A more complex MIDI-DIN Sync convertor?

[Dan Gendreau]

Byron G. Jacquot Wrote:

> I'm drawing this form ancient memory here, but if I remember
> correctly, pin
> 4 is actually a reset line.  I think it goes something like this:
>
> When run/stop rises high, the machine will start up.  If reset is active
> then, it starts at the beignning of a pattern.  Otherwise it starts from
> where ever it was when it stopped.

As you can see below, The definition of pins 4,5 are inconsistent. Some
machines use pin 5 for "Continue", the TR808 uses pin 4 for "Reset/Start"
(not even sure if it works the same). The TR808 uses pin 5 for "Fill In".
The TB303/TR606 use Pin 4 for "Tap".

MC202: "4=N/C, 5=Continue"
TB303: "4=Tap, 5=N/C"
TR606: "4=Tap, 5=N/C"
TR707: "4=NC, 5=Continue"
TR808: "4=Reset/Start, 5=Fill In"

I do know that if you leave the pin 4 and 5 unconnected or at 0V, all DIN
gear will behave in an overall sane manner, except that they always start
from step 1. Perhaps I could break out a Reset/Continue signal from the
converter chip and leave it up to builders how they wish to wire it.

> Hmmm...2x might be tricky...if the TR808 only sees fairly wide
> clock pulses,
> it wouldn't be too long before it stopped chasing properly.

The TR808 needs at least a 3ms on-time for its pulses, which means it wont
work reliably above 400BPM. This means you cant run the TR808 above 200BPM
at double speed. The only ppl who want to do this are Rotterdam house
producers, so I think its not a bad limitation. hehe :) But seriously, if 2X
speed is too fast for the TR808 to sync reliably, then I would use the old
doctor's catch phrase "then dont do that". :)

> You might also
> need to do a little math/predictive timing to get the in-between pulses in
> the right places...if the input clock were of 50% duty, then you could
> simply send a shorter pulse on every rising or falling edge.

I have already worked all that out. It will be done with a software Phase
Locked Loop. That will make 50% duty cycles and shuffle possible. The only
side effect is that during rapid tempo changes, the PLL output may not be in
perfect sync with the midi clock (may lead/lag by several ms), but once the
tempo evens out, it will lock back in phase(ie it will not drop pulses). By
rapid changes, I mean extremes like winging the tempo from 100BPM to 300BPM
in 1/2 second. Like a hardware PLL it will have a limited slew rate on its
output frequency, but it will be well with acceptable limits.

> Look for the Atmel AVR dev kit, the STK200.  You could easily do the above
> with it.
I'm glad to see I made the right choice. Already have it ordered. Thanks! :)

-Dan G.