Archiving ST-506 drives

On 30/11/2011 21:48, Chuck Guzis wrote: It doesn't work like that... Acquiring data like that is wasteful -- the width of the pulse which the drive outputs when a flux transition is detected is basically irrelevant and carries no useful information. However, if we know the time between the leading (active-going) edges of the pulses, we can reconstruct the transition stream. Once we have the transition stream at the desired sample rate, we play it through a software implementation of a data separator. Now we have a stream of '0' and '1' bits. Finally we separate these into clock and data bits. The clock bits are compared against the output of an MFM encoder, and used to calculate a "confidence factor" -- the assumed possibility that this really is a valid MFM stream. Dividing the number of "valid" clock bits by the total number of bits produces a confidence factor ranging from 0..1 (floating point). The higher this value, the more likely that the data is MFM. The data bits are separated out, and reconstructed into bytes. These bytes are saved to disc. There's a bit of extra logic 'in the middle' to detect sync words (0x4489 encoded -- SYNC-A1) but that's about it. Cheers, -- Phil. philpem at philpem.me.uk http://www.philpem.me.uk/