# Differentiator design - Meeting design requirements

#### Elevon

Joined Jun 7, 2020
18
If I needed a differentiator amplifier to have a maximum voltage output of +12V and a maximum signal slope of 0.2 Volts per microsecond, what components or component values can I use to achieve this?

#### WBahn

Joined Mar 31, 2012
30,034
There are many differentiator circuits out there. Some use opamps, some use transistors, some use vacuum tubes, some use other things. Might it not be a good idea to first come up with a circuit that, conceptually, performs the function you want and then determine what specifications you needs from the components that make up that circuit?

#### crutschow

Joined Mar 14, 2008
34,396
I assume you mean "differential"?

#### Elevon

Joined Jun 7, 2020
18
I assume you mean "differential"?
Not sure if that's the same thing or not but the question states a 'Differentiator Amplifier'. It involves the circuit shown below:

I'm using the value for signal slope (0.2) as the rate of change of input voltage signal with respect to time, and I'm assuming the output voltage is 10V across the Rf terminals, I just don't understand how the current is split across the parallel components.

#### crutschow

Joined Mar 14, 2008
34,396
That circuit is a combined high-pass (R1C1) and low-pass (CfRf) (bandpass) filter, not just a differentiator
A pure differentiator would consist of just C1 and Rf.

What is the exact signal you are trying to differentiate?

#### LvW

Joined Jun 13, 2013
1,753
A pure differentiator would consist of just C1 and Rf.
Yes, that is correct - IF IT WOULD WORK!
In most cases, such a "pure" differentiating circuit is unstable or very close to instability.

#### crutschow

Joined Mar 14, 2008
34,396
Yes, that is correct - IF IT WOULD WORK!
In most cases, such a "pure" differentiating circuit is unstable or very close to instability.
It may be noisy but why would it be unstable?

#### LvW

Joined Jun 13, 2013
1,753
It may be noisy but why would it be unstable?
Because it is very probable that the loop gain drops with 40dB/dec around the 0dB-cross-over.
(Open-loop gain as well as the feedback factor both drop with app. 20dB/dec).
Hence, for stabilizing the circuit, a resistor in series to the input capacitor is necessary.

#### crutschow

Joined Mar 14, 2008
34,396
Hence, for stabilizing the circuit, a resistor in series to the input capacitor is necessary.
Okay, thanks.