• Please review our updated Terms and Rules here

Today: a ND-100/CX Compact computer

Small update: today I managed to write an image (on a FreeBSD machine with a 5.25 inch floppy drive) that was close enough that I got the "RUNNING" indicator to light up on the panel of the machine. But, sadly it didn't boot completely. Probably some more magic is needed.
Here are the commands I used to format the floppy and copy the image:
Code:
root@kg-t2# fdcontrol -f 1232 /dev/fd1
root@kg-t2# dd if=/home/tingo/dl/nd/floppies/Testdisk.raw of=/dev/fd1 bs=1024
1232+0 records in
1232+0 records out
1261568 bytes transferred in 230.682619 secs (5469 bytes/sec)
(the FreeBSD machine was running FreeBSD 8.2-stable, and it is an i386 machine)
I then took out the floppy, put it into the ND-100/CX and entered:
Code:
1560&
(without pressing ENTER)
The machine read a bit (one sector? one cylinder? Something else?) from the floppy and turned on the "RUNNING" indicator. After that, nothing happened. No reading from the floppy, no prompt on the console, nothing.
 
I'm following your log (here for those who missed it in the other thread). It's great to see that the machine gets some love :)

Is the hard drive completely dead? I managed to boot the drive twice (with some failed attempts in between) into some kind of system I was unable to interact with at all (no prompt).
 
The 1560& procedure will attempt to read the first 2048 bytes from the floppy (and then start execution from address zero, unless my memory fails me. But that's documented in the floppy controller manual I believe, so it can be checked there).

-Tor
 
Is the hard drive completely dead? I managed to boot the drive twice (with some failed attempts in between) into some kind of system I was unable to interact with at all (no prompt).
Well, so far it is just .. unresponsive (doesn't react to the START button, nor to '&' on the console). I haven't done much to test / fix it yet; just re-seated the connectors. More info to follow.
Good to know that you managed to get it to life, then there is still hope.
 
Last edited:
Did more testing on the hard drive today; unfortunately without much success at all. I am wondering; perhaps it doesn't spin at all? I took the hard drive out of the machine, and unscrewed the bottom circuit board (it's hinged one one side) and lifted it (carefully). I can't see anything wrong with it (no burnt or visually damaged components), and the flywheel (?) at the bottom turns easily around.
 
The flywheel is the motor. Check if it actually gets 12V & 5V on the connector of the power supply. Old HDD's sometimes have SMD fuses near the power connector as well. You can also try to power up the HDD using a standard PC PSU, as long as the connector is the same. Could also be that the HDD's does not spin up until the controller tells them to (Like SCSI).

I don't know anything about ND; so the above is just general troubleshooting.
 
I forgot to write about that: I measured the voltages (the drive is located with the connectors pointing out of the frame, so it's easy) on the power connector. +5V was spot on, +12V was 11.61V.
The drive is ST-506, a 85Mbyte Micropolis 1325. Does anyone know if ST-506 drives generally starts spinning when powered up, or if they need a signal for that?
 
I can't offer much insight on the hard drive. But the if I remember correctly the PSU has points of measure and buttons for adjusting the voltages.
 
I forgot to write about that: I measured the voltages (the drive is located with the connectors pointing out of the frame, so it's easy) on the power connector. +5V was spot on, +12V was 11.61V.
The drive is ST-506, a 85Mbyte Micropolis 1325. Does anyone know if ST-506 drives generally starts spinning when powered up, or if they need a signal for that?

They spin up as soon as they get power, I think. It is not a SCSI variant of the drive? If yes, then it might have the delay_spinup jumper enabled, which causes the disk not to spin up unless the controller tells it to.

You can also take the HDD out of the case, and probe around with a multimeter to see if the 12V and 5V reaches further into the circuit board.
 
Last edited:
Fixing a Micropolis (ST-506) drive

Fixing a Micropolis (ST-506) drive

From a discussion on the cc-tech mailing list, I got a description on how to fix the "sticky bumper" problem on a Micropolis ST-506 drive. These drives have rubber "bumpers" as end-stops for the drive head. And after many years this rubber becomes sticky, resulting in the head not having enough power to "unstick" itself from the bumper when powered up.

The solution is (for me at least) quite scary: open up the drive, move the head and put a piece of paper between the rubber "bumper" and the head part. Then all will be well (if the drive doesn't have other problems).

So I did. First I had set up everything I needed: a psu to test the drive with, some post-it paper, screwdriver, scissors, a small flashlight.
The top cover has ten screws, and six of them is partly or totally obscured by the label. I simply peeled away the label above the screws.
After unscrewing, I used a small flat-blade screwdriver to loosen the cover; the gasket made it stick to the housing. Then I lifted the top cover carefully off (on the inside, the airflow duct is connected to it), and started to look for the bumpers.
Since I haven't seen the inside of one of these drives before, it took a while before I realized that the "bumpers" where inside the head assembly; there is only two small slits where you can see them when you move the head. There is also a locking mechanism that prevents the head from moving freely if the drive isn't powered on.

When I moved the head assembly with my fingers, there was very little resistance; I started thinking that maybe there was something else wrong with this drive.

But I cut a small strip of post-it paper (maybe 2 or 3 mm wide), moved the head a bit and put it into the hole between the head and the bumper that the head rested on when turned off. Simply releasing the head made the paper stick to the bumper.

Testing; I powered up the drive (yes, with the cover off) and now the head moved out when the drive had finished spinning up. Before the paper, the head hadn't moved at all. Aha!
It was the "sticky bumper" problem after all. I re-assembled the drive, powered it on, and let it be on for about four minutes. Success!
 
