Analogue-sequencer

> Some while ago I told you that the midi2din was working perfect, but 
> not on the cr-8000. The cr-8000 "swinged" like hell, but not in a 
> musical way. You told me that I should compare the clock signals on a 
> scope from the devices I run my cr-8000 with. I did and I found out 
> that the sbx-10 (Roland syncbox) has a "square wave pulse" as a clock 
> signal, and that the midi2din produces a really narrow pulse as a 
> clock signal. You told me that it was not a problem to make the pulse 
> less narrow, but that that would affect the timing accuracy, or 
> something. I would really like a less narrow pulse so how should I do 
> that? I use an external powersupply with the thing, so power is not a 
> problem.

The original Roland clock oscillators in something like a TR808 or TB303
generate a true square wave.
I did a fair bit of research into the importance of this, especially
with respect to the TR909, as I knew it was doing it's shuffle at a
48ppqn resolution, and I thought it may have been using the trailing
edge of the din clock as a reference. But I found that the 909 only
cares about the start of the clock pulse, and interpolates the middle
position.
For the other devices I looked at, I found that as long as the pulse was
more than 2ms long, it would work fine. I had reports of some older
Roland gear having trouble, so I increased the length of the pulse to
4ms.
Do you know if your m2d2m is doing 2ms, or 4ms ?
Making a longer fixed length pulse, which reduces the maximum possible
tempo, is a poorer solution than doing a 50% duty cycle square wave
clock, but that is probably beyond the capability of the PIC 12C508 I
used for the convertor.
Funnily enough, the P3 outputs a true 50% duty cycle square wave sync
pulse, as it has to do clock multiplication of the incoming midi sync to
it's internal 48 ppqn resolution, so it's an easy matter just to toggle
the din clock line on the off at every other 48ppqn pulse, whether
running from midi or internal clock. Does the Cr8000 work OK from P3 ?
I guess you could use another P3 just for this ;-)
Or you could plead with me to do a simple sync convertor with true 50%
duty clock output.

Cheers,
Colin f