Author |
Message |
mediatechnology
Joined: May 10, 2006 Posts: 80 Location: Oak Cliff, Texas USA
|
Posted: Wed Jul 09, 2008 6:48 pm Post subject:
VCA and Op Amps Form Simple VCO/Function Generator |
|
|
I've been toying with the idea of using THAT2180 and THAT2181 series VCAs as exponentially-controlled current sources and sinks. By placing a reference current into the audio input of a THAT VCA and controlling Ec, one can produce a current output that is exponentially related to Ec. This has a lot of potential uses.
One of the first things I wanted to try was a simple VCO "function generator" capable of producing triangle and square waves. The following is not a precision instrument but only a crude proof-of-concept. By inserting a VCA between a comparator and integrator the operating frequency can be varied over a fairly wide range at low cost. Accurate scaling, temperature compensation and precision were not initial objectives.
U1B and the 2181C VCA form a voltage-controlled integrator. U1A, configured as a comparator with positive feedback hysteresis, detects the zero crossings of the integrator output. The output polarity of U1A determines the polarity of the reference current to the VCA and whether or not it sources, or sinks, current into the virtual ground input of U1B. This causes the feedback capacitor to either charge or discharge. As the output of U1B crosses zero, the polarity of U1A's output reverses. In this configuration the THAT VCA inverts polarity. (The VCA in this circuit has no output I-V converter op amp to re-invert the internal OTA feeding the log/antilog gain cell.) Due to the inversion, the inputs of U1A are reversed compared to typical two-op amp circuits where the control element is a resistor.
With the values shown the hysteresis band around U1A is very narrow. Lowering the 1M to 100K produces a larger triangle wave output and lowers the maximum operating frequency. I would have changed this on the drawing but the frequency measurements shifted.
Due to the linear dB properties of the THAT VCA, the voltage at input Ec has an exponential relationship to the output current. The sign of the output current is determined by the sign of the input current.
As mentioned this is a proof-of-concept and it works quite well. There are obvious improvements to be made in the comparator, reference stability (currently set by U1A's clip points), scaling accuracy etc. But, it works. I think it might, with suitable improvements and scaling, make a good LFO. Last edited by mediatechnology on Sun Feb 12, 2012 11:38 am; edited 1 time in total |
|
Back to top
|
|
|
synthmonger
Joined: Nov 16, 2006 Posts: 578 Location: flada
Audio files: 3
|
Posted: Thu Jul 10, 2008 6:04 am Post subject:
|
|
|
Neat! I wonder what would be considered the simplest vco...and one with expo control. |
|
Back to top
|
|
|
mediatechnology
Joined: May 10, 2006 Posts: 80 Location: Oak Cliff, Texas USA
|
Posted: Thu Jul 10, 2008 6:26 am Post subject:
|
|
|
Thanks synthmonger.
I've got another one in mind using a CMOS 555 and THAT VCA. It might be simpler. No triangle wave though.
I forgot to mention that the previous circuit would also need PTAT compensation. Looks like another job for a PT-146. |
|
Back to top
|
|
|
frijitz
Joined: May 04, 2007 Posts: 1734 Location: NM USA
Audio files: 54
|
Posted: Thu Jul 10, 2008 7:17 am Post subject:
Re: VCA and Op Amps Form Simple VCO/Function Generator |
|
|
mediatechnology wrote: | I've been toying with the idea of using THAT2180 and THAT2181 series VCAs as exponentially-controlled current sources and sinks. |
Similar to what Roman Sowa did, right?
http://www.sowa.synth.net/modular/vco4m.gif
Ian |
|
Back to top
|
|
|
mediatechnology
Joined: May 10, 2006 Posts: 80 Location: Oak Cliff, Texas USA
|
Posted: Thu Jul 10, 2008 7:26 am Post subject:
|
|
|
Thanks frijitz. I hadn't seen that one but yes the same concept only his is about 9 years earlier.
Topologically they are different but the current source generation principle is the same.
That one uses a resettable integrator with sawtooth the generated waveform and the reference polarity positive. The VCA sets the slope of the charging ramp. There doesn't appear to be any reason why a THAT VCA couldn't also be used there as well and his circuit is far more refined.
In the circuit I posted, the VCA sets both the charge and discharge ramps with the native waveform being triangle.
One of the interesting things about using VCAs as current sources is that the audio input, in this case the reference, can be of either polarity. That opens some doors for both the THAT2180/2181 and ADI SSM2164 parts.
Two VCAs could be used with one setting the charge rate and a second the discharge. |
|
Back to top
|
|
|
frijitz
Joined: May 04, 2007 Posts: 1734 Location: NM USA
Audio files: 54
|
Posted: Thu Jul 10, 2008 8:12 am Post subject:
|
|
|
mediatechnology wrote: | Thanks frijitz. I hadn't seen that one but yes the same concept only his is about 9 years earlier.
Topologically they are different but the current source generation principle is the same.
That one uses a resettable integrator with sawtooth the generated waveform and the reference polarity positive. The VCA sets the slope of the charging ramp. There doesn't appear to be any reason why a THAT VCA couldn't also be used there as well and his circuit is far more refined.
In the circuit I posted, the VCA sets both the charge and discharge ramps with the native waveform being triangle.
One of the interesting things about using VCAs as current sources is that the audio input, in this case the reference, can be of either polarity. That opens some doors for both the THAT2180/2181 and ADI SSM2164 parts.
Two VCAs could be used with one setting the charge rate and a second the discharge. |
Right. Roman has done some very nice and innovative work, although he hasn't been active for a few years. It's worth looking at all his circuits, as he worked quite hard at finding apps for the SSM chip. I always thought it was a shame that more folks didn't follow up on his ideas.
I like the idea of using a tri waveform instead of a saw. It has several advantages -- for example, leakage errors cancel to first order, since one side ramps faster and the other slower.
It would be most interesting to see what kind of tracking accuracy you could get with this approach. Please keep us posted if you pursue this.
Ian |
|
Back to top
|
|
|
mediatechnology
Joined: May 10, 2006 Posts: 80 Location: Oak Cliff, Texas USA
|
Posted: Thu Jul 10, 2008 8:26 am Post subject:
|
|
|
Ian - thanks again for your reply.
If the Ec vs. gain (actually attenuation) conformity of the THAT VCAs is any indication, the tracking ought to be excellent once PTAT, a better DC-performing op amp, and comparator with controlled clipping amplitude are used. The circuit I posted doesn't do a lot of things e.g. it has no sync. Roman's work would be good to build upon. Anyone who wants to run with this is certainly welcome to and I can keep doing basic blocks.
I'm also looking at National's application note AN-263.
http://www.national.com/an/AN/AN-263.pdf
It's got some nice sine shaping circuits.
A polarity-flipper driven by a comparator based on the concept here is also worth exploring though I think I would flip polarity without the FETs:
http://www.national.com/an/AN/AN-263.pdf#page=10
Altering each polarity at the reference input would allow duty-factor control. The VCA input, like Roman's, is also a good place to apply linear FM. |
|
Back to top
|
|
|
|