Testing the machine

Testing the machine

After fixing the drive; I had to test if it worked.
I put it back in the machine, connected it up, and powered up the machine. I waited maybe a minute or two, but nothing happened, so I pressed the "START" button.
Almost immediately, the "RUNNING" light on the panel turned on. This hadn't happened before!

After a few more minutes, I got this on the console (transcribed):
Code:
BATCH NUMBER =  1
13.14.12     15 SEPTEMBER 1987
SINTRAN III - VSX / 500 K  ******** VÄLKOMMEN TILL DEMO SIMULATORN ********
STANDARD CONFIGURATION: C
GENERATION (WORK MODE NO.):  312B
REVISION (PATCH FILE NO.): 11400B
CPU TYPE: 102
CPU NUMBER: 3608
GENERATED: 13.14.00     15 SEPTEMBER 1987
SINTRAN III RUNNING -

PAGES FOR SWAPPING: 1357B

ND-100 PANEL CLOCK INCORRECT

1987.09.15  13:14:21 ** XROUT: xmsg  version L03 (87.11.26) started **

*** 13.14.40 TERMINAL 670:
***INFORMATIONS SIMULATORN ÄR NU FÄRDIG ATT ANVÄNDA ***

Yes! The machine is working again!
Pressing ESCape got me the "ENTER" prompt, and I could log in as user SYSTEM (without password):
Code:
OK
SIM-INFO..@
Happy now. :D
 
Awesome! It looked the same the few times it worked for me!

Also, I don't think I gave you the systems history. I got it from a university which ran a big ND-500 to control a simulator for of a small Nuclear Reactor. The simulator was quite neat, complete with control room and everything. The software on the ND-500 had been ported to a GNU/Linux system. So the ND-500 was donated to a friend of mine and later made it's way to another collector in Sweden. The ND-100 was sitting next to the 500 and according to the donor had never been used for as long as he could remember. However, I suspect it is somehow related.

Congratulations on getting it going again!
 
Also, I don't think I gave you the systems history. I got it from a university which ran a big ND-500 to control a simulator for of a small Nuclear Reactor. The simulator was quite neat, complete with control room and everything. The software on the ND-500 had been ported to a GNU/Linux system. So the ND-500 was donated to a friend of mine and later made it's way to another collector in Sweden. The ND-100 was sitting next to the 500 and according to the donor had never been used for as long as he could remember. However, I suspect it is somehow related.

This is very interesting. Do you know if there is any information available on this simulator?

pontus said:
Congratulations on getting it going again!
Thanks. It was a really satisfying moment. :)
 
I wonder how they transferred data to and from this machine?
- it has TCP/IP software installed, but not Ethernet card (and no cabinet wiring or connectors on the back plate for that either)
- the only terminal is the console; not other terminal interface cards in the machine (again: no wiring and no connectors for that on the back plate)
- there are no other communication cards (HDLC, Megalink or others) in the machine

But the TCP/IP software is configured with a list of systems to talk to...
 
It sounds like there must have been an Ethernet interface card in the machine at some point. The ND-100 we had at work had an Ethernet card and that was the main form of connection (well, the first years we had a bunch of terminal card interfaces in the machine and we wired current loop to all offices). Admittedly there was more room in our non-compact-model ND-100, but there's room in the crate of your machine for more cards than it currently has, no?

-Tor
 
Yes, there is plenty of space in the card crate:
  1. 3033 - PCB 3033 ND100 CPU, Print: S, Eco: V
  2. 3104 - PCB 3104 ND100 Mem. Man. II, Print: G, Eco: N
  3. (empty)
  4. (empty)
  5. 3041 - PCB 3041 ST506 Disk Contr. Print: E, Eco: K
  6. 3112 - PCB 3112 8" + 5 1/4" fl. + str., Print: B, Eco: J
  7. 3009 - PCB 3009C Local I/O Bus, B-interface address 151X
  8. (empty)
  9. (empty)
  10. (empty)
  11. (empty)
  12. 3042 - PCB 3042 ND100 2 Mby RAM, Print: B, Eco: C


But I am puzzled by the lack of wiring; I can unsderstand that they used the cards elsewhere; but the wiring?
 
This is very interesting. Do you know if there is any information available on this simulator?
I found this document (PDF). I don't know if it is the correct one, but it has interesting information about the Compact Simulators:
the first so-called Compact Simulators were build in Studsvik in the early 1980s.
...The old system used a so called minicomputer
of the Nord 500/100 type, a two-CPU system with a size of about two large
refrigerators. A four meter long operator’s panel was connected to the Nord
computer. An external computer terminal was used to communicate with the
Nord system. The system featured five large CRT colour screens, controlled
from the Nord computer, via a display system based on a microprocessor and a
several graphical electronics cards under the panel.
and about the software:
The present software is built around a condensate of the old Fortran code from
the Nord system. All the main physical processes in the reactor simulation
uses the old code, modified in order to run on a GNU/Linux system on the
x86 architecture. The Fortran modules are built into a framework of C-code,
handling the general control of the software, including timing, a graphical user
interface, and a display system for the five LCD screen on the operator panel. A
dedicated kernel driver written in C that controls the communication between
the parallel port of the PC and the external interface card is loaded into the
Linux kernel at system startup.
The Fortran modules responsible for the different parts of the reactor plant
use a total of between 6500 and 7000 process parameters (variables and con-
stants). These are shared between the modules by a number of large COMMON-
blocks defined in Fortran.
Very interesting.
 
I don't know much about the simulator. But the PDF you found is definitely related. How do I know that? Well, the physical description is spot on, I know the fellow who ported the software to GNU/Linux and I know the simulator was used for teaching students about nuclear physics.

Also, If you'd like a network card or other parts I could get you in touch with Göran in Umeå. He is the mother lode of ND in Sweden :)
 
Back
Top