0

I'm trying to achieve automatic gain control in an audio circuit, which requires me to use both diode pins of each amplifier available on the LM13700 chip. The diode pins accept a current (Idiode), which is produced from the rectified and buffered output of the first (top) OTA in the circuit below. I'm concerned that feeding this output to both pins will cause a discrepancy in the current received by both OTAs, simply due to the nature of Kirchhoff's current law and uneven resistances in the connections.

I initially thought that a current buffer/follower would suffice (as seen in the circuit) but now I'm thinking that a current divider (with equal resistances: 1/2x) combined with a current amplifier (2x) is a more reliable solution, but this seems redundant. How would you achieve the simultaneous biasing?

enter image description here

meekeech
  • 1

1 Answers1

1

I would try to use a current mirror with two outputs:

schematic

simulate this circuit – Schematic created using CircuitLab

I'm assuming that the Lowpass filter outputs a voltage. That voltage needs to be converted into a current, this is done by R2. Note that you might need a voltage buffer at the output of your lowpass filter.

The current through R2 is pulled from the input of the current mirror. As long as you make R3a, R3b and R3c the same value and the 3 PNP transistors are also the same type, then the currents sourced by Q2 and Q3 will be the same as the current through R2.

The value of R3a, R3b and R3c should be such that you get at least 200 mV across them, then the matching of the currents should be good enough.

I give you kudos for using the (ancient) LM13700 as a volume control. That was about the only way to do it when I was a learning about electronics. These days we have digitally controlled volume adjustment ICs and that would be much easier to use. Personally I would be lazy and use such an IC instead of an analog LM13700 based solution. I would not worry about the stepped gain control of such a "digital potmeter" as the steps can be as small as 1 dB which is too small to notice I think.

Bimpelrekkie
  • 80,139
  • 2
  • 93
  • 183
  • Thank you for your detailed reply Bimpelrekkie. There is indeed a series resistor in the low pass filter which allows the current flow. A voltage buffer would modify/increase the current, which is in the range of 100's of microamps, as is desirable for the OTA diode input. As far as your current mirror solution: that looks promising but I'm not clear on whether that would allow a variable current, which is dependent on the OTA's input voltage. I might be missing something but to me that seems like it would have a constant current. – meekeech Sep 21 '21 at 17:10
  • In terms of the LM13700, I'm trying to replicate automatic gain control described in a patent from the 70s, hence the ancient tech. I initially wanted to use a digitally controlled amplifier but I haven't been able to find one that support gain steps with the desired precision of around ~0.1dB. – meekeech Sep 21 '21 at 17:12
  • *I'm not clear on whether that would allow a variable current, which is dependent on the OTA's input voltage* Not sure what you mean, the current mirror copies the current. If the current varies then the output current will vary as well. Looking at your drawing again, I notice that you're shorting the + and - inputs of the OTA feeding a signal to the power amplifier. That is not how you use an OTA. This way, no signal will get through. – Bimpelrekkie Sep 21 '21 at 19:54
  • *support gain steps with the desired precision of around ~0.1dB* Why do you think you need 0.1 dB steps ??? I claim that 1 dB steps are enough but feel free to show evidence that I'm wrong an that 1 dB is not enough. I could also ask why not have 0.01 dB steps ?? It is "easy" to write down a number that must be met. It can sometimes be **VERY HARD** to come up with the number that is actually needed. – Bimpelrekkie Sep 21 '21 at 19:56
  • My mistake then, I'm sure your setup is fine in that sense. I'll review my knowledge of current mirrors and run some simulations before making any further assertions. Also this diagram is very high level, so my apologies on the confusion, I'm definitely not shorting the + and - in real life. – meekeech Sep 21 '21 at 21:01
  • Also, my application is not related to music but rather to the calculation of acoustic immittance, which requires the fine adjustment of volume throughout the procedure. My calculations at the microphone side have shown changes as low as 0.1dB between two measurements, so I'm striving to adjust the speaker volume by the same amount to maintain constant sound pressure within the ear canal. – meekeech Sep 21 '21 at 21:01