[sdiy] Crumar Bit-1 DCO question

[Tom Wiltshire]

Hi All,

I've been having a look at the Crumar Bit 1 DCO schematics. It's quite an interesting design. The Bit99/Bit01 synths are an update of this.

	https://archive.org/details/synthmanual-crumar-bit-1-schematics-late-revision 

The gist of it is that there's 8253 counters generating a clock at 16 x the note frequency. This clock then drives a 4520 4-bit counter. The clock itself and the four counter outputs are added together with weighted resistors to make a staircase approximation to a ramp wave. Since the highest square wave is at 16 x the note frequency, the first missing harmonic is the 32nd at -30dB, so it's not too bad as a quick approximation to a ramp. The ramp is fed to a comparator to generate a PWM output in the usual way, and there's a single-transistor fullwave rectifier to create a "triangle" waveform too. It's supposed to be another Mario Maggi design, and it does seem to have his quirky fingerprints.

However, there's one thing I don't understand. The 8253 counters are clocked at 2MHz. The output is 16 x the note frequency, so a highish C (MIDI Note 84) needs a frequency of 16.744 KHz. For that frequency, the counter is using a division value of 119, but the previous note 83 uses 127, so there are only 8 frequencies per semitone - clearly audible. From there on up, the situation gets even worse.

Basically, I can't see how there's enough frequency resolution, especially for any sort of modulation (and LFO pitch modulation is possible).

Any ideas what's going on?

Thanks,
Tom

==================
       Electric Druid
Synth & Stompbox DIY
==================