GEC-MARCONI RUGGED COMPUTER FOR LACS

(A tale of computer archaeology investigating a machine of the middle Clansman era)

Introduction

I obtained this computer via e-Bay with the intention of using it as a simple data terminal alongside my Clansman radio sets (as it uses a Clansman style power connector and had what look like Clansman audio and RTTY interface connectors along the back). It was sold as an ex-RAF machine that only got as far as the BIOS setup screen reporting "unexpected amount of memory" (a fairly usual symptom if the BIOS RAM backup battery has expired). It was found on arrival to boot into aminimal version of ROM-DOS (an MS-DOS clone optimised for running from ROM) on a built in ROM drive A: and to have a corrupt 130 Mbyte Calluna PCMCIA disk fitted. The ROM-DOS seems to just install PCMCIA ATA Disk drivers and attempt to mount the C: Drive then run C:\Autoexec.bat.

A visit to a windows XP laptop and running the WinXP disk check utility uncorrupted the disk sufficiently to read it, although all attempts to boot afterwards resulted in an "ERROR: MENU_CFG.PIF not found" message that was inescapable (despite a prompt to press any key to continue). I found that if I booted before inserting the PCMCIA disk I could then navigate around it and examine or run the batch files.

The C:\Autoexec performed some integrity checks (which were failing) and then offered a menu of different BATES sub programs from those installed. It looks like there were about a dozen roles of which this machine has two - CI04 and CI10. CI04 seems to be a battery command post and CI10 is BMETS the weather station. Most of the software is in the BATES directory comprising an upgrade/overwrite utility which I have never dared run, and a subdirectory for each installed BATES role. Each subdirectory contains two disk images, a RAM image for the sub-CPU card installed in the machine, a main program and a batch file that first loads the sub-CPU, then runs the main program after setting environment variables DISK0: and DISK1: to point at the image files. Doing

A:\> C:
C:\> CD BATES\CI04
C:\> SK04

(or the same for CI10 with SK10.BAT) allowed the BATES or BMETS program to be launched successfully.

Some background on BATES

Before continuing with the technical detail it is worth giving some background on Artillery control and communications in the British Army since the 1960s. The best source I have been able to find is Nigel F. Evans' British Artillery Fire Control and I am also grateful for some private communications from the webmaster of that site for what follows.

The first widely deployed computer system for artillery fire control was FACES in the 1970s. This was based on a GEC-Elliott minicomputer with punched tape storage and used Larkspur combat net radio communications. There was an associated weather data system called AMETS. In the late 1970s FACES and AMETS were replaced by BATES and BMETS which ran on a dedicated 8086 processor (same as the contemporary IBM PC) with separate smart VDUs and 20Mbyte Hard Disk Cartridges. The VDU appears to have used a custom 74 columns x 43 rows format. BATES and BMETS used either Ptarmigan trunked radio or Clansman combat net radio for messaging between sites ("Cells" in BATES terminology). The system included a dedicated link encryptor (which I understand was called PALLADIAN) and connected to UK/VRC-353 Clansman radios or Ptarmigan single channel radio access at low/mid VHF frequencies. Each cell had a number of serial peripherals, a number of VDUs, a teleprinter, and one or more radio interfaces. All communication was based on structured messages including that from the CPU to the VDU. All VDU screens were forms based and resulted in messages being added to or removed from a queue (I think processing or sending also removed messages from the queue). BATES was a heavyweight system usually fitted in a FV432 APC chassis. In the late 1980s it was decided to commission a lighter system from GEC-MARCONI called LACS. It appears that my computer is a relic of LACS.

LACS

LACS replaced the bulky 8086 processor and VDU of BATES/BMETS with a laptop-like clamshell computer (although one would need a lot of padding or muscles on any lap to be used with it!) running a form of MS-DOS. It is also unlike a modern laptop in that it depends on an external power supply. The machine consists of an Intel 80386 (Windows 3.1 era) PC motherboard connected to a custom riser board that links to the external connectors (via a flexible circuit board) and one ISA (PC-AT bus) expansion slot. In the expansion slot is a card comprising an 80186CPU (more integrated version of 8086 with most of the supporting chips of the original PC built in), four dual port AMD serial communications chips, and a dual port memory for connection to the ISA bus. Along the back are (from left to right looking into the sockets) a Clansman 2 pin power input, six Clansman 7 pin Pattern 104 male connectors, and one Clansman 7 pin Pattern 104 female connector keyed the same as the ones on the ATR teleprinter-radio interface. I have yet to determine the function of the (presumed) serial connectors - I suspect subject to electrical measurements that the female connector is the teleprinter interface for use with a low voltage teleprinter (Trend 813 or similar).

As far as I can see the way the LACS computer works is that the core BATES CPU and its serial communications are implemented on the custom 80186 card and the storage and VDU are in the laptop main CPU. To run a particular BATES role or BMETS the necessary BATES software and disk image files are first installed into the 80186 card and subsequently the VDU messaging protocol runs between the two processors. It is noticeable that BATES uses the MSDOS 80x43 VGA text mode and never uses columns 75-80. The startup process goes as follows:

Select Disk
Select Image and load mode
Fill in Own, Step-Up and Alternate station info
Select task from main menu

I don't have enough understanding to do this without "inconsistent data" errors except by trial and error yet.

It is also noticeable that the shutdown process is quite involved because of the need to either flush all queues (send all messages) or update the disk images on the laptop disk with any database changes since boot time which would otherwise be lost.

LACS and BATES screens are divided into five areas:

The top four lines are a status area
The next 4 or 5 lines are used for interactive forms
The middle area is used to display menus and data
The Bottom Left corner displays error messages
The Bottom Right corner displays context-sensitive help and Accept/Return/Abort menus when a form is completed

The GEC Laptop has some custom keys to support BATES & BMETS

MENU returns to the root menu abandoning anything in progress
MSG-Q displays the message queue abandoning any previous screen in progress
PROC-Q hasnt done anything in any form I have visited yet
SKIP+ skips to the next field in a form
SKIP- skips to the previous field in a form
HELP sometimes gives help on the current form field or error

Next Steps

I need to finish exploring all of the menu options in BATES and BMETS. As I am aware of several people who intend to display BATES/BMETS (or LACS) vehicles I intend to write a simulated BATES/BMETS/LACS user interface using the 4DOS shell in FREEDOS to provide a system that can boot on period hardware but is free of GEC copyright code or MOD data that may reside in the genuine disk (I have no plan to publish disk images as they may contain ballistic data).