MIDI-CV converters (was: Building my first synth...)

gstopp at fibermux.com gstopp at fibermux.com
Mon Feb 5 18:07:53 CET 1996

     The thing that is most intimidating about a home-brew MIDI-to-CV 
     converter is the fact that you need a development system for uP 
     designs. This usually consists of a PC with a PROM programmer 
     attached to it. We got loads of this stuff here at work (top of the 
     line, I might add) but I can't risk doing non-work-related stuff 
     here - it's bad enough I'm on these mail lists!
     I think that such a project is right on the border of being 
     practically implemented in hardware without using a uP. A basic 
     hardware-only unit that would get the job done wouldn't be too hard, 
     but the kicker is that if you use compiled code written on a 
     development system, you can have so much more functionality for free - 
     as long as you have to get the basic thing running, all you have to do 
     is type some more to add features.
     Not only that, if you want the device to function flawlessly in 
     complex MIDI data stream environments, there's quite a bit of 
     algorithm stuff that should be added to keep things from getting 
     confused. For example, there's the two kinds of note-offs at &H8x and 
     &H9x, running status handling, and real-time messages. I've written 
     data deciphering programs for MIDI that can handle everything, so I 
     know the algorithms required, and if you go with the just-hardware 
     approach it would be quite hard to implement these completely.
     I just had an idea - I have an old XT laying around, with an original 
     MPU-401 in it. If I made a D/A circuit board, and it was controlled by 
     the PC hardware, it should be possible to develop the code on the PC 
     and run it to make a very high-powered MIDI-to-CV controller. The D/A 
     could be an 8- or 16-bit discrete R-2R ladder, driving an big analog 
     mux into a bunch of S & H's, with some gate latches as well. It would 
     probably be easy to go to sixteen or thirty-two analog voltages. The 
     software could do all of the demux scanning probably, reducing the 
     hardware requirements. There could be a graphic user interface for 
     setting up the channels etc. This all could probably be done in 
     something easy like Visual Basic or Visual C....
     Any thoughts on this? How should it be connected to the PC? If it's 
     built on a card that fits into a slot in the PC, there may be 
     connector density problems as well as EMI pickup from nearby spikey 
     digital stuff. Maybe an external box, attached through the parallel 
     port? How about using the serial port? Now it has a UART in it, but 
     now it only has to decipher D/A setup data instead of MIDI traffic.
     Whaddaya think guys.
     - Gene
     gstopp at fibermux.com
______________________________ Reply Separator _________________________________
Subject: Re: building my first synth
Author:  "J.D. McEachin" <jdm at synthcom.com> at ccrelayout 
Date:    2/5/96 8:57 AM
On Sun, 4 Feb 1996, Jonathan Mayer wrote:
> I seem to recall seeing a 8051-compatible uProc with a PLL UART built 
> into it that would be ideal for a MIDI interface.  Slap a DAC onto one of 
> the i/o ports, and you've got a cv-box, right?
Most if not all variants of the 8051 (8031, 8052, etc.) have a built in 
UART that is ideal for MIDI.  You can lift the circuitry off the Jupiter6 
or the JX3P schematics...

More information about the Synth-diy mailing list