# Detecting Pico-Amp AC signals at 100KHz

Discussion in 'Physics' started by will6459, Aug 6, 2013.

1. ### will6459 Thread Starter New Member

Aug 6, 2013
5
0
I'm working on a project where I will be attempting to detect a change in current (of the order of pico-amps) at ~100 KHz between a capacitor (dielectric is simply a vacuum). I'm looking for any pointers anyone may have in circuit design for measuring such low signals at high frequencies. I have modelled the system and have a strong idea of the signals I'm looking for but have little background in amplifying circuits.

Will

2. ### WBahn Moderator

Mar 31, 2012
23,563
7,212
Could you explain a little more clearly. I don't know what it means to detect a change in current between a capacitor. The word "between" implies two things.

3. ### will6459 Thread Starter New Member

Aug 6, 2013
5
0
Sorry my description was a little vague. Essentially I'm trying to detect ions as they rotate between two parallel plates, the ions induce an image charge on each plate as they rotate between the two. So the change in charge over time can be measured as an alternating 'image current'.

4. ### WBahn Moderator

Mar 31, 2012
23,563
7,212
What is the capacitance?

Do you have control over the rotation? In particular, can you provide a signal that is locked to the rotation rate?

The thoughts that come immediately to mind are to look at lock-in amplifier techniques and perhaps correlated double sampling.

If the capacitance is small enough, you might be able to look at the capacitor voltage more-or-less directly. If you have 1pA on a time scale of 10μs, that's a change of roughly 10aC. Pretty small. But if the capacitance was 1pF, that would still give you a voltage change of about 10μV. So not entirely out of the question.

5. ### will6459 Thread Starter New Member

Aug 6, 2013
5
0
Well its not strictly a parallel plate system (slightly curved) which makes calculating the capacitance a little more complex, however a good approximation is ~0.05pF.

The rate of rotation is used to accurately calculate properties of the ions. However I'm trying to detect a range of frequencies (bandwidth of around 200KHz), so I'm guessing using a single signal wouldn't be suitable?

6. ### WBahn Moderator

Mar 31, 2012
23,563
7,212
That would get you up into the range of 200μV, which is probably enough to work with.

[/QUOTE]

No, not to use lock-in techniques. For a lock-in approach, you would need to have a signal that can sample reliably at the different phases of the signal. There's a reason they call it a "lock-in" amplifier.

My first suggestion would be to look for a FET-input differential preamplifier with good performance specs. With a bandwidth of only a few hundred kilohertz at a gain of several hundred to a few thousand, that shouldn't be too hard to find.

7. ### will6459 Thread Starter New Member

Aug 6, 2013
5
0
Perhaps something like this? http://pdf1.alldatasheet.com/datasheet-pdf/view/56673/BURR-BROWN/INA111.html

Looking at some amplifier circuit designs I've found they use a three stage design.

Firstly with the above instrumentation amplifier (INA111) using a 10MΩ resistance at each input (apparently a gain of 10 is achieved).

Secondly a single-pole low-pass RC network with a high frequency breakpoint of 500KHz. I don't really understand what this is/does.

The third and final stage uses another INA111 to provide a further gain of 50.

Thanks for the feedback, its greatly appreciated.

8. ### WBahn Moderator

Mar 31, 2012
23,563
7,212
That may work -- probably worth looking into. I haven't looked at the actual specs. When I tried to access the datasheet via that link I got a really strange message:

Your require pages is cannot open by blow