Top Octave Generator chips and dividers

[PRR]

> divider network that does 185/196.

Does anybody remember the Rate Multiplier?

Apparently TI still does.

Good old-old CD4089 is 4-bit, not enough. It's stackable, and at $0.18 for PDIP you can't whine about cost, but pin-wiring labor would be a chore.

SN7497 is 6-bit cascadable. Still not enough for the Hammond approximation.

[~arph]

Does anybody remember the Rate Multiplier?

Yes, I always thought that name was hugely misleading.. as it divides.

[amptramp]

The rate multiplier provides some oddball waveforms depending on the ratio in use.  It may or may not be usable.  But I do remember avoiding it.

[R.G.]

I was thinking about this, and what about doubling instead of dividing? Would that be simpler than using a divider chip? Looking at a triangle wave, if you full-wave rectify it, it will double the frequency without distorting it much. Setting an oscillator for a lower frequency and doubling it for each next octave seems like it might be simpler.

There's not much simpler in the electronics world than dividing a square wave by two. Analog or digital doubling are way worse by the time you actually get to the end result you need.

Twelfth-root-of-two is a very hard number to compute 11 times in 500 microseconds (1/2000th of a second, period of a high note)...

It's worse than that.

You don't have the period of a high note to compute it. You have zero time to compute it in some cases. The notes at different frequencies will have transitions that they must make on time to get the proper notes out. Those transition times will inevitably correspond to less than a clock tick at some point, and that is effectively zero time to a digital device.  That may or may not be audible, or you may be able to do some special casing, whatever.

This kind of thinking is what led me to decide that getting one close-enough frequency, and divided-down outputs out of a uC per note was probably the most effective.

PLDs and the hard-logic 50240 and its ilk are effectively parallel processors, one per note because the divider chains are independently being clocked down to the particular dividers. There may be some commonality at the first divider stages by factoring the divide-by chains and using as many common early stages as possible, but the divider chains at the end become independent, not locked as they must be in a single-thread computation.

[amptramp]

Many years ago, a scheme was suggested whereby there would be a divider chain that would use a broadband phase-locked loop with 185/196 dividers which would be augmented by a narrowband phase-UNlock loop that would push the VCO away from an exact multiple.  It is the slight deviation from accurate phase lock that provides an acoustic instrument sound in place of a Lowrey organ sound.  Of course, which direction to push off the logic divider frequency may be somewhat in doubt.

[PRR]

Total over-kill:

3.5 GSPS Direct Digital Synthesizer with 12-bit ADC
•    3.5 GSPS internal clock speed
•    Integrated 12-bit DAC
•    Frequency tuning resolution to 190 pHz
•    16-bit phase tuning resolution
•    12-bit amplitude scaling
•    32-bit parallel datapath interface
•    Programmable modulus
•    Automatic linear and nonlinear frequency sweeping capability AD9914 1K price $135.58
   Learn more about the AD9914 GSPS Direct Digital Synthesizer, download the datasheet and order samples here
   datasheet

*pico* Hz?? Boy, can't complain about pitch error!