[sdiy] x0x memories (TB-303, TR-606, TR-808)

Colin f colin at colinfraser.com
Thu Jan 20 23:19:38 CET 2005


 
> > At 12:49 20/01/2005, Mattias Rickardsson wrote:

> >> The TB-303, TR-606, and TR-808 have a microprocessor 
> (containing also 
> >> the program code, I believe) and some memory ICs beside it 
> (holding 
> >> the pattern/song data, I believe).
> >>
> >> Is this external memory used in such a way that it could be 
> >> "emulated" by some other thing connected to the same 
> microprocessor 
> >> pins?

The CPUs in the 303, 606 and 808 run at very low speeds compared to modern
microcontrollers.
A memory emulator should be fairly easy to do.
I was thinking of doing one myself to help with a reverse engineered 303 CPU
project, with a logging interface so I can see exactly what's going on.

> >> What I mean is basically this:
> >> Does the 808 read the external memory ICs continously 
> (like a sheet 
> >> of music notes) while playing, so that an emulator could decide at 
> >> any time what the next step should contain? Or is the 
> external memory 
> >> read in larger chunks, stored in the microprocessor while playing, 
> >> and nothing can change it while playing?

The CPUs are very simple, I don't imagine they are doing any caching of
current patterns.

> Richard Wentk wrote:

> > I'm pretty sure the drum machines are step based. But the 
> 303's glide 
> > feature would need some lookahead.

That's an internet propogated myth, I'm afraid.
The 303 slide operation is very straightforward.
The slide circuit is activated by the same line that loads the DAC latch.
If a note is not slid, the line is pulsed. If it is to slide, the line goes
high to strobe the DAC then stays high to enable the slide.
This happens on the first clock pulse of the beat, same as for a non slid
note.

In fact, the 303 unnecessarily re-writes the current note to the DAC even if
the note was tied from the previous beat, which hints at the general
simplicity of the code.

> > What are you trying to do? If it's customised pattern 
> > sequencing/creation, it might be easier to go the other way round - 
> > throw out the microprocessor and the memory completely, and 
> replace it 
> > with trigger circuitry of your own design, keeping the 
> audio circuitry 
> > intact.

But if you change the microprocessor, and develop a more advanced UI, you
lose the character of the instrument.

Cheers,
Colin f





More information about the Synth-diy mailing list