Amiga ADF "Ziriax 100%" fails to load

[keirf]

See EAB forum thread for details:
http://eab.abime.net/showthread.php?p=1221573

The ADF itself:
Ziriax (1990)(The Whiz Kidz)[cr ross][t +6 ross].zip

[keirf]

Quick read of the trackloader indicates a 33ms max latency from read-start to sync-found. This is in addition to a 15ms settle delay if any head-step occurred.

Worst-case to find a sync should be sector-latency + track-gap = (544*16 + 4256) * 2us ~= 26ms.

So there should be plenty to play with here. I will debug on real hardware next week.

[keirf]

LA trace shows 200ms+ between retries (seek to track 0 and back) so this is not the 33ms disk-sync timeout. Consistent number of steps to and from track 0 so not missing step pulses.

[keirf]

Bug in Ziriax track loader: skips 300*2 bytes of MFM at the track gap.
This assumes track is at least (544*11 + 300)*2*8 = 100544 bitcells long.
However a DD track is 100000 bitcells +/- perhaps 2%. Usually drives run a bit slow and you end up with a long enough track, but this is certainly not guaranteed.

[keirf]

It has been pointed out that an Amiga generates long tracks because bitcell timing is derived from CCK and thus is a little faster than 2us (precisely 14/7093790 us on PAL Amiga). And some loaders depend on this (Ziriax is an existence proof). Hence it makes sense to make FlashFloppy generate longer ADF tracks.