[Coco] DriveWire issues with Coco 1 F/285 Board

Torsten Dittel OS-9 at TRS-80.CC
Tue Jul 2 14:09:53 EDT 2013


I assume it strongly depends on the particular capability of your server's
serial interface whether it can reconstruct the signal messed by the slow
op-amp (reduced voltage level for "isolated" bits, triangle pulse shape
instead of rect). What interface are you using (e.g. chipset of USB to
RS-232 converter)?

Does the highest baudrate (I assume that's 57.6 kBaud) require the 1.5x high
speed POKE65495,0? In that case I assume the DW driver needs to run from
ROM, because this POKE doesn't speed up the CoCo1/2 if in all RAM mode. At
least MIDI speed (31.25 kBaud) worked fine at normal speed, I remember the
bit length was 27.5 cycles or so (the LYRA driver didn't work accurate here,
AFAIR it was using 28 cycles for all 1 bits and 27 cycles for all 0 bits,
which only worked for bytes with more or less an equal number of 1s and 0s -
after writing a patch which rather used alternatingly 27 and 28 cycles it
solved the issue of some notes "hanging" during playback).

Regards,
Torsten





More information about the Coco mailing list