[sdiy] Interesting article on top octave generators.

[brianw]

One disadvantage of a CPU or FPGA solution is that the circuit doesn't do anything until you program it.

A board with 12 counters would be hard-wired to do the job, and it would function as soon as you power it up - no programmer needed. If you're only making one of these, the programming step probably doesn't make the project take much longer. But if you're manufacturing thousands of TOG replacements, then that programming stage could be significant. It requires extra connectors. Then there's alway the chance that marginal programming or bit rot in the Flash memory causes the board to "forget" how to generate anything at all. Hard-wired logic gates just work.


I see the point of a different solution, though. Twelve counter chips take up a lot of space. Plenty of CPU chips have 8 counters with available output pins. I'm surprised that there are any chips with 12 counters, but perhaps these are chips with eight 32-bit counters that can be split into sixteen 16-bit counters. I'm still surprised that there are 12 pins for counter outputs.


One question I have was already asked by someone in the comments for the original article.

> 2) What’s the smallest FPGA or CPLD that these (presumably) 12 counters would fit in? (Always a fan of open source FPGA tools myself).

I tend to frown upon the way that an FPGA seems to be thrown at so many problems these days. They're not exactly power-efficient. But in this case an FPGA would be much smaller than a board full of counter chips.

I'm guessing that there are plenty of small, cheap CPLD or FPGA parts that have 12 output pins and enough logic blocks for 12 counters. Your idea of 61 individual outputs would necessarily require a larger part with more pins, and it would certainly change the schematic of a vintage organ.

Brian


On Dec 27, 2024, at 4:01 AM, The SynthiMuse wrote:
> Yea, using a processor seems like overkill but 'C' on a little dev board is so accessible.
> 
> I suppose a good choice might be one of those 'Blue pill'-style FPGA boards that would allow you to implement a logical clone of the original TOG chips.
> 
> A further advantage would be that if there were enough outputs, you might be able to implement a custom divider for every note and get closer to the integer numerator/denominator that Don mentioned was inherent on the tonewheel organs.
> 
> My last VHDL experience was 20 years ago so well beyond the 3.5-year 'half life of knowledge' I've heard about.
> However, I can imagine perhaps VHDL coding a single numerator/denominator module then instantiating it 12 times for a TOG or 61 times for a whole keyboard.
> I can imagine the board layout could be a challenge as you might have the background breakthrough of 60 other discordant notes for every key you pressed. Maybe the divider module block would need an enable line but that would double the number of pins needed.
> 
> Gerry