Circuit Design for a fractional -amplifier

[iqjump123]

Circuit Design for a "fractional"-amplifier

Hello all,

I am right now in the process of designing a simple device that will help me measure the voltage response of a piezoelectric element attached to a specimen after it has been hit with a projectile. The obvious issue is that the amount of voltage picked up, although short in duration, will be very high (~300V), and high in frequency values. I wanted to build an amplifier circuit that will take the voltage detected by the piezo element and decrease it down (while keeping the signal quality) to a +/- 5V signal. In addition, it would need to have a high frequency response, up to about 100kHz.

What I was thinking of doing was to use a charge amplifier circuit shown as
http://www.mmf.de/instrumentation.htm"

One issue that I am envisioning is the main reason to build a charge amplifier is to detect the small amount of charge coming from a piezo element and amplifying it, NOT to actually decrease it, and therefore choosing the right component that will give a high frequency response AND can exhibit fractional gain can be tough.

would it be just best to hook the system up to a divider circuit and then amplify that signal through a typical amplifier?

Thanks for your help in advance.

iqjump123

 

[cmb]

If it's just a high impedance 5V you need, then what's wrong with just a resistor divider (of many MOhm total)? If it is a low impedance 5V drive you need, what is it going to feed?

Maybe you might be also/more interested in integrating the signal with an op-amp/capacitor?

 

[iqjump123]

If it's just a high impedance 5V you need, then what's wrong with just a resistor divider (of many MOhm total)? If it is a low impedance 5V drive you need, what is it going to feed?

Maybe you might be also/more interested in integrating the signal with an op-amp/capacitor?

Hey cmb, thanks for your reply.

Right now, the signal I record from the shot will go from my circuit to an oscilloscope/DAQ system, so what I want is an amplifier that will serve to (1) view the output response in the screen, but (2) also serve as a buffer, so that the detected voltage spike won't go directly to the DAQ system and damage it.

Isn't the charge amplifier circuit links involve integrating the signal using opamps and capacitors? (Unless you were suggesting that I use the actual circuit design for an "integrator")

Also, Using a resistor divider will indeed decrease the voltage, but I fear that since a piezoelectric element is a capacitor that stores charge as well, current might come into picture- even though the voltage will be stepped down from the divider, not having a separate buffer stage will cause the current or charge to enter the recording device. Also, since I will be recording to a device, I would want a low impedance output.

Please feel free to correct me.

Thanks for the fast reply as well.

iqjump123

 

[cmb]

I'd not at all put myself in a postion of 'well-informed/professional' on this (.. hoping others will join in here.!..) I 'd class myself as only 'amateur-practitioner'.

I've built a few feedback amps, from discretes, to measure units-10's worth of picoCoulombs (as ~uA over ~10uS). There are opamps that do femtoAmps but I imagine that from a piezo you would be amplifying noise rather than a signal at those levels.

Need someone with more experience to confirm this, but if I was doing this I'd start by testing a 100MOhm/1Mohm divider. The 'scope should be enough impedance for you to check, first of all, you are getting the signal you want and expect. Then use an opamp/discretes amp buffer to follow the voltage, 1:1, for output to your DAq.

 

[es1]

Totally agree with cmb, start with a divider. Depending on the probe, buffer the divided node with a follower.

Only if it doesn't work go to the more complicated integrators.

My only caution is 100M/1M is pretty big and even if there is only a few 10s of pF on the follower node (due to the opamp or probe) the measured signal will be low passed into KHz. As ~100KHz values are desired you might want to try lower value resistors.

 

[cmb]

Thanks for the back-up. Yes, of course, once you see anything at all (or if you don't), back off the resistances until you optimise your signal/noise, without compressing the signal.