• State
  • Locked Locked
  • Replies 3 replies
  • Subscribers 47 subscribers
  • Views 133 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

OPA928EVM: Application support for

ANDREW SAUTER
Intellectual 280 points

Part Number: OPA928EVM
Other Parts Discussed in Thread: OPA928, DDC232, IVC102

I am working with this sensor Unisense hydrogen sensor.  The sensor requires a 100 mV bias, and the system will operate between 10°C and 30°C. We require maximum thermal stability and environmental isolation; minimal motion and EMI are expected.  

We aim to measure current with a 0.1 fA resolution. Our signal has negligible high-frequency content (near-DC). We anticipate replacing the standard feedback resistor with a high-value glass-encapsulated resistor.

What is the most effective way to implement a small form-factor solution for this high-gain stage and subsequent digitization?

  • 0
    TI__Genius 15045 points

    Hi Andrew,

    The OPA928 is the right device, and the EVM is a good place to start.  I'm not sure if "glass-encapsulated resistor" and "small form-factor" really go well together - I don't have much advice on how to make it all smaller but we do have some good collateral on minimizing the effects of things like board-level leakage and dielectric absorption when using the OPA928,  I would check out this app. note:

    https://www.ti.com/lit/an/sboa597/sboa597.pdf

    The methods implemented in the app. note above are some of the tricks we used to characterize the OPA928, our resolution was on the 0.01 fA level.

    Things to keep in mind to minimize noise:

    1) no DC/DC converters nearby
    2) minimal trace lengths when possible
    3) use the guard pin to "guard" sensitive traces and also to drive any shields on cables
    3) be careful with ground connections; any ground loop will easily couple in

    Let me know if I am headed in the right direction here or if you were looking for different information.

    Regards,
    Mike

  • 0
    Intellectual 280 points

    Very helpful.

    Do I need to discharge the cap with another circuit? or should i swap the nodes on the front end so that it charges and discharges?

    Do you think we can get 0.1 fA without charge couting?

    the parasitic capacitican vulnerability makes me very concerned about prototyping this with EVMs!

  • 0 in reply to ANDREW SAUTER
    TI__Genius 15045 points

    Hello,

    Yes the parasitic capacitance is an issue with the charge-counting method, along with the capacitor non-idealities (voltage coefficient, leakage).   Our team designed the EVM to minimize parasitic capacitance (guard traces, eliminating ground plane) but this does need consideration.  Discharging the capacitor is another issue; any switch added to the capacitor for discharge can add leakage.  Generally the lowest possible leakage is from relays; most MOS-based switches will have too much leakage.  If you reverse the connection, that will also need a switch so I don't see how that could be any better.  I can check with the team to see which relays they have found to be the lowest leakage if you are interested in this.

    Using a TIA-style approach (which is the other circuit on the EVM, the "transimpedance amplifier"), you will get a constant voltage proportional to the large feedback resistor and the measured current.  This will have a significant amount of noise, though, that would then need time to average out to end up with the desired resolution.  This highlights the tradeoff between the charge-counting method and TIA; in one case, you wait for some time while charge accumulates on a capacitor.  But for a TIA, you can see (once the amplifier has settled) a voltage that is proportional to the current flowing into the amp. at any point in time; however, this may have so much noise that the resolution is effectively very low and needs to be averaged, during which time if the measured current changes it just gets averaged anyway.

    So I can't say I know a clear-cut answer besides to say that the highest resolution current measurement techniques tend to use use charge-counting.  There are other devices also that feature this (IVC102, DDC232) for very high resolution, whereas faster current measurement techniques with lower resolution will use a TIA (such as optical communications).

    Hope this helps!
    Mike