Simulating a Charge Amplifier with LTSpice

[johofz]

Hello dear EEVblog-Community,

I'm currently designing a charge amplifier for a piezoelectric force sensor. I found a lot of material about charge amplifiers so I feel comfortable with my design.
But since I'm a cautious person, I'd like to simulate it with LTSpice and confirm it's behavior.

Here is the data-sheet for the sensor: https://www.kistler.com/files/document/000-630e.pdf?callee=frontend
It has a sensitivity of -3.7pC/N and a capacitance of 65pF.
I want my amplifier to produce an output of 1V @ 10kN and 0V at 0N.
For the modelling I went with a voltage source with a 65pF capacitor in series, since it seemed like the easiest way. From my understanding I need to calculate the voltage for the voltage source by dividing sensitivity with capacitance of the sensor. That gives my 56.9mV/N. I went with an AC-source that oscillates around 284.6V with +-284.6V amplitude (like a signal from 0N to 10kN would do). I choose a 500Hz frequency.

For the feedback-capacitor I choose 39nF, since that should produce roughly the desired output of 1V at 10kN output. (10000N * 3.7pC/n / 39nF = 0,9487V).
I calculated the high-pass freq as 1 / (2*pi*1GOhm*39nF) = 0,0041Hz
Low-pass freq without cable is calculated as 1 / (2*pi*1kOhm*65pF) = 2.45MHz

Here is a picture of the circuit:


And here is what I get as a result of the Simulation:


Am I missing something? I Expect V_OUT to be a sinusoidal curve bewteen 0 an d 1V. Is there a problem in my circuit?

Any help is greatly appreciated

greetings
johofz

[johofz]

I just found a massive bug in my quick and dirty example. I forgot the power supply for the opamp. Here is the corrected result:


It's still not what I expected. It seems like the output has an offset of 0.5V.

[jmelson]

It's still not what I expected. It seems like the output has an offset of 0.5V.

Yes, of course.  You are powering a bipolar circuit (ground-referenced sine waves) from a unipolar power supply.

You need to shift the reference positive, through the non-inverting input of the op amp, or provide a negative supply also to the op amp.

Jon

[johofz]

Yes, of course.  You are powering a bipolar circuit (ground-referenced sine waves) from a unipolar power supply.

This is the output of the ac voltage source:


I think my error comes from not fully understanding the whole concept of charge amplifiers yet. I thought the offset on the sine wave would produce only positive charge so I wouldn't have to bias the opamp.

I used this document from Analog Devices as a reference: https://www.analog.com/media/en/reference-design-documentation/reference-designs/CN0350.pdf
They do bias the opamp and the sensors negative terminal from the same potential.
In my application I know I will only get compression on my sensors axial force axis, so I thought it's fine to not bias the opamp

[johofz]

After digging into it a bit more, I think I've got it correctly now. The problem was that I had a wrong idea about the way the voltage-source produces charge trough the capacitor. The DC-offset on the ac-voltage source doesn't do anything. Only the amplitude has an effect.

Here is my result with an additional adc-preamp: