Joshua's VAX page

So, on 20080214, while bored, pyxystyx, zmccord, mjrosenb, 8 and I went perusing campus, and found a whole bunch of VAXen being pitched on Wean 8. In particular, we found a VAXstation 3200, a VAXstation II, and a small workstation-size VAX. Here are my notes as I'm bringing these things up.

The VAXstation 3200

The machine appeared to be in a state of disarray when we found it. We powered it on, and after a while, the hard drive started chunking. It refused to give us any output on the LCD or on serial, no matter how we set it. The front panel lights seemed to do things, though. We moved on to the VAXstation II.

The culprit, by the way, was that the console panel was disconnected. When we pulled the machine apart, we found that not only was the console panel disconnected, though, but the CPU board was disconnected too! I have no idea how it was ever chunking. Anyway, after reconnecitng that sutff, it seems to do a countdown.

I haven't enumerated all the boards in it yet, but there don't seem to be many.


The VAXstation II

This machine came right up when we poked it -- no troubles at all. It did the countdown and booted right into VMS -- that is, after we made a console cable for it.

We had no idea what the VMS password was, so we went to install NetBSD. After we reconnected the DEQNA, we managed to boot it from the network, although I can't reliably do so. The boot process is finicky.

It's about the best feeling ever to see text come down your hacked up serial cable, after hearing this delicious noise.


KA630-A.V1.3

Performing normal system tests.

  7..6..5..4..3..
  
Tests completed.
  
  
>>>

Poking at the firmwares

The hard drive was really slow, and I kind of wanted to poke at the firmware of the various add-in cards. I used the instructions in the Dilog manual to poke at it:


KA630-A.V1.3

Performing normal system tests.

  7..6..5..4..3..

Tests completed.


>>> d/p/l 20088004 80000001
>>> D/P/W 20001F40 20
>>> D/P/W 2000146A 3FFF
>>> S 200

---------------------------------------------------------------
   DILOG On Board Disk Formatter       IP/SA Address = 772150
   Version: A-E-3   Model: DQ696       Boot  Address = 773000
---------------------------------------------------------------


          NO DRIVE SELECTED                                DECIMAL


Are you using a (P)rinter or a (C)RT?

[...]

---------------------------------------------------------------
   DILOG On Board Disk Formatter       IP/SA Address = 772150
   Version: A-E-3   Model: DQ696       Boot  Address = 773000
---------------------------------------------------------------


   ESDI DRIVE 02 (DU01) SELECTED                                 DECIMAL

Display Drive Configuration
---------------------------

DRIVE IS HARD SECTORED         DRIVE DOES NOT HAVE MOTOR CONTROL

DATA TRANSFER RATE             > 5MHZ <= 10MHZ

NUMBER OF CYLINDERS =          1224

NUMBER OF HEADS =              0015

NUMBER OF USER SECTORS/TRACK
   (NOT INCLUDING ONE SPARE) = 0035       INTERLEAVE = 01

USER DRIVE CAPACITY (MBYTES) = 328.6      USER RECORDS = 00642075

Press RETURN to continue
I reformatted drive 2 with an interleave factor of 2.

---------------------------------------------------------------
   DILOG On Board Disk Formatter       IP/SA Address = 772150
   Version: A-E-3   Model: DQ696       Boot  Address = 773000
---------------------------------------------------------------


   ESDI DRIVE 02 (DU01) SELECTED                                 DECIMAL

Format Selected Drive
---------------------
(CTRL-A ABORTS TO NEXT STEP, CTRL-C ABORTS TO MAIN MENU)

INTERLEAVE FACTOR [1] ?2       INTERLEAVE = 02  CORRECT(Y/N)? Y

DRIVE AND ECC TEST

WRITING HEADERS
CYLINDER ADDRESS    1222

WRITING HEADERS
CYLINDER ADDRESS    1223

WRITING DATA
CYLINDER ADDRESS    1223

INITIALIZING RCT TO NO DEFECT STATE

WRITING DATA
CYLINDER ADDRESS    1222

READING DATA
CYL=0082   HEAD=0010   SECT=0031   LBN=00043431  READ DATA ERROR (CORRECTABLE)
                                   CORR PAT=080000(1 BITS CORR) CORR VEC=0199
CYLINDER ADDRESS    1222

Press RETURN to continue
I tried that on drive 1, but it's dead. How sad.

Installing an OS

Since we don't have VMS media yet, we'll try netbooting into NetBSD. It doesn't like doing this while it's on resnet, so I had to connect it to my laptop for this. Also, recent NetBSDs make the machine very sad.


>>> b/100 xqa0

  2..
Bootfile: mopboot

  1..0..


>> NetBSD/vax boot [Oct 23 1999 16:44:26] <<
>> Press any key to abort autoboot 2
Press '?' for help
> boot netbsd.ram.gz
Trying BOOTP
Using IP address: 192.168.1.201
myip: exvax (192.168.1.201), gateip: 192.168.1.1, mask: 255.255.255.0
root addr=192.168.1.1 path=/export/vax
@@@@@@@@@708488+1205248+145848 total=0x1f6d44
Copyright (c) 1996, 1997, 1998, 1999, 2000, 2001
    The NetBSD Foundation, Inc.  All rights reserved.
Copyright (c) 1982, 1986, 1989, 1991, 1993
    The Regents of the University of California.  All rights reserved.

NetBSD 1.5.3 (INSTALL) #5: Mon Jul  1 23:41:35 CEST 2002
    he@turing.urc.uninett.no:/usr/src/sys/arch/vax/compile/INSTALL

MicroVAX II
total memory = 13300 KB
avail memory = 9228 KB
using 191 buffers containing 764 KB of memory
mainbus0 (root)
ibus0 at mainbus0
uba0 at ibus0: Q22
dz0 at uba0 csr 160100 vec 304 ipl 17
mtc0 at uba0 csr 174500 vec 774 ipl 17
mscpbus0 at mtc0mscpbus0: can't init, controller hung [joshua's note: which is this? is this the emulex or the TTI?]
uda0 at uba0 csr 172150 vec 770 ipl 17
mscpbus1 at uda0: version 3 model 13
mscpbus1: DMA burst size set to 4
ra0 at mscpbus1 drive 0: RA80
ra1 at mscpbus1 drive 1: RA80
qe0 at uba0 csr 174440 vec 764 ipl 17
qe0: deqna, hardware address 08:00:2b:03:73:ba

WARNING: you must update your boot blocks.

md0: internal 1138K image area
boot device: <unknown>
root on md0a dumps on md0b
Clock has lost 11870 day(s) - CHECK AND RESET THE DATE.
root file system type: ffs

This time, we'll install to ra1, since ra0 is now dead according to the Dilog. Oops. NetBSD confirms it by spitting errors when it tries to read from it. We picked the default partitions, and let it newfs. We ignored the warning that it couldn't ping anything, since it just fails at life, evidently. Since NetBSD 1.5 is way out of date, we had to change the location that it looked on the network for things -- we changed it to pub/NetBSD/NetBSD-archive/NetBSD-1.5.3/vax/binary/sets.

Once we got it set up to FTP, it actually got quite a respectable data rate from the network -- 134KB/sec! It spat out a lot of qe0: discarding oversize frame (len=1966), but don't worry -- like any reasonable people would do, we ignored it.


The extraction of the selected sets for NetBSD-1.5.3 is complete.  The system
is now able to boot from the selected harddisk.  To complete the
installation, sysinst will give you the opportunity to configure some
essential things first.

Since the timezone selection wasn't going so well, I decided to hit ctrl-c. Unfortunately, then sysinst exited. It said it was done, though, so I typed 'reboot', and got this:


Loading system software.

  2..1..0..


>> NetBSD/vax boot [Jan  6 2002 22:13:30] <<
>> Press any key to abort autoboot 0
nfs_open: must mount first.
open netbsd.vax: Device not configured
> boot netbsd
1174268+57032+195096+[85608+100959]=0x189ca7
[ preserving 186567 bytes of netbsd a.out symbol table ]
Copyright (c) 1996, 1997, 1998, 1999, 2000, 2001
    The NetBSD Foundation, Inc.  All rights reserved.
Copyright (c) 1982, 1986, 1989, 1991, 1993
    The Regents of the University of California.  All rights reserved.

NetBSD 1.5.3 (GENERIC) #5: Mon Jul  1 23:34:54 CEST 2002
    he@turing.urc.uninett.no:/usr/src/sys/arch/vax/compile/GENERIC

MicroVAX II

Success! Well, it then said, /etc/rc.conf is not configured. Multiuser boot aborted.. But, well, I went through the steps in the INSTALL guide to configure my /etc/rc.conf. If you want to see a log of the machine booting, you can take a look.

You can, indeed, compile things on the machine. In my experience, running configure on thttpd looks to be taking around 20 minutes to half an hour. Building it is pretty slow too -- it took about an hour or so. But, it runs! Behold the fruits of our labor, as brought to you in a fashion that only a DEQNA can.

For VMS, we should probably find a second drive, or bring up the TTI. The other VAX has a drive; although I'm not sure why the first drive suddenly failed.

The insides

This machine seems to have the following internal configuration:
[empty] M9047 Grant Continuity Card Emulex QD32
On the back, it says "PWB QU3210701". If it's anything like the QD33, this might work for getting into the firmware.
The DIP switch is configured as 0101; he hasn't anything connected to him.
Update: We have a manual for it.
Update 2: The other DIP switch on the board is configured 0001110000.
M9047 Grant Continuity Card M7504 DEQNA Ethernet Card
It's a shame we don't have a DELQA. The board appears to have some reworks. It wasn't connected to the panel when we got the machine, but we connetced it and it seems to work fine.
A bank of memory.
The memories are 16pin MSM4256 chips. The board is an MV2RAM/8, made in 1986 by Clearpoint, Inc. It's an 8MB RAM board.
Another bank of memory.
The memories are 16 pin MB81256 chips. The board is an MV2RAM, made in 1985 by ClearPoint, Inc. I'm not sure how big an MV2RAM is without a / after it. What a cool board!
M7606 CPU board, which is a KA630-A.V1.3.
[empty] A Transitional Technology, Inc. SCSI card.
The only thing I can find about it is this post on port-vax. You have to connect a console to it to configure it -- as opposed to invoking it by jumping to a specific address. I'm thrilled.
Update: I tried connecting a serial cable to it, and I got no love. I should beep out the pins to the max232 and to a ground, and build a cable, although there's still no guarantee that I know how to poke it to activate the serial. I observed that when I type 'boot duax' for any x, then the light on the TTI blinks; I wonder what'd happen if I added a SCSI disk.
Dilog DQ696-15 ESDI/MSCP controller
DIP switch set to 00000100. It has a Z80! It's pretty slow, but this thread on cctalk implies that if I up the interleave to 2, I'll get MUCH better performance. It also appears to need a console cable to configure :\ we'll see. pwg found this manual for the Dilog.
M9047 Grant Continuity Card M3106 (DZQ11) 4-line Async Card
Looks similar in configuration to the DHV11.

Random notes

Quotes from setting this thing up


Miscellaneous stuff

Console cable

VAXen don't have the same pinouts as PCs, but they do speak RS-232, and they do, confusingly enough, have a DB9 connector. We clipped apart an RS232 straight through connector. Unfortunately, it didn't have the correct color code, so we had to beep it out with a multimeter. We included the colors in our diagram. Our cable had 2 and 3 tied straight through, pins (4, 5, and 6) and (8 and 9) shorted on the VAX end, and (1, 7, and 8) shorted on the PC end. (We got our directions from
a post on the NetBSD mailing list. Here's our diagram:
hacked cable
Color     VAX       PC
Gray      9--+     +9
             |
Brown     8--+   +--8
                 |
Blue      7----+ +--7
               |
Yellow    6+   | +--6
               | |
Green     5+   +-(--5
                 |
Red       4+     +--4
                 |
Black     3------(--3
                 |
Orange    2------(--2
                 |
White     1+     +--1

VAXen give you ADD