correct volume contour for THAT 2162 VCA?

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

dfiction

Member
Joined
Feb 8, 2016
Messages
14
Hey everyone!

I found this wonderful forum and I've been lurking for a year or so. I have a question about a VCA circuit I'm designing, and I think that I am far enough along to ask for help.

I'm doing some work with the THAT2162, and in my experiments and trials I've managed to get a very good sounding, quiet VCA design that I'm quite happy with. I've designed a circuit that controls the gain sent to 6 power amps with a single potentiometer, by using 6 different VCAs, implementing the THAT2162. I'm using a single 12v supply, and the whole vca section of the circuit is running in reference to vcc/2. I opted to use differential drive because I wanted to maintain a wide dynamic range in my circuit. This might have been overkill, but I ended up with a pretty great circuit, using some rather expensive semiconductors.

While I have a great sounding circuit with wide dynamic range and low noise, I am having a lot of trouble controlling the contour of the volume pot, which should not scale linearly, but rather with an approximately log curve, so that I can use the single potentiometer to control some "long fadeouts" with the volume knob. I did measure the gain at the Ec- port My circuit tracks linearly, scaling between -120 -> +6 dB at the full extremes of the volume pot (the extremes are configurable by two resistors in the control circuitry).

But the volume pot doesn't track correctly, the way a master fader VCA should, and I don't know how to modify my circuit accordingly. At first I kept increasing the resistor R5, which lowers the threshold of the "off" position of the volume knob P2, and I thought that the right approach would be to raise the minimum threshold to around around -70 db or so (R5 to 100k). But finally I have learned that I want a log response of the pot, which would enable me to keep my dynamic range, and give the circuit a "good response" when the user adjusts the volume control.

So, here's my circuit. I have already read the THATCORP design note 116, which details the diode breakpoint application, but I'm not sure how I should adapt that principle to my current circuit, for P2 is controlling a very small span of approx 300 mV. Where to interrupt, or stick a diode?

I don't really need a perfectly logarithmic response in this circuit. I just need the volume knob to be "useful" in the lower positions—right now the linear response I'm getting means that most of the action is in the top rotation of the pot, not very good for fades.

Any wisdom you might have would be much appreciated.

best,
Daniel

 

Attachments

  • VCA_ControlVoltageGen2.png
    VCA_ControlVoltageGen2.png
    51.3 KB · Views: 43
> here's my circuit.

The attached circuit shows a LME49724, not a THAT2162.

It also has no audio input. It has two pots with no clue which does what. You speak of 6 outputs but no sign of where anything splits six ways.

Ignoring that.... the THAT VCA is really based on dual-supply thinking. A single supply is sure possible but everything must reference to it. R5 appears to reference to "ground", not V/2. I think a dual-supply could be cleaner; and at 6 paths, not a big cost item.
 
The gain control on the THAT2162 is log already. If you want an even more dramatic control of the curve, I would just use the split curve diode circuit in DN116.

But if you're not convinced, then let me point out that your circuit has problems. One problem is that you're not using the full voltage range. You're using tiny voltages. That will result in inaccuracy and drift.

The second problem is that your control voltage is not low-passed so is going to be noisy (the noise will be small but even a tiny noise will affect the VCA) which will result in distortion in the audio. The number one absolutely most vitally important things about VCAs like these is that any noise on the control port will be output as distortion in the signal. You should use a simpler circuit like one of the ones from the THAT design notes. The circuit in DN116 has a 100n cap to low pass the output which your circuit does not have. Without low-passing at the control port, you will get noise which equals distortion in the audio.

Also, you definitely don't need the differential drive. You only need -500mV to 200mV on Ec+. Using diff drive is not going to have any impact on dynamic range.

Regarding your bias supply, the voltage divider cap is a little small (only good down to 30Hz) and the 10R buildout resistor is doing nothing (buildout resistors prevent the op amp from freaking if the circuit tries to source / sink too much current) because in this case you're only using it for small currents.
 
dfiction said:
I am having a lot of trouble controlling the contour of the volume pot, which should not scale linearly, but rather with an approximately log curve,
The actual control law is 6mV/dB, so it is perfectly logarithmic.


But the volume pot doesn't track correctly, the way a master fader VCA should,
That's because a "master VCA" is not perfectly log, as can be seen in its scale, that expands the +/-10 range. An actual audio potentiometer is an hybrid between a true log and linear.


So, here's my circuit. I have already read the THATCORP design note 116, which details the diode breakpoint application, but I'm not sure how I should adapt that principle to my current circuit, for P2 is controlling a very small span of approx 300 mV.
I see that in fact it's less than 100mV.

