Using sample-and-hold for 100% analog tap-tempo?

[bioroids]

Be prepared to test several Cd4066 (and CD4016), because in pwm applications, they tend to differ a lot in the amount of noise and modulation range. Use sockets!

Miguel

[Eb7+9]

turns out to be not such a difficult problem to solve if you step away from the same old control methods ... a no-brainer really, the other issue is about keeping your sample voltages steady for a long time (minutes) so the rate doesn't drift - which is also easily solved using regen-hold circuitry ... it's been done, works great, doesn't use many parts at all - a fun challenge any capable circuit designer can solve ... of course the main drawback to the whole idea is you won't be able to do in-time triggering for super fast LFO rates unless you scale the sampleTimeDifference-to-speed conversion, again easily doable ...

[R.G.]

turns out to be not such a difficult problem to solve if you step away from the same old control methods ... a no-brainer really, the other issue is about keeping your sample voltages steady for a long time (minutes) so the rate doesn't drift - which is also easily solved using regen-hold circuitry ... it's been done, works great, doesn't use many parts at all - a fun challenge any capable circuit designer can solve ... of course the main drawback to the whole idea is you won't be able to do in-time triggering for super fast LFO rates unless you scale the sampleTimeDifference-to-speed conversion, again easily doable ...

Kewl! Then you won't mind posting for us to see the pre-existing ("it's been done"), functional "works great", economical ("doesn't use many parts at all") version, presumably one you yourself ("any circuit designer") designed and have working, will you? Since this is after all a simple ("not such a problem", "a no-brainer") little thing, I'm sure you have one in your hip pocket, right? So post it.

[Paul Perry (Frostwave)]

Yeah, it's like those math texts - just when it gets interesting, it is "left as an exercise for the reader"
Mind you, that is what helps make readers smarter.. .

[earthtonesaudio]

Over the weekend I got all but the PWM part working.  I'm not sure yet how to implement a voltage-controlled PWM using a Norton amp.  If all else fails, I could split up duties so the Norton amps are doing one thing, and a comparator or schmitt trigger is doing the rest, but I like the idea of doing it all with the Norton amps.

[R.G.]

I'm glad you're working through it. Don't let Eb's note trivialize what you're doing.

[earthtonesaudio]

Thanks R.G.

Eb7+9, I too would like to see the circuits you're talking about. 

[earthtonesaudio]

Okay, I scrapped the idea of the PWM/LED/photocell/whatever, in favor of an adjustable current sink.

Since pin 6 of the pt2399 is a VCO, and the voltage on pin 6 stays at about 2.5V regardless of the resistance value, it must be controlled by the current.  There are some appnotes in AN-72 for voltage-variable current sources/sinks, so that ought to do it.  Once I figure out the max/min currents required by the pt2399, I'll be able to properly scale the current sink, and then the speed of the ramp voltage increase.

Since the whole thing will be so dependent on power supply voltage, I'm thinking it would be a good idea to include a regulator also.

This page helped a lot: http://www.homebuilthardware.com/index.php/projects/pt239x-delay/
Fingers crossed...

[bioroids]

Don't ever think to get away with this without precise power supply regulation!

[slacker]

Once I figure out the max/min currents required by the pt2399, I'll be able to properly scale the current sink, and then the speed of the ramp voltage increase.

Be careful not to pull too much current, I killed a PT2399 experimenting with current sources whilst developing the Echo Base. The method I ended up with to modulate the PT2399 using a PNP transitor seems fairly linear or at least fairly predictable so maybe doing something based on that would save you some work.

[earthtonesaudio]

Thanks for the input Ian.  I'll have to check that out.

Edit: checked it out. 

Huh?  That PNP is hooked up backwards!?  I guess I don't understand what's going on there.

[slacker]

Huh?  That PNP is hooked up backwards!?  I guess I don't understand what's going on there.

I'll be completely honest, I don't either. I just experimented and it worked better that way round
It works better that way round in my simulator as well so it's obviously not a fluke.
It's essentially working as a voltage controlled resistor, the resistance is proportional to the voltage on the base.

[earthtonesaudio]

Well in any event, do you happen to know which way the proportion works? 

Meaning, if the voltage at the base goes higher, does the collector-emitter resistance decrease or increase? 

[slacker]

Resistance increases as base voltage increases. If you need the opposite response you can use an NPN transistor instead.

[earthtonesaudio]

Okay that makes sense.

That might just work for me.  From what I've read the max current you should put through pin 6 is 3ma or so.  Shouldn't be too hard to set the max/min currents, with some resistors.

That only leaves the truly hard part, translating the hold time into the proper voltage, then into the proper current.

[earthtonesaudio]

Actually, I wonder if I could just ramp from ~zero to 2.5V, and rely on the amplifier to sink/source the proper amount of current as needed...?
It occurs to me that the PNP transistor on the echo base might just be acting as a pair of diodes, and the oscillator op-amp is sinking the current from pin 6 as it modulates.  Maybe.

[slacker]

Yeah that will work fine. By the time the LFO signal has gone through the voltage divider to the base of the transistor it's sweeping between about 1 and 2.5 volts at maximum depth.

You could be right about how the transistor is working, hopefully someone who knows their stuff will fill us in.

[Eb7+9]

Okay, I scrapped the idea of the PWM/LED/photocell/whatever, in favor of an adjustable current sink.

you're on the right track, current control is the way to go IMO ... linear conversion to current of your time reference via sample/hold is easy  - there are several options ... from there I just use your garden variety linear control topology (app. notes) - then my sine shapper, again several possibilities ... for triggers and shunts I use a modified Boss astable switcher circuit to do it all from one physical push switch ... I been mucking with touch-plate trigger also but haven't gotten around to finishing that part yet ... I'd like to see what you guys come up with - don't give up, it's doable ...

[R.G.]

you're on the right track, current control is the way to go IMO ... linear conversion to current of your time reference via sample/hold is easy  - there are several options ... from there I just use your garden variety linear control topology (app. notes) - then my sine shapper, again several possibilities ... for triggers and shunts I use a modified Boss astable switcher circuit to do it all from one physical push switch ... I been mucking with touch-plate trigger also but haven't gotten around to finishing that part yet ... I'd like to see what you guys come up with - don't give up, it's doable ...

turns out to be not such a difficult problem to solve if you step away from the same old control methods ... a no-brainer really, the other issue is about keeping your sample voltages steady for a long time (minutes) so the rate doesn't drift - which is also easily solved using regen-hold circuitry ... it's been done, works great, doesn't use many parts at all - a fun challenge any capable circuit designer can solve ... of course the main drawback to the whole idea is you won't be able to do in-time triggering for super fast LFO rates unless you scale the sampleTimeDifference-to-speed conversion, again easily doable ...

Kewl! Then you won't mind posting for us to see the pre-existing ("it's been done"), functional "works great", economical ("doesn't use many parts at all") version, presumably one you yourself ("any circuit designer") designed and have working, will you? Since this is after all a simple ("not such a problem", "a no-brainer") little thing, I'm sure you have one in your hip pocket, right? So post it.