Scor­pio News

The IBM card (and I guess the others) has some simple and stupid dedicated hardware features. The major ghoulie is two input latches, one which says “80 track/40 track” which relates physically to the drive selected and the other which says “High/​Normal “density”. This could give us four permutations which should cover the three cases we want and one we don’t.

Unfortunately, when “80 track” is selected, it jams the other latch so that it’s automatically high density so any drive assigned with 80 tracks is automatically assumed to be high density. Also, the head step hardware always sends two head-step pulses to the dive but in the case of a “High density” being selected a hardware ‘divide 2’ divider is inserted in the head-step line to ensure one step per tack. If an “80 track” drive is selected and “Normal “density” is asserted, then the divider is switched out to provide 40 tracks on the 80 track drive, and the data rate switches to 250K bits/s. Thus formatting a 360K disk. This explains my finding of a normal 80 track drive double stepping, despite being an 80 track drive and running normal density.

If “40 track” as selected, then it jams the “Density” latch to low and forces normal 360K operation.

That isn’t the end of the story, I’ve since discovered a controller card which claims to solve the problem, although I haven’t tried it (I’m not spending more money on something I’m prepared to live with, although I’d rather not – live with it, that is). The card is the “UDC card” from UPIC of Dunstable.

The story is continued by John and what follows is an edited version of a long letter which I received in the middle of July. It appears to get round the problem with one particular controller (Everest) and throws some beginners light on to some of the other peculiar aspects of MS-DOS.

John Parrott Continues

There are many IBM XTs and ATs at the company where I used to work and the word was that 360k disks written on the high density drives were notoriously unreliable. Of course that made some sense to me, knowing that the difference in the head areas does cause a reduction in the signal to noise ratio when a 96tpi recording is played back on a 48tpi drive, because of the unrecorded areas (or even worse, unerased 96tpi track) “seen” by the 48tpi heads. But having said that, I have to say that with the Gemini one could get away with almost anything most of the time. Anyway, the computer services people were therefore equipping their ATs with 360K 48tpi drives so that disks could be shared around the company.

The time came for me to buy an AT (why, how and which are another story) and the addition of a 360KB drive was an essential item on my shopping list. I was also looking forward to playing with one of those high-density drives – should be a doddle to make them do any format – I thought. I thought a Japanese clone manufactured by Precise from CAS Computers. They fitted the extra drive while I waited and then demonstrated the machine.

I asked for some disks to be formatted and some programs run. All the programs ran and the Norton SI gave a system performance figure of 13.5 (although it typically gives 11.5 now that my system is fully configured). The formatting and disk copying were, however, providing very unreliable even when using the /4 option and a duff 360K drive was suspected, even though I had seen it removed from it’s brand-new packaging. Before changing the drive, we tried formatting in the HD drive. This was also unreliable. Some “experts” were called and they said “ah ha, you’re using quad density disks and they won’t work at 48tpi”. I explained that these were fresh unformatted disks and that my old fashioned CP/M system had no problem in formatting them in any format chosen. I tried to explain that the HD drive was 96tpi anyway etc. but to no avail. They thought it was my disks, I thought there was evidence of a bug is MS-DOS relating to the HD disk and I wasn’t sure about the 360k drive.

I finally persuaded them to try another 360k drive and that initially gave the same results as before. This pointed me towards there being an MS-DOS bug in respect of all floppy drives. I then quickly discovered that if I read a particular type of disk, I could successfully format and write disks of the same format. Diskcopy remained a 50/50 affair (as it had been with the company’s XTs and ATs) but that didn’t matter. I was satisfied that all the machine’s hardware was working and the “experts” were happy that it had all been because of my disks. I left, clutching my new toy.