Author Topic: Testing transimpedance amplifiers  (Read 2351 times)

0 Members and 1 Guest are viewing this topic.

Offline new299Topic starter

  • Regular Contributor
  • *
  • Posts: 119
Testing transimpedance amplifiers
« on: June 16, 2020, 01:56:43 pm »
I've looked at transimpedance amplifiers in the past for various applications. Generally these would be used with a photodiode or another sensor (for example, used in a scanning tunneling microscope). I'm less than clear on how these amplifiers should be independently evaluated.

If I'm looking at doing measurements in the pico to nanoamp range, I've tried a sticking 100Meg Ohm resistor on the output of a function generator. This seems like a relatively common approach. It seems to work relatively well. And I assume is fine because the input impedance is low?

However I've also seen people suggesting building a current source using an LT1167 [1] for testing.

Alternatively, Keithley make programmable picoamp current sources. Like the Keithley 220 for example.

What I don't understand well is the tradeoffs between these various options. Is using a function generator just as valid as the other options? Or are there issues that it could cause?

[1] http://images.100y.com.tw/pdf_file/33-LT-LT1167.pdf figure 7.
 

Offline RoGeorge

  • Super Contributor
  • ***
  • Posts: 7012
  • Country: ro
Re: Testing transimpedance amplifiers
« Reply #1 on: June 16, 2020, 02:46:33 pm »
In theory, no difference between a voltage source + a big resistor instead of a current source.  There is even a way to convert from one to another:  Thevenin/Norton equivalence for (Vsource with serial Rs) <===> (Isource with parallel Rp).

In practice, it depends.  No component is ideal, so a resistor is expected to have stray capacitance, inductance, temperature coefficient, etc.  It depends what parameters matter or not for a given test, and what value ranges are expected for a given test.

The specs for a tunneling microscope amplifier are not common knowledge.  Might help answering if you briefly explain what do you want to test for a transimpedance amplifier used in a tunneling microscope.  Ranges for current, noise, frequency, temp, specific peculiarities, etc.
 
The following users thanked this post: new299

Offline new299Topic starter

  • Regular Contributor
  • *
  • Posts: 119
Re: Testing transimpedance amplifiers
« Reply #2 on: June 16, 2020, 02:55:31 pm »
So, typically using either a 1M, 10M or 100M feedback resistor on the transimpedance amplifier. I'm interested in picoamp or nanoamp ranges. I'd like to be able to characterize down to 10s of picoamps of noise (at worst), bandwidth maybe upto ~10KHz in the picoamp range. ~100Khz in the nanoamp range. Room temperature.
 

Offline RoGeorge

  • Super Contributor
  • ***
  • Posts: 7012
  • Country: ro
Re: Testing transimpedance amplifiers
« Reply #3 on: June 16, 2020, 07:59:38 pm »
100M\$\Omega\$ is not a common value, might be hard to find, or expensive.  That alone can be a good reason for a very low current source.  Some resistors might be noisier than other, that could be another reason.  Large impedance variations of the load (compared with the required series resistor - not your situation), maybe some well designed active source current sources can deliver highest output impedance than a simple resistor for a decently low voltage, and so on.

Why are you looking for a controlled current source, did you get unsatisfactory results with the series resistor method?
 
The following users thanked this post: new299

Offline nfmax

  • Super Contributor
  • ***
  • Posts: 1624
  • Country: gb
Re: Testing transimpedance amplifiers
« Reply #4 on: June 17, 2020, 07:20:40 am »
Drive a triangle wave through a very small capacitor into the input. This will give you a square wave current, and with readily available components you can set very small currents.
 
The following users thanked this post: new299

Offline new299Topic starter

  • Regular Contributor
  • *
  • Posts: 119
Re: Testing transimpedance amplifiers
« Reply #5 on: June 17, 2020, 03:19:42 pm »
100M\$\Omega\$ is not a common value, might be hard to find, or expensive.  That alone can be a good reason for a very low current source.  Some resistors might be noisier than other, that could be another reason.  Large impedance variations of the load (compared with the required series resistor - not your situation), maybe some well designed active source current sources can deliver highest output impedance than a simple resistor for a decently low voltage, and so on.

Yes, I've used some of the large Ohmrite through-hole resistors. They state "low inductance", but I'm less clear on their other specs.

Why are you looking for a controlled current source, did you get unsatisfactory results with the series resistor method?

The results look fine. I'm mostly concerned that they look too good. I.e. does this not model my system well, where I have a fixed bias voltage and, as I understand it, varying resistance (the tunneling junction). So in the real system, I'm measuring the current through that varying resistance.

In this test setup I have a varying voltage and a fixed resistance. So I imagine this is equivalent. But (likely due to my own poor understanding) it concerns me that there might be issues I'm unaware of.
 

Offline new299Topic starter

  • Regular Contributor
  • *
  • Posts: 119
Re: Testing transimpedance amplifiers
« Reply #6 on: June 17, 2020, 03:20:48 pm »
Drive a triangle wave through a very small capacitor into the input. This will give you a square wave current, and with readily available components you can set very small currents.

Thanks, I like this approach and will also use it for testing.
 

Offline RoGeorge

  • Super Contributor
  • ***
  • Posts: 7012
  • Country: ro
Re: Testing transimpedance amplifiers
« Reply #7 on: June 17, 2020, 04:33:04 pm »
So, if I got it right, you need the pA source not to test/measure a transimpedance amplifier, but to simulate the tunneling microscope.  I don't know enough about the tunneling to help you with that, but given the fact that a capacitor is a derivator, looks possible to integrate the desired waveform, and to apply it as varying voltage through a series capacitor.

Regarding how to generate small currents, there are all kind of tricks, including linear voltage through a capacitor mentioned by nfmax, or even mechanically vibrating capacitors https://doc.xdevs.com/doc/_Metrology/femto_ampere_current_source.pdf

Offline FreezeSSC

  • Contributor
  • Posts: 24
  • Country: us
Re: Testing transimpedance amplifiers
« Reply #8 on: June 18, 2020, 02:02:02 pm »
One thing to keep in mind is that most transimpedance amplifiers have to deal with peaking so you'll want to try and test for that with your setup, not testing for that can lead to oscillating op amps. 
 
The following users thanked this post: new299

Offline new299Topic starter

  • Regular Contributor
  • *
  • Posts: 119
Re: Testing transimpedance amplifiers
« Reply #9 on: June 21, 2020, 09:49:24 am »
One thing to keep in mind is that most transimpedance amplifiers have to deal with peaking so you'll want to try and test for that with your setup, not testing for that can lead to oscillating op amps. 

Thanks, yes I generally look at the Bode plots created using the built in waveform generator on my Keysight scope. I then add caps to the feedback to remove the peaking.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf