[sdiy] Tap tempo & pulse/clock ideas needed

[WeAreAs1 at aol.com]

In a message dated 3/21/04 1:54:52 PM, dunker at invalid.ed.ntnu.no writes:

<<  I am looking for a clock (trigger pulse) generator with a tap tempo
function. I imagine  that I would want this clock to be able to output
"musical" multiples of the tapped beat frequency for direct triggering of a
sequencer, ADSR events etc. I also have the external triggering of my Korg
Polysix's arpeggiator in mind here.
 How big a deal is it to build something like this? >>

These days, it would probably be easiest to build a tap-tempo clock generator 
using some kind of microprocessor and some code.  However, it certainly would 
be possible to build something from discrete logic circuits.  The old Hammond 
Concorde organs had a tap-tempo control on their clunky internal drum 
machine.  This instrument came out around 1977 or so.  The Hammond tap-tempo circuit 
and concept were patented at the time, and you can see the patent at the big 
online U.S. Patent database (it doesn't show the exact circuit, but it explains 
how it's done).  Sorry, I don't have a link for it -- try searching the 
patent database for the Keyword "tempo", and the Patent holder as "Hammond" (or 
"CBS"?).

Here's a rough block diagram of how such a thing could be achieved with 
simple logic counters and gates (similar to the Hammond circuit):

(view this with a proprtional font)

Tap tempo sensor--->F/F----|
                           V
Fast fixed rate clock--> Gate--->  Binary Counter A with 4-bit or greater 
output
                       |                             |
                       |--->  Divide by N --->  Presettable Binary counter B

The binary coded output of Counter A is used to set the preset inputs of 
Counter B.  You tap the sensor once, and the F/F (flip-flop) tells the Gate to 
open, allowing Counter A to start counting up.  You tap the sensor again, and the 
F/F changes state, telling the gate to close, and stopping the count.  The 
fast fixed-rate clock drives the clock inpout of both counters, but it also goes 
through a divide-by-N circuit before hitting the clock input of Counter B.  
Counter B outputs the actual clock that you use to drive your noisemakers 
(you'd use the final bit of its up-counting sequence as your tempo clock).  The 
actual value of N would be determined by the frequency of the fast fixed clock 
and by how much resolution your binary counters allow.  (faster clock and bigger 
counters = better tempo resolution.  You do the math - I'm too lazy, and 
unless I really try, too dumb.  Obviously, this is just a bare-bones 
simplification of what would be needed to make it wourk -- there would necessarily be a lot 
of miscellaneous logic needed to glue the sections together properly.  
Unfortunately, to really make this work well, it would wind up being a fairly 
complex and unwieldy kludge.  Much cleaner to do it with code and a PIC, or 
something like that.  The code could be basically a software version of the above 
circuit.

Michael Bacich