Voltage controlled gain

[barry]

Hi all,
I need an inexpensive voltage(or current) controlled gain stage. Basically, I need to control three op-amps with one pot, and a 3-gang pot is too expensive. I know I can use a JFET as a voltage-controlled resistor, but does anyone have any real-world experience with this approach? Since this will be going into a product, it needs to have repeatable performance. How much variation can be expected from batch-to-batch of JFETs? What other alternatives are there to the JFET approach? I've seen a number of VCAs out there, but they all seem to be geared toward high-frequency applications (and the associated price tag). I need something to work in the audio range; I'm not TOO concerned about noise or linearity. I need maybe a 5:1 or 10:1 gain control.

Any thoughts would be appreciated.

Thanks,
Barry

 

[LvW]

Barry, what about Transconductance amps (OTA`s), which can be controlled via a current (Iabc) to be fed into a separate input node ?

 

[barry]

Hmmm, good thought. I'd forgotten about those; I didn't even know anyone still made them-the LM13700 is about 6000 years old. Any other devices you'd recommend?

 

[crutschow]

You can also use analog multipliers, which are very repeatable from unit to unit.

 

[LvW]

According to my knowledge:
LT1228, MAX435/436.

By the way - you can use the OTA as the main amplifier or as a grounded controllable resistor (in a classical opamp circuit).

Remark: The LT1228 has an output buffer onboard (CFA with a voltage output).

 

[barry]

The MAX435 is obsolete, as well as a lot of the others (CA3080, CA3280,SSM2164). Seems that manufacturers have obsoleted these parts with no alternatives. The LT1228 looks good, except it's a little pricey-about US$4 in quantity. (I'm talking about a small-market consumer product using about 5K units/year).

Multipliers would work but they're REALLY expensive.

The hunt continues...

 

[FvM]

FETs have large type variations, the only practical way to get along with a reasonable matching is to use dual FETs with one channel as reference, as discussed in this old edaboard thread. https://www.edaboard.com/threads/129234/

LvW already mentioned OTAs, LM13700 is an industry standard part with a linearity benefit over standard OTAs due to the linearizing diodes.

Analog multipliers are always an option, but not particularly cheap.

VCA have been used for studio automation in the analog age, e.g. various SSMxxxx chips, but have been completely superseeded by digital level control solutions, because they simply didn't meet the increasing SNR and THD requirements.

LDR analog couplers are another option.

I understand from your post that you are actually targetting to manual control. In this case, you should consider to attach multiple single channel potentiometers to a common shaft, e.g. a propular Piher type available from many catalog distributors http://www.piher-nacesa.com/pdf/14-PT15v03.pdf

I have previously used it in a tunable active crossover design.

 

[schmitt trigger]

How about digital potentiometers? Microchip and Analog Devices offer an extensive line of devices.

If you use three devices with an up_down/clock/select interface, those could be driven in parallel from the same control lines, you could achieve the excellent tracking you require.
Some devices keep their latest settings in memory, so it will remember their position even when powered down.


EDIT: I know that you requested for voltage controlled gain....but perhaps you could have an alternate control mechanism.

 

[barry]

This is a strictly analog board and digital pots would require generation of the digital signals/clocks; otherwise, I like those digipots. I like FvM's idea of using the individual pots connected to a common shaft. If these are inexpensive enough (and if Piher ever gets back to me), that would be the way to go, since the user interface will be a pot anyway. Depending on what my client wants, I might be able to get away with a 3-pole, 3-position switch that would just select an appropriate resistor for 3 different gain settings. (The other hard part of this project is getting my client to make up his mind.)

 

[schmitt trigger]

"(The other hard part of this project is getting my client to make up his mind.)"

Been there, done that.
And once that the design has started, that a customer comes back with: "By the way, could you add this simple feature?"

But I can't complain............. my daughter is an architect, and you wouldn't believe the "simple requests" she gets from customers, that require substantial structural modifications.

 

[FvM]

Piher PT15 is e.g. stocked by Mouser. I don't know if it's in the US stock, because the series is primarly popular in Europe. But it can be surely ordered through Mouser US.

 

[barry]

Piher PT15 is e.g. stocked by Mouser. I don't know if it's in the US stock, because the series is primarly popular in Europe. But it can be surely ordered through Mouser US.

I see the pots, but what about the shaft for connecting all three together? Do they (Piher) have shafts, or do I need to design my own? I don't think I want to go down that road.

 

[FvM]

There are PT15 types for different shaft forms, round with double flattening is the standard type, hexagonal is available too (a popular form for brass pre-product rods). To be honest, I don't remember where I got the shafts for my designs (quite a long time ago). Probably from a mechanical workshop around the corner. Usually you have some requirements, e.g. a specific control knob mount, possibly a bearing, a specific length. It's not a big thing to get it made, either for a prototype or series.

You might get useful shafts from Piher as well, but I never tried.