Where to interrupt, or stick a diode?
You would need to do that at a point where the voltage is higher, and attenuate after.
I think the diode thing is too complicated to set up and adjust.
I have a better solution. That's called "law-steering". Loading a potentiometer with a resistor between wiper and bottom makes it behave almost logarithmically (actually hyperbolic); the ratio between the pot's nominal value and the resistor determines how "bent" is the taper. Conversely, putting the resistor between wiper and top makes it behave reverse-log.
I suggest you do that.
But I also suggest you increase P1 and decrease R4, so you control a few volts insted of some millivolts. You set the attenuation later in the chain, e.g. around the LME48724, by increasing R8 and R10.
 
Thank you all for your prompt and generous suggestions! I'll start with law-steering, and the latest comment by abbey road d enfer.

I have a better solution. That's called "law-steering". Loading a potentiometer with a resistor between wiper and bottom makes it behave almost logarithmically (actually hyperbolic); the ratio between the pot's nominal value and the resistor determines how "bent" is the taper. Conversely, putting the resistor between wiper and top makes it behave reverse-log.
I suggest you do that.
But I also suggest you increase P1 and decrease R4, so you control a few volts insted of some millivolts. You set the attenuation later in the chain, e.g. around the LME48724, by increasing R8 and R10.

The volume control in my circuit is P2, which is connected to the output of an op amp stage, as well as R5, the "minimum attention" setting resistor.

I tried putting in the law steering resistor, getting this advice from multiple people here and as well on another forum. I read this sheet for more into on it: http://www.proaudiodesignforum.com/images/pdf/Resistors_Provide_Nonlinear_Pot_Tapers_Mark_Rumreich_EDN_November_13_1986.pdf
Seemed like the right idea based on how many people have been saying it!

But trying it, I haven't has as much luck. P2, which is the volume control, doesn't connect to ground, but rather R5. The link I posted shows the voltage divider version connecting to ground. When I connect a resistor between the wiper and the R5 junction, it makes no observable change. I started with a 1k resistor, tried 2k, 4.7k, and 330 ohms. No difference.
Well, connecting between the wiper lug and the other pin of the potentiometer does make a change—but it reduces the swing of P2 to be basically 0v and -6 db—and then the volume never shuts off.
 
dfiction said:
But trying it, I haven't has as much luck. P2, which is the volume control, doesn't connect to ground, but rather R5.
You really need to revise your plan. It's not a good idea to use a resistor at the bottom of a pot, because the tolerance on potentiometers is quite high. You should sum the voltage from the pot's wiper with a fixed volkatge via resistors. The fixed part will define the lower limit and the pot will define the range.
 
abbey road d enfer said:
You really need to revise your plan. It's not a good idea to use a resistor at the bottom of a pot, because the tolerance on potentiometers is quite high. You should sum the voltage from the pot's wiper with a fixed volkatge via resistors. The fixed part will define the lower limit and the pot will define the range.

This makes a lot of sense to me—I'm using an alps pot, but it is not particular precise, so I get what you mean about the tolerance. I'd like to try this approach, but i'm having trouble visualizing exactly what you mean. Is there any way you could draw the schematic you're suggesting?
 
Thank you for the drawing, and your prompt reply!

However, I'm not sure exactly where this drops in the schematic, because there isn't a 5v input anywhere in my circuit. Can you clarify which part of the circuit you're suggesting I modify? Do you mean to start at P2, or P1? P1 just defines a small +6db gain boost so the top end of the volume port VCA can be configured for a slight increase.

But the volume adjustment is P2, which varies between Ov and -200mv or so. anyway, I marked up your drawing with values so i might better understand it.
 

Attachments

  • VCA pot1.jpg
    VCA pot1.jpg
    15.9 KB · Views: 18
dfiction said:
However, I'm not sure exactly where this drops in the schematic, because there isn't a 5v input anywhere in my circuit.
5V is just an example of a reference voltage. It could be your 12(11?) volt rail.

Can you clarify which part of the circuit you're suggesting I modify?
All the circuit.

Do you mean to start at P2, or P1?
You start with "my" circuit, fed with either your supply rail, or a regulated voltage if the supply is not regulated. Work out the resistor values for correct range.

P1 just defines a small +6db gain boost so the top end of the volume port VCA can be configured for a slight increase.
Make the circuit work, then add a resistor to the summing node, fed by another potentiometer. Scale the resistor for 6dB increase with the pot fully CW.
 

Latest posts

Back
Top