If I remember right, the encoding scheme is standard FM, though I don't remember how it writes data on the two tracks. Data is written into two registers named track 1 and 2 which does not necessarily mean they directly correspond to two tracks on the tape. There was a topic on decoding the MDV signals which ended up resulting in a product that directly emulates microdrives from SD cards.Brane2 wrote: ...
DO different driving strategy, different modulations ( I like my dual signal, QAM-like scheme) etc.
Before that, I have to try it on microdrive.
So I dug one out of my shelves. But motor was loud, so I opened it to lubricate it, at least enough to get trough tests.
...
I took a look at writing signals and they don't look that sophisticated - it's driven by pure digital signal with couple Vpp.
There is PLENTY to optimize there.
Now, replacing FM with MFM would have doubled the density of data on the tape, i.e. increased capacity twofold. Moving to one of the RLL codes like RLL 2,7 would add about 50% on top of that. There are also various versions of GCR which would have also increased data density without increasing the density of flux transitions on the tape, so that's still staying with pure digital recording. Not to mention that the theoretical increase in data density does not have to be used to increase the storage capacity, but rather include automatic error detection and correction encoding, which would tolerate one erroneous bit in each byte.
One odd thing is that RLL 2,7 never caught on on floppy drives, while it became completely common as an upgrade to MFM hard drives of the time, instantly increasing the data capacity of most of them by ~50%.
Then there are still many (x-of-y) coding schemes such as EFM, MEFM. These are interesting because they can encode one byte (256 bit combinatioons) but actually have more than 256 possible codings so it is simple to insert codes to delimit various kinds of data within the bit stream without having to use 'illegal' codes.
Wikipedia is actually quite useful on explaining how encodings work:
https://en.wikipedia.org/wiki/Run-length_limited
https://en.wikipedia.org/wiki/Group_coded_recording