erelder
New Member
Hi, I worked for MDS in the Netherlands from 1978 until the end in 1989, mostly as a technical support engineer. This thread brought back many memories ...
After the 1100 and 6400 data recorders, MDS made the 2400 and 1200 series as mentioned before, with a whole host of peripherals like 2,5 MB harddisks (2470, 71) and the fast 10MB Marshall harddisk 2478, vacuum tape drives (2433, 35, 37, 39), chainprinters (2444, 45) and (very fast) drum printers (2446), paper tape devices etc. After that came the standalone 2300 series. We had 2400 systems in many banks. Up to 24 ladies would type in bank transactions simultaneously, filling up the 2,5 MB harddisk during the day. After hours, the disk would be "rolled off" (copied) to tape, the tape sent off to a mainframe (mostly IBM) and the disk wiped for the next day. Other customers were departments of our government.
The Series 21 was a really nice machine. The same 2,5MB harddisks (2173 and 2174) that were used in the S21 and the 2400 and 1200 were also used in tanks, very sturdy and with enclosures that either allowed the harddisk package to be exchanged or with enclosures that were closed with a metal front. The only problem we saw was that we at one time had the wrong diskpacks. Diskpacks existed with different numbers of hardcoded sectors, little slots that indicated the start of the next sector. The newer 2175 harddisk system supported a fixed and a variable pack as described before. This fixed pack could crash in a dusty environment. We had some customers right on the seacoast where the packs sometimes crashed because of the salt in the air. It took a few hours to replace all the heads and the pack itself and properly align it. The newer 2160 CPUs also allowed for one or more of the more modern 60MB 5,25 inch harddisks with a special sasi interface board mounted on the drive, sasi being a predecessor of scsi, as far as I can recollect almost identical to scsi except for the lack of hardware parity. These 2160s were amazing if you take into account that several people could be doing data entry or other applications while another person could be compiling a program, all on the same 6MHz Z80. Many tasks in the series 21 were performed in hardware, scrolling of the lines of the terminal was done by a hardware table that listed all the lines and their addresses. One line was reserved for error messages so it was easy to display an error message like #33 in the top left corner by flipping a single address and continuing by restoring the original address in the table. There was a keyboard buffer in hardware, the 2841 fifo chip was used to store max. 64 keystrokes and release them when the system was ready for them. The 2192 screens were driven by a special 93 ohm coax cable, differing from the usual 50 or 75 ohm cable. The newest systems would have memory consisting of the well known 2164 dynamic memory chips and I think that a max of 4 DULS boards could connect up to 8 screens to one 2160. Peripherals not mentioned above: tape drives 2181/82, printers 2141/45. The OS consisted of many separate modules.
In my spare time I wrote some games in Mobol, like lunar lander and a few more, basically to learn the language and also because we only had one game until that time, a horse racing game. It is a pity that no pictures of the Series 21 seem to have survived on the internet.
Later, just before the end, MDS introduced the NGEN (Hero) systems, very modular with separate CPU, Disk etc. building blocks that could be simply connected by plugging them together like a big row of blocks. Every block with their separate power adapter.
This is all from memory so apologies for any inaccuracies.
Best regards, EE
After the 1100 and 6400 data recorders, MDS made the 2400 and 1200 series as mentioned before, with a whole host of peripherals like 2,5 MB harddisks (2470, 71) and the fast 10MB Marshall harddisk 2478, vacuum tape drives (2433, 35, 37, 39), chainprinters (2444, 45) and (very fast) drum printers (2446), paper tape devices etc. After that came the standalone 2300 series. We had 2400 systems in many banks. Up to 24 ladies would type in bank transactions simultaneously, filling up the 2,5 MB harddisk during the day. After hours, the disk would be "rolled off" (copied) to tape, the tape sent off to a mainframe (mostly IBM) and the disk wiped for the next day. Other customers were departments of our government.
The Series 21 was a really nice machine. The same 2,5MB harddisks (2173 and 2174) that were used in the S21 and the 2400 and 1200 were also used in tanks, very sturdy and with enclosures that either allowed the harddisk package to be exchanged or with enclosures that were closed with a metal front. The only problem we saw was that we at one time had the wrong diskpacks. Diskpacks existed with different numbers of hardcoded sectors, little slots that indicated the start of the next sector. The newer 2175 harddisk system supported a fixed and a variable pack as described before. This fixed pack could crash in a dusty environment. We had some customers right on the seacoast where the packs sometimes crashed because of the salt in the air. It took a few hours to replace all the heads and the pack itself and properly align it. The newer 2160 CPUs also allowed for one or more of the more modern 60MB 5,25 inch harddisks with a special sasi interface board mounted on the drive, sasi being a predecessor of scsi, as far as I can recollect almost identical to scsi except for the lack of hardware parity. These 2160s were amazing if you take into account that several people could be doing data entry or other applications while another person could be compiling a program, all on the same 6MHz Z80. Many tasks in the series 21 were performed in hardware, scrolling of the lines of the terminal was done by a hardware table that listed all the lines and their addresses. One line was reserved for error messages so it was easy to display an error message like #33 in the top left corner by flipping a single address and continuing by restoring the original address in the table. There was a keyboard buffer in hardware, the 2841 fifo chip was used to store max. 64 keystrokes and release them when the system was ready for them. The 2192 screens were driven by a special 93 ohm coax cable, differing from the usual 50 or 75 ohm cable. The newest systems would have memory consisting of the well known 2164 dynamic memory chips and I think that a max of 4 DULS boards could connect up to 8 screens to one 2160. Peripherals not mentioned above: tape drives 2181/82, printers 2141/45. The OS consisted of many separate modules.
In my spare time I wrote some games in Mobol, like lunar lander and a few more, basically to learn the language and also because we only had one game until that time, a horse racing game. It is a pity that no pictures of the Series 21 seem to have survived on the internet.
Later, just before the end, MDS introduced the NGEN (Hero) systems, very modular with separate CPU, Disk etc. building blocks that could be simply connected by plugging them together like a big row of blocks. Every block with their separate power adapter.
This is all from memory so apologies for any inaccuracies.
Best regards, EE
