• State Resolved
  • Locked Locked
  • Replies 4 replies
  • Subscribers 47 subscribers
  • Views 4292 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

IVC102 for nA measurements

Gabriel Koutilellis
Prodigy 50 points
Other Parts Discussed in Thread: IVC102

I have carefully read the datasheet of this transimpedance amplifier. I have followed the pcb outlines, but I am against some weird behavior.

Power supply is +/- 15V and capacitance used is 100pF. So V = - i * T / C = - i * T / (100 pF).

When I measure positive currents the transfer function seems to be V = -1.002 * i + 0.006. That`s pretty well expected and understood.

When I measure negative currents the transfer function seems to be V = -1.208 * i + 0.02. This is something I don`t like firstly, as I expected symmetry, and secondly I cannot understand why is there such a different behavior.

Is there something that could help me understand?

  • 0
    TI__Prodigy 130 points

    Gabriel,

    Without knowing a bit more detail on the switching sequence and timing, I may not be able to answer all of your issues on this circuit, but let me try and shed some light on probable sources of error. 

    The diffent slopes on the transfer functions that you showed are most likely just due to the IVC102's bias current.  The bias current of the integrator will appear to increase slope of currents that are in the same direction as the bias current,  and decrease the slope of currents in the opposite direction.  I do not know what currents you are trying to measure,  but if they are in the neighborhood of +- 1pA,  the input bias current of the IVC would give you a ~10% error in slope.   Another potential error that would cause an identical error would be any consistant leakage to the input of the IVC.   This could be due to humidity or other contaminants on your printed circuit board between the input and any voltage source in the vicinity.

    Now the difference in y axis intercept;  This is a harder error to explain.   One possible cause of this could be stored charge on the input pin.  If the source is connected to the input with the input sample switch in the open position,  the pin will "charge" up to some extent.  When the sample switch is closed this charge will be distributed to the integration cap and apear as a change in charge injection.   So, if a different souce is connected in this same fasion,  it is reasonable to expect a different impact on the charge injection.

    I would recomend doing a null measurement with your specific timing.  By this I mean a measurement with the input of the IVC not connected to any current source.   This can go a long way in canceling most errors,  including bias current,  board leakage,  and charge injections.   Subtract the null from the actual unknown current integration and you should see improved results.

    Bubba

  • 0 in reply to Bubba
    Prodigy 50 points

    first of all the switching timing is exactly the same for positive and negative currents. I use the switched-input measurement technique. I have a micro-controller doing it and I have already checked it with an oscilloscope!

    The measured currents for the data I posted are in the range of +/- 12uA, 1 pA is a really difficult current!

    It's out of my interest!

    I have also made some tests in the nA range with exactly the same effect observed.

    Something I have also noted is that the voltage output of IVC102 when measuring (capacitor charging) positive currents is a straight line that declines until I reset the integrator. When measuring negative currents the straight line becomes a curve like the one of a charging capacitor that augments of course, until the reset.

  • 0 in reply to Gabriel Koutilellis
    Prodigy 50 points

    Also thank you for your answers!!!

  • 0 in reply to Gabriel Koutilellis
    TI__Prodigy 130 points

    Gabriel,

    Ahhh...  100pF cap with 12 uF current is pusing the slew rate of the amplifier..   If you are really looking at that magnitude of current I would recomend using a much larger integration cap.   This will have two effects.  Your charge injections will be smaller and you will not be pushing slew limit of the amplifier.   I would suggest upping the cap two decades.  Right now you should be integrating at ~1.2V/uS  so in a dozen uS you will hit the rail of the integrator..   I don't know what integration time you are trying to use or what the speed requirnments are for your application, but 1.2V/uS is ripping...   You can use any low leakage cap in parallel with the internal cap.     I think this should make many of your issues go away.   If you still see different slopes for the transfer function of positive and negative currents....look for leakage!

    Bubba