PDA

View Full Version : IBM Floppy Drive Behaviour (486 era)



Zelandeth
October 27th, 2009, 09:45 AM
This seems an ideal place to ask a question that's had me scratching my head for a while now.

Ten years or so ago, my main PC was an IBM PS/ValuePoint 433DX/Si.

Old 33MHz 486DX based machine, was well obsolete at the time with an equally aged SCSI harddrive and double speed CD-ROM drive, but it did what I needed it to, and was far better than nothing!

One thing I became aware of as I started to learn more about computers was that the floppy drive sounded a bit different to that in other computers. I since then figured out that it's that the machine seems to run the stepper motor to position the heads more quickly than other machines I've had.

This has the result of making the seek process all but completely silent - only a soft click being heard when the drive heads start/stop moving.

This doesn't seem to be anything to do with the drive itself, as the behaviour has been the same with any other drive tried in the machine - though none have been quite so impressively silent (or forgiving of marginal discs) as the original IBM one.

Is this a behaviour anyone else has noted on some computers? If so, I'd be really curious to know if anyone can explain why it's the case. No reason aside from curiousity!

...Though a way to make the external drive attached to this thing quieter would be nice!

Chuck(G)
October 27th, 2009, 10:07 AM
Much of the "noise" you hear has to do with the matchup of the mechanical characteristics of the positioner with the rate of step pulses being issued by the controller. When the matchup is good, the drive is very quiet. When the controller issues step pulses that too slow, the operation is noisy. Too fast, and you get positioning errors, so one errs on the side of being too slow.

Unfortunately, the NEC 765-based controller found in PCs has to adopt a "one size fits all" approach in determining step rate. Even if that were not the case, the granularity of the step rate adjustment in the FDC is fairly coarse.

I've been playing with a microcontroller driving a variety of disk drives. Even an old Qume 242 5.25" drive can be made to operate nearly silently if the step rate is tuned accurately enough.

IBMMuseum
October 27th, 2009, 10:18 AM
...Ten years or so ago, my main PC was an IBM PS/ValuePoint 433DX/Si...

...One thing I became aware of as I started to learn more about computers was that the floppy drive sounded a bit different to that in other computers. I since then figured out that it's that the machine seems to run the stepper motor to position the heads more quickly than other machines I've had.

This has the result of making the seek process all but completely silent - only a soft click being heard when the drive heads start/stop moving.

This doesn't seem to be anything to do with the drive itself, as the behaviour has been the same with any other drive tried in the machine - though none have been quite so impressively silent (or forgiving of marginal discs) as the original IBM one.

Is this a behaviour anyone else has noted on some computers? If so, I'd be really curious to know if anyone can explain why it's the case. No reason aside from curiousity!

...Though a way to make the external drive attached to this thing quieter would be nice!

Yes, it wouldn't have been the drive itself, because the ValuePoint line at least saw IBM returning to the conventional 1.44Mb floppy drives. I hadn't noticed a noise difference myself, but maybe you had more prolonged exposure than I did (I still have several ValuePoint systems around, so I might try to check the sound levels soon). The external drive that you are referring to, it wouldn't happen to be the IBM 4869 5-1/4" would it?

By the external cable the interface has to allow for the longer timings. In particular if you have to copy floppy (internal) to/from floppy (external) for multiple sectors, the system will run only one drive at a time, and the 4869 has a nice "clunk" to it as it engages the mechanism again and again. Copies using the external floppy to/from a hard drive are much less noisy.

Chuck(G)
October 27th, 2009, 12:01 PM
...Though a way to make the external drive attached to this thing quieter would be nice!

The answer is to patch the driver or the BIOS (whichever is applicable) to use a faster seek time. (Long cables have nothing to do with that aspect, as step rates are measured in milliseconds). In particular, look to the first byte of the FDC "specify" command.

Zelandeth
October 28th, 2009, 10:41 AM
Interesting.

Definitely down to something different being done by the controller rather than the drive itself - though it IS a pretty quiet drive anyway.

Even when doing the normal test of the drive during the post, the frequency of the "buzz" from the head motor is appreciably higher on the old IBM than most of the Pentium/later class machines I have here.

Acorn A5000 here seems to behave similarly - it also seems to actually be willing to jump between tracks more quickly than my current PC, but I realise that the hardware config in the case of the Acorn is very different to your average PC, so it may well handle hardware control of the floppy drive a lot differently too. Should point out that the comparison there was being made using a 1.44Mb formatted DOS disc, rather than anything which might be treated differently such as a 1.8Mb ADFS one.

In the case of this PC, it's just got an external Iomega floppy drive (the internal one having been booted in favour of a Zip100 I found somewhere), it just appears to be a somewhat rattly drive - and the frequency that the head motor buzzes at appears to be precisely that at which every metallic component inside it resonates at, so it's rather noisy sometimes. ...Worse actually dealing with HD discs rather than DD ones.

Unfortunately the IBM in question is currently in bits scattered throughout my parents house at the moment, so isn't really available for a direct comparison.