that the jumpers are from W6 at the top downwards to W1 at the bottom:

The technical drawings number the jumpers from W6 at the top down to W1 at the bottom, which puts W2 second from the bottom.
The technical drawings number the jumpers from W6 at the top down to W1 at the bottom, which puts W2 second from the bottom.

So we need a jumper on the 2nd thingy from the bottom. I decided to add a small switch for that so that we can easily revert without soldering on these fragile board too much:

A small switch fitted in the W2 position instead of a soldered link, so the setting can be reverted without working on the board again.
A small switch fitted in the W2 position instead of a soldered link, so the setting can be reverted without working on the board again.

The CSR address needs to be set using the following schema:

The base address scheme. 774400 has W12, W16 and W7 in; reaching 774410 means switching W13 on as well.
The base address scheme. 774400 has W12, W16 and W7 in; reaching 774410 means switching W13 on as well.

We will need to toggle W13 ON. I added the same switch there after removing the wire wrap posts from that location.

Running the tests - controller 2

ZRLG

ZRLG on controller 2: one pass, no errors at all.
ZRLG on controller 2: one pass, no errors at all.

ZRLH

This test fails:

ZRLH stops in test 002 with a device fatal error. RLCS reads 112313, with composite error, operation incomplete and header not found all set.
ZRLH stops in test 002 with a device fatal error. RLCS reads 112313, with composite error, operation incomplete and header not found all set.

cs 112313 (94cb h):

  • composite error (15)
  • header not found (12, bit 10 is set)
  • operation incomplete (10)
  • controller ready (7)
  • interrupt enable
  • function code: 101 (write data)
  • drive ready

ZRLJ

ZRLJ fails its seek test at test 007, sub 153, with the read header reporting cylinder 637.
ZRLJ fails its seek test at test 007, sub 153, with the read header reporting cylinder 637.

Controller 1

That was depressing.

I swapped the controller. Controller 1 still uses its original addresses, so I ran the xxdp software through the Unibone using RX02 emulation. This caused some trouble: for some reason the 11/44 would start running immediately after a "pwr" command, and it would not boot from the Unibone provided files.

After some investigation I think this was because my 11/44 was set to boot automatically after powerup using settings on the M7098 Unibus module. I switched autoboot off, and that made the start from rx02 emulation work. For details see Boot proms.

The following tests work on this controller (2 passes each):

  • zrlg e0
  • zrlh b0
  • zrli d1
  • zrlj c0

Testing the drive with RT11

Next step was to try to use the drive under RT11. I booted RT-11SB V05.07 from the Unibone (/rt11v54_ry0.sh) and checked that it has the RL02 driver:

RT-11SB  V05.07  

.R MSCPCK

.show
TT  (Resident) 
DY  (Resident) 
    DY0 = DK , SY 
LD   
SL   
DL   <-- rl02
VM   
SP   
LS   
NL   
13 free slots

Then tried to access the disk:

.dir dl0:
 
?DIR-F-Invalid directory

Not RT-11 format, obviously, so trying to format:

.init
Device? dl0:
DL0:/Initialize; Are you sure? Y
?DUP-W-Replacement table overflow DL0:
Type <RET>, 0, or nnnnnn (<RET>)
Replace block # 
?DUP-W-Replacement table overflow DL0:
Type <RET>, 0, or nnnnnn (<RET>)
Replace block # 0

.dir dl0:
 
FILE  .BAD     1                 
 1 Files, 1 Blocks
 20381 Free blocks

Apparently some bad block? Copying data however failed with I/O errors. I tried to init the disk again:

.init dl0:
DL0:/Initialize; Are you sure? Y
?DUP-F-Error reading bad block replacement table DL0:

.init/REPLACE:RETAIN dl0:
DL0:/Initialize; Are you sure? Y
?DUP-F-Bad block in system area DL0:

I tried several other packs, but they all failed in the same way (I have only 2 without a marking that says they have a read error at the start).

Testing with ZRLM..

A tip from vfced was to try ZRLM, this contains a full disk scan:

ZRLM, the utility that formats bad sector files. Option 4 verifies the pack read-only.
ZRLM, the utility that formats bad sector files. Option 4 verifies the pack read-only.

Running option 4 delivered first:

The first full read: the pack completes in 1 minute 48 with no soft and no hard errors.
The first full read: the pack completes in 1 minute 48 with no soft and no hard errors.

After that retrying causes an odd error:

Retrying reports a soft error at cylinder 511, sector 63, head 1, followed by two ILL INTER 100 traps.
Retrying reports a soft error at cylinder 511, sector 63, head 1, followed by two ILL INTER 100 traps.

Restarting the machine and re-running reports an error now:

After a restart the same error at the same address, but now at a time of 00:00:00 instead of after a complete read.
After a restart the same error at the same address, but now at a time of 00:00:00 instead of after a complete read.

Completely switching off and retrying made the read run successfully twice, after that there was another error:

After a full power cycle the read runs twice, then reports that same cylinder 511, sector 63, head 1 error at 3 minutes 30.
After a full power cycle the read runs twice, then reports that same cylinder 511, sector 63, head 1 error at 3 minutes 30.

Trying option 5, "Write pack with worst case data pattern" failed with:

Option 5, the worst case write, gives up after one second at cylinder 001, sector 63, head 1.
Option 5, the worst case write, gives up after one second at cylinder 001, sector 63, head 1.

What is also odd: running ZRLGE0 after all this also aborts immediately:

ZRLG now aborts on startup with ILL INTER 100 traps, before it runs a single test.
ZRLG now aborts on startup with ILL INTER 100 traps, before it runs a single test.

Restarting the whole bunch and running zrlg:

After restarting everything ZRLG reaches test 037 and fails on the difference word: 000400 expected, 000000 present.
After restarting everything ZRLG reaches test 037 and fails on the difference word: 000400 expected, 000000 present.

Something is rotten in the state of Denmark..

Last modified: 2026/09/06
© 2024 by Frits Jalvingh. All information in here can be freely used (MIT License)