80-Bus News


November–December 1983, Volume 2, Issue 6

Page 6 of 67

Testing Disk Suitability

One way to test the suitablity of a disk is to do a bit of software twiddling, not so easy as you have to understand the disk drive primitives, but consists of counting disk read failures. Not may people will be prepared to try this as an idea, and it does require a lot of disks to produce a valid statistical sample, so the only way is to gain experience over a period of time. I’ve noted, long term, that certain make of drive X will always be less reliable with disk brand Y than with disk brand Z.

Temperature and Humidity

Finding media errors this way may suggest that it is the magnetic coating which is unsuitable, again not necessarily true, as this assumes that all other things are equal. We have discounted temperature and humidity. Temperature first. If we take the worst case, double sided 96 track per inch drives, the track width is about 7 thou. Now at the outer edge of the disk this means that the head mechanism has to move the head to a known position +/  about 2 thou, not a difficult engineering job provided expansion is ignored. Taking expansion into account lifts drive and disk design from the mundane to the extremely difficult. The base material of the disk is usually mylar, a plastic of well defined characteristics, but one which non-the-less expands with temperature. Likewise, the drive head cantilever (or whatever type of support mechanism is used) also expands with temperature, and being made of different materials, like as not at a different rate. Shugart, in some of their earlier drives took the logical step of making the head support mechanism of the same material as the disk, unfortunately, one of the properties of mylar is flexibility so what they gained by matching the coefficient of expansion, they lost by the head wobbling all over the place. These days nylon head supports are usual, and the shape and length of the support, coupled with the length and material of the lead screw or cam used to move the head are combined to ensure that the head moves with expansion at a uniform and equal rate to the expansion of the disk, OVER A LIMITED TEMPERATURE RANGE. So knowing this temperature range is important, it’s certainly between 15 and 35 C, and most drives are better. If the computer is cased and ventilation is inadequate, it’s surprising how high the temperature can rise within the box. This wouldn’t be the first time I’ve heard of obstinate media/​drive faults causing disk misreads being entirely cured by cleaning out six months accumulated muck from the fan filter!!! Another ‘aside’ on the subject of temperature, why are diskettes always black? I’ll let you work that one out yourselves.

Humidity. Now that’s one I didn’t expect. During the hottest and stickiest two weeks of summer a couple of years ago, we experienced a complete and utter failure of the works Gemini Galaxy to ‘boot’. Once we’d got it booted, then the number of disk errors which were recovered by the retry logic were out of all proportion to those previously experienced or experienced later. We tried the lot, cleaned the filters, ran the thing with the lid off to keep it cooler, changed the drives, everything. Still no better. We noted that the disks read more easily towards the middle of the disk than the outer tracks, and if we PIPed to a newly formatted disk (preferably an old one) everthing was Ok. It all pointed to the disks being in trouble, but why? This situation went on for some days until a wandering disk rep. happened to drop by. After greeting him with my usual banter specially reserved for getting rid of people trying to sell me things, I told him his blankety-blank disks were useless anyway as we couldn’t get them to boot. His cool reply was that, ‘What did I expect when the temperature was in the eighties, and the humidity was 99%!!’ It turns out (quite logically) that disks absorb atmospheric moisture

Page 6 of 67