Differentiator

[cmb042]

I have an amplifier circuit I am tryin to get more gain out of. I increse Rf and the signal is attenuated. I don't understand why. I first thought that if Rf gets to be too much bigger than Ri, then the capacitor can't discharge fast enough for the input to register the next pulse. I tried increasing, then decreasing the capacitor C and then Ri but I didn't notice any effect. The negative supply rail of the opamp is grounded, and the noninverting input is supplied with about 1V DC. The input signal compsist of 30-200 negative pulses per second of microsecond duration.

Maybe some sort of phase problem? Not sure where to start on a test for that.



crappy sketch:

---Rf---
| |
| |\ |
---C--Ri----|-\ |
| >---
+1VDC--|+/
|/

 

[IRstuff]

You need to provide more information. What op amp is it? What's its bandwidth? What values of Ri and Rf are you using? What's the output current capacity of the op amp?
etc...




TTFN

FAQ731-376

 

[cmb042]

OP15 6MHz bandwidth, closest parameter I can see on the data sheet to output current capacity is Supply Current 2.7-4mA.

c= 470pF Ri=1k Rf=130k

 

[Noway2]

I just ran a quick spice simulation of this circuit and got different results. If you haven't done so, you may want to download a spice tool and simulate your circuit.

Also, your values (1K and 470pF) form an RC with a .47uS time constant which is about 1/2 of the duty cycle of your waveform. In other words, the capacitor will get a significant charge per pulse that will build voltage until a steady state is achieved. In all cases, I got a sawtooth waveform (assuming a very sharp transition pulse) that oscillated between 1 and 1.5V, at least once steady state was achieved, though there was a small amplitude change with different values. The value of the Rf and Ri had a major effect the transient response, i. e. settling time and damping.

I think you might be having a form of PWM - switchmode regulator actiong going on here and that the circuit is not behaving entirely as a differentiator amplifier in the terms you are thinking.

I also think you should analyze the LaPlace transfer function for your circuit and then reconvert this to the time domain. The equation should show you why you are getting the results you are.

 

[IRstuff]

And, have you considered that at the gain that you are trying to run, you're not going to get 6 MHz bandwidth, which is the GAIN-BANDWIDTH product.

A perusal of TPC21 in the Analog Devices datasheet shows that 42 dB gain results in only about 50 kHz bandwidth.

TTFN

FAQ731-376

 

[cmb042]

Ok, so if I understand this correctly what is happening is as I increase Rf, I am also decreasing the frenquency at which the circuit can respond. So I must be at the edge of the bandwidth.

So, there is no way I can maniputlate the circuit (short of changing to a higher bandwidth opamp) to get more gain for this signal? I think my next course of action will be to design a second stage linear opamp.

Thanks for the help.