VCA

hi everyone,

I am now building the VCA circuit, 

I would like to ask it is possible to build the VCA with LM386 or LMH6505?

or I should use the typical IC LM13700?

please help. 

thank you so much.

Not.

The gain is set by external resistor, not voltage. Moreover the range of control is too narrow for  such application (G=20...200).

Ilya is correct - You need some form of voltage or current gain control to implement a VCA.

Simplest is probably an opto-coupled resistor or Fet.

A H11F1 Fet opto-isolator is, IMO, one of the best / simplest ways to implement a VCA.. More current through its LED, the lower its Fets impedence - Down to about 100 ohms with 8mA through the LED, about 10 Meg ohms with no current through the LED. The only thing to be careful about is that the voltage across the FET is kept small, otherwise there is massive distortion (it becomes extremely non-linear) - Over drive it a tiny bit for nice tube type distortion.

See "Isolated Variable Attenuators" in the data sheet applications.

You can use these passive circuits with whatever amplifier you wish.

Fred.

TDA 7052A (out of production but still good availability). The advantage is a logarithmic volume characteristic with a linear control voltage given.

If one wants a really good, available (in production) Log CCA, the THAT 2181 is by far the best I have found.. Particularly if you dont want SMD.

Many VCA's are in fact CCA's (Current controlled amplifiers rather than voltage controlled amplifiers) - A CCA turns into a VCA by placing a resistor in series with the current input (The LM13700 OTA operates like this)

One of the reasons I suggested the H11F1 is simplicity - Someone who does not understand the basic operation of a VCA and asks about using the LM386 or LMH6505 is probably (IMO) better going for the simplest "passive" Current controlled attenuator they can find - Provided you dont put excessive current through the LED or FET, the H11F1 is really hardy, and with 4 operational pins (two unused pins) one can get a VCA up and running in minutes - You can fiddle about with resistors and have almost none of the potential problems you get if you mess about with say a LM13700 when you dont know what you are doing.

Apart from which, these parts doo make damn good CCA's, particularly for musical instruments where sometimes controlled distortion is a bonus.. For applications like mixing consoles, one would probably never use a H11F1 (but one would never use a LM13700 in a high-end console), and I use the THATS parts now - The SSM VCA's I used to use are no longer in production, sadly.

Fred.

Fred, your recent switched capacitor thing could (after demodulation of the signal) also act as VCA when controlled with a variable pulse width.

Edit: The LM13700 is not as bad as its reputation. During my recent Theremin Cello development, I discovered that one may obtain very good results when biasing it correctly through both input diodes and when NOT using the darlington transistor pair at the output but feeding the current from pin 5 directly into an inverting opamp (I used the TL074) as a current-voltage converter.

"The LM13700 is not as bad as its reputation. During my recent Theremin Cello development, I discovered that one may obtain very good results when biasing it correctly through both input diodes and when NOT using the darlington transistor pair at the output but feeding the current from pin 5 directly into an inverting opamp (I used the TL074) as a current-voltage converter." - Thierry

Yes, agreed - BUT..  You need to know what you are doing, and for hobbyists, getting the LM13700 to behave well requires understanding most dont have.. Also, its a linear VCA / CCA, If you want (as you recomended) a log VCA, then the THATS part is far simpler and probably cheaper if you factor in the log converter.

IMO, the problem with a combined power amplifier and VCA (like the TDA7052A) is that one does not have a way to get a clean line-level output.. Using a seperate VCA or buffered H11F1 before the power amp means that you can have both - A line signal without the power amps nasties, and a power amp to drive a 1W internal speaker if you really want this. The cost of the extra parts is nil or trivial.

"Fred, your recent switched capacitor thing could (after demodulation of the signal) also act as VCA when controlled with a variable pulse width." - Thierry.

The circuit I presented was NOT a "switched capacitor thing" "(as in, in terms of a switched capacitor filter which could be modified to a VCA) and is not suitable for use as a VCA - I clearly showed this, and argued this, And YOU AGREED that it was not a "switched capacitor thing"(as in a switched capacitor filter) !  - Yes, my circuit does swap capacitors, but all it is, is a highly efficient SAMPLE AND HOLD! - its not a PWM controlled multiplier or attenuator or filter! -

Changing the PW into the (Ref) clock does not change anything - Provided the REF input is longer than the CLK inputs required duration, and at the correct logic level, ANY PW into it can be applied without any change to the circuits function - Sampling occurs at the rising edge of this input signal only!

Each rising edge of the input (Reference oscillator) clock is converted to a change of state from the divider following this, and sampling occurs at every change of this dividers output state. Equal (50:50) spacing between these changes of state is essential for the circuit to work correctly.

If PW into the SAH circuit is not 50:50 sampling occurs wrongly (one is effectively sampling with irregular intervals)  it unbalances the charge/hold cycles on the SAH capacitors and implements probably some form of SinC function if pushed to extremes where it does behave a bit like some horrible multiplier.. But the divider following the REF Clk is there to ensure that the above does not happen! I was designing a sampling system NOT A MULTIPLIER!

When balanced, with  50:50 SAH switching cycle, it behaves like a sampler! - and this is how it was designed to operate.. the whole purpose of swapping the capacitors is to implement sampling and avoid the problems that narrow (or any non 50% PW) pulses bring with them when operating with a SAH required to sample rapidly changing signals...

My SAH samples for an entire cycle, and holds for an entire cycle, and does both of these operations simultaneously - the "sampling pulse" does not actually exist! All that exists is the transition time as the capacitors are swapped.. Go and look at the waveforms in the last posting on this page!

Now, take your Henk mixer, and PWM a pulse into the gate of that, and feed AF into it, and you get what you are suggesting - But why bother? There is no advantage in this approach - You need to turn the CV into a PWM, you need to demodulate the output side - you have potential problems of aliasing between any RF on the audio input and your PWM.. May just as well bias the fet as a voltage controlled resistor and get the CV to vary its resistance - But its a lot easier to use a H11F1 which effectively does this without the hassle of biasing the fet.

Anyway, switched capacitor circuits (or switched resistors followed by a filter) are complex and, quite frankly, horrible unless great care is taken in their design - Even using dedicated switching ICs, use of switching filters or attenuators for audio is not something I would advise - To me, they usually sound horrible!

Give me a simple CCA any day - A cheap LM13700 is likely to be much better suited and simpler than a switched gain / attenuation circuit for the above application, and, IMO, the THATS VCAs are worth the extra money for high-end audio or when log law is required, and the H11F1 is great for musical instruments where optional addition of musical distortion is not taboo . (I have actually fitted them in some synth retrofits, just to get their musical distortion)

Fred.