Voltage-controlled amplifier

[elki]

Hi all! I would like to get an advise on the following circuit (attached). The idea is to control the amplitude of the input pulse from 0 to the supply rail voltage with the help of a voltage-controlled amplifier. Is it possible to achieve with such circuit? I guess I have some fundamental issue in how I approach it.

[xvr]

If you need only 2 gain values than this schema will work. If you expect gain controlled by voltage value - it will not. MOSFET is not a linear element.
Take a Variable Gain Amplifier in form of IC - https://www.digikey.ie/en/products/detail/analog-devices-inc/AD8337BCPZ-REEL/1218098 for example

[jonpaul]

pulse Vlo, Vhi, risetime, PRF?

j

[elki]

0-5V, 5ns, 100 Hz

[Terry Bites]

Simply use an analog switch and set the output voltage to a reference level of your choice. Vtop in the attached.
eg a TC4S66F or similar spdt.
If Ron is an issue a buffer amp can be added to the output.
You'll lose some risetime but I doubt that will be a problem at 100Hz.

[elki]

Thanks. If I understand correctly, this would change the reference of the output signal but not the amplitude wrt zero?

Simply use an analog switch and set the output voltage with a reference level of your choice. Vtop in the attached.

[Terry Bites]

The switch pulls the output to zero when the input is zero and to Vtop when the input is positive.
If Vin is too low for the switch to operate then add a comparator.

[elki]

Thanks for your suggestion. I am trying with DG419 running on single supply. I guess VL should be connected to GND?

The switch pulls the output to zero when the input is zero and to Vtop when the input is positive.
If Vin is too low for the switch to operate then add a comparator.

[Terry Bites]

Vl set the thresholds at which the switch operates and should be set to the same voltage as your logic or pulse generator.
If your input was 3.3V, say from an arduino, then you'd tie Vl to the 3.3V rail.
 

[elki]

Thanks again. This works well. However, for short pulses < 400 ns (my main application), the typical turn-on time is of the order of 100 ns or so, which screws up the output pulse shape. Any suggestions on fast analog switches?

[elki]

This is the signal at the output of the analog switch. As can be seen, the two peaks are the turn-on and turn-off times.

[Zero999]

A few comments regarding the original poster's schematic:

The power supply to the op-amp is only 5V, yet the input voltage is also 5V, so even if it has a rail-to-rail input and output, it won't output any more than 5V.

M1 will need to be a logic level MOSFET with a low threshold voltage.

The MOSFET will have some capacitance, causing the gain to tend towards 1+R3/(R1||R2) at higher frequencies, when it's off. This will result in over shoot on the output, as the higher frequencies are amplified more.

What's going on with C1? That'll AC couple the signal, so the output will swing negative, as well as positive.

[Terry Bites]

That looks like charge injection. I'm gusesing that the only load is a scope probe, possibly with a poor ground.
Add a small cap to see the effect. Put a realistic load on the switch common and buffer the output with an opamp.
For the DG419 the charge injection is about 10pC. Q=CV so a few pC in a few pF is a few volts. With a high impedance probe, the charge has no where to go leading to a long ugly transients. You could just us a small signal mosfet (2N7002) to pull Vtop to ground. Opamp optional.

A fun way to get control is to replace a gain set resistor with an LDR (often under a buck each). Shine a LED on it adjust theh LED brightness to set gain. So simple.
AKA "Vactrol".

[elki]

Thanks a lot. This is very inspiring and educating.

That looks like charge injection. I'm gusesing that the only load is a scope probe, possibly with a poor ground.
Add a small cap to see the effect. Put a realistic load on the switch common and buffer the output with an opamp.
For the DG419 the charge injection is about 10pC. Q=CV so a few pC in a few pF is a few volts. With a high impedance probe, the charge has no where to go leading to a long ugly transients. You could just us a small signal mosfet (2N7002) to pull Vtop to ground. Opamp optional.

A fun way to get control is to replace a gain set resistor with an LDR (often under a buck each). Shine a LED on it adjust theh LED brightness to set gain. So simple.
AKA "Vactrol".