Actually it's not the controller, it is the fact that the head gap on the 48 tpi drive is twice as wide as on the 96TPI drive.

So the actual head sees the magnetic fluxes from two different tracks,  This looks like noise to the data separator on the controller.

The way around this was to magnetically erase a disk and format it and write it on a double stepped on a 96tpi drive and only read it in a 48 tpi drive.

This is what we had to do when I wrote the floppy formatter for Gimix 6800 and 6809 Flex and OS/9.  The floppy driver also had an option to double step the drive for normal Flex operation.

On 2/27/2023 2:55 PM, r.stricklin via cctalk wrote:
On Feb 27, 2023, at 10:21 AM, Mike Katz via cctalk <cctalk@classiccmp.org> 
wrote:

the drive would see half the new data and half the old data.
I think that explaining it this way can easily lead to an incorrect inference 
on the part of an arbitrary hypothetical neophyte that what is going on in the 
drive in such a case is that the head can equally well read either the old data 
or the new data but the controller can’t distinguish which is which, or might 
return old data, or might return new data, or might indiscriminately return 
some old data and some new.

What the drive reads in such a case is noise, because the wider head picks up a 
superposition of the old (wide) 48tpi track data AND the new (narrow) 96tpi 
track data, simultaneously.


ok
bear.


Reply via email to