ND-110 Satellite 9883.21238

From NDWiki
Jump to: navigation, search

An ND-110 Satellite model T9, upgraded to T17, donated by Gandalf.

Configuration

  • Hard disk: 170 MB (126MB formatted) Micropolis SCSI disk, SE 50pin, ND designation ND 10 62 80
  • Floppy: 1.2MB 5.25" floppy, model TBD
  • Tape streamer: Tandberg TDC 3640, ND designation ND 11 02 17, serial number 379392
  • Slot 1: 3013 ND100 8-TERM I/F, print: L, eco: N
  • Slot 2: 3094 ETHERNET IF. II - print: E, eco: G
  • Slot 3: 3095 ND110 CPU&MM 48B, print: B, eco: H, 39.3216 MHz crystal
  • Slot 4: 3201 SCSI/FLOPPY, print: B, eco: K
  • Slot 5: 3042 ND100 2 MBYTE RAM, print: B, eco: D
  • Slot 6: 3015 ND100 HDLC W/A.LO. print: S, eco: U (TBC)
  • Slot 7: 3013 ND100 8-TERM I/F, print: L, eco: N

Power Supply

  • Wiener DN 03, 390 W

Labels

SYSTEM 9883.21238
FLOPPY: 187-3784
STREAMER: 229-1282
DISK: 228-1940

The disk itself has label '228-319', so it is likely that this is a replacement/upgrade disk. The ND 9883 models are also known as T9, and like the T17 model they were originally spec'ed with 125MB disks. But this one has a 170MB disk (probably unformatted capacity though - see below) as well as an additional terminal board, an Ethernet II board, and an HDLC board (so that all positions are now in use in this particular system). In that sense this T9 system has been upgraded to a T17 system (due to the extra terminal board), with extra communication.

CPU

This system has an ND110/II (110 version 2) CPU, with 48-bit floating point hardware (note the '48B' on the 3095 CPU board designation)

Operating system version

SINTRAN III/VSX version K patchlevel 10200B

History and trivia

  • The last time the system was booted appears to be at 23:09:33, 21 August 1994.
  • The boot process took 6 minutes and ten seconds.
  • It may have been shut down at 23:24:48 the same day, but this cannot be verified.
  • This system was part of a network of many remote ND computers, and Tingo's ND-110 Satellite 9883.21005 was also a member of this network.
  • It is likely that ND-110 Satellite 9883.21251 was also part of this network, as its configuration files include the CPU of this system (21238) even though 21251 isn't mentioned in 21238's configuration. And they share the same system password, which is another clue. But it's possible that 21251 was decommissioned a couple of years before 21238, and that's why 21251 isn't part of 21238's configuration.

DISK

The disk is a Micropolis SCSI disk stamped with "170MB", but after imaging the disk I see that it's more likely that this is actually the disk spec'ed as 125MB for the T9 and T17 ND 110 Satellite series. The actual capacity is 132MB in 'modern' disk size speak (132 million bytes), but during ND's time the industry hadn't yet changed into the numbers war and the original KB=1024, MB=1024 KB interpretation was still used. So that disk would be 126MB around then, conveniently described as "125 MB". The stamped size of 170MB would be the unformatted size then, which sounds likely after some back-of-the-envelope calculations.

fdisk and scsi info from Linux computer used for imaging:

Vendor                           : NDMICROP 
Model                            : 1375   
Rev                              : B0C
ANSI SCSI revision               : 00
Sector size (logical / physical) : 1024 bytes / 1024 bytes
Capacity                         : 132 MB, 132415488 bytes
Disk identifier                  : 0x81ddc629
No partition table

The Micropolis 1375 disks can still be found on ebay and even (much more expensively) on Amazon. It's often described as '145 MB formatted', or '153 MB formatted' or something else, but pictures confirm that these drives have the same '170MB' stamp on their printed media defect table. So it's the same drive. Specifications can be found here, and a PDF file is on bitsavers.In the latter document we find that formatted capacity would typically be 150MB with 1024 byte sector size and 146MB with 512 byte sector size. However, from the same document it is also explained how to calculate the formatted capacity with various options. It looks like ND used a very conservative setup with 3 spare sectors per track (the maximum), and possibly 10 spare cylinders. That leaves us with only 96 sectors not accounted for.

The disk was advertised at $2295 w/PC controller in 1988 (source: Infoworld, Mars 7 issue, 1988).

The system password found on the disk is the same as for ND-110 Satellite 9883.21251, 'FIXSEG', so it's likely that this satellite also originated from the same administrative unit.

TODO

  • Image the SCSI disk before attempting a boot DoneDone
  • In addition to getting a backup of the disk, imaging it will also get access to the password DoneDone
  • Make a C15 power plug by violently abusing a C13 plug Acquire a C13 to C15 power plug adapter DoneDone
  • Check the power supply condition, start with the test measuring points. DoneDone Measures perfectly.
  • Make a cable for the console terminal DoneDone
  • Disable UE-LOGIN so that I can actually log in DoneDone
  • Set up a wi-fi console connection. By chance I found a basically NOS current-loop to RS232 converter on ebay (on a 1 euro bid), so the plan is to use that one and a serial-to-wifi converter to give me remote console access.
  • Start testing SCSI2SD as an alternative to the physical disk.
  • New: Fix the keyboard of the Nokia Notis-terminal to make it usable again.
  • New: Try to figure out what's going on with the front panel. It doesn't work properly.

I didn't notice right away that the computer needs a power cord with an IEC 320 C15 power plug, which I don't didn't have. I don't think I've seen one of those ever used. But I hacked a normal C13 into the same shape, so I'm OK for now. I also got a C13-->C15 adapter from Hong Kong, it arrived surprisingly quickly. Looks great, and it lets me use any standard IEC 320 power cable.

Status

Booted, after setting the memory card thumbwheel to 0 0 0.

The system is a bit difficult to power on, for some reason. It needs two or three tries before the console or the front panel buttons respond. After the second or third time I can press the rightmost button followed by the leftmost button and get the front into 'advanced' mode (where it shows 'macl', 'opcom' and 'load' buttons). After the first power on the START and STOP buttons are lighted up, but nothing responds. Power off, wait a bit, power on, and the front panel and console works. Strange. I don't have an ND test floppy so I can't boot from floppy and test the system (well I have a bootable SINTRAN install floppy, but I don't want to boot that one).

Unfortunately, User Environment is enabled on the console, and the normal system password doesn't log me in. So I'm still a bit stuck until I figure out how to bypass UE. I should have a UE manual somewhere, but due to time constraints it will have to wait for a while.

As of 2020-04-09 the Nokia terminal keyboard has developed some issues, some keys stuck easily, the space bar in particular. This creates challenges when I'm using PED to edit files. Still managed to edit a little in LOAD-MODE:MODE today, but I have to do some maintenance on that keyboard. The computer itself has some problems booting, today I managed to boot it once (and then everything looked normal), but tries both before and after failed. It just didn't boot, no output. The front panel clearly has some problems, I can (usually) get MCLR to work (master clear), but LOAD doesn't seem to work. So I have to look at this as well. Disk is a bit noisy and I really would like to get SCSI2SD working, particularly as it could be difficult to find a usable spinning rust hard disk, ref. those 1024 sectors etc.