• State Resolved
  • Locked Locked
  • Replies 18 replies
  • Subscribers 51 subscribers
  • Views 9268 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.

Choosing a transimpedance amplifier to replace OPA111

Wayne Chu
Prodigy 120 points
Other Parts Discussed in Thread: OPA627, INA111, OPA827, OPA320, OPA140, OPA2320, OPA2140, LMP7721

Hi TI Engineers,

I obtained a reference design of a transimpedance amplifier that uses two amplification stages to achieve the following:

  • Input range below pA to nA
  • Gain of 10^8 - 10^10 (V/A)
  • Max bandwidth up to 100 kHz
  • Noise below sub-pA with 100 pF input capacitance (around 1kHz)
The design is used as a pre-amplifier for STM tunneling current, thus they would be on the order of pA to nA.  I am trying to reproduce the design with some new size and power requirements.  The original design calls for OPA111BM and OPA2111KP, which have a large footprint.  Ideally the components are surface mounts, because I need to fit the entire design in an area measuring 3/4" x 5/16".  Also, the available DC power is ~ 10V.
Do you have a recommendation on the precision amplifiers that will fit these criteria?
Thanks,
Wayne
  • 0
    TI__Guru 62745 points

    Hi Wayne,

    Would you please provide a schematic of the design you're referencing?  It would be easier to recommend replacements if we could see the circuit.

    Thanks!
    Collin Wells
    Precision Linear Applications

  • 0 in reply to Collin Wells
    Prodigy 120 points

    Hi Collin,

    I have attached the schematic from the paper.  Let me know if this helps.  Thank you very much!

    Source: A low-noise & wide-band ac boosting I-to-V amp for STM (2005), [DOI: 10.1063/1.1841873]

    Wayne

  • 0 in reply to Wayne Chu
    TI__Guru 62745 points

    Hi Wayne,

    The second stage composite amp is a topology that I'm not familiar with.  I'll have to go play with it when I get some time to figure out how it works. 

    For replacement parts I kept the search to parts that support smaller package sizes, +/-15V supplies, JFET inputs, and bias currents close to the OPA111.  With this search I found the following parts:

    OPA124
    OPA129
    OPA627

    If you are willing to change the power-supply voltages then some other parts become available as well.  You can search though all of TI's JFET / low Ib parts using the main parametric search tool found below:

    http://focus.ti.com/paramsearch/docs/parametricsearch.tsp?family=analog&familyId=1478&uiTemplateId=NODE_STRY_PGE_T

    Hope this helps!
    Collin Wells
    Precision Linear Applications

  • 0 in reply to Collin Wells
    Prodigy 120 points

    Thanks Colin,

    I will definitely check out OPA124, OPA129, and OPA627.

    We were also examining the instrumentation amplifier INA111.  Can this part produce the similar requirements for us (gain = 10^9, noise around pA)?  The spec sheet of INA111 said it has the gain from 1 to 1000 (V/V).  I couldn't find the gain-bandwidth product on the INA111.  I had always thought instrumentation amplifier to be a different beast, maybe this is why?  Are precision amplifiers pretty much all JFET's?

    Wayne

  • 0 in reply to Wayne Chu
    Genius 5430 points

    Wayne;

    This is an "integrating front-end", a transimpedance amplifier with a very high feedback resistance to minimize its current noise and a quasi-differentiating second stage to flatten the overall frequency response. As the TZA gain rolls off, the 2nd stage gain begins to increase, giving a flat frequency response. It is used for increased SNR at the expense of Dynamic range.

    I'd look at an OPA827 for the TZA; it has very low noise for a FET op amp and good BW. for the 2nd stage, you can use the same type of op amp or even consider a low noise bipolar op amp.

  • 0 in reply to Neil Albaugh
    TI__Guru 62745 points

    Good suggestion Neil, and thanks for the explanation of the circuit it makes sense now.  If you place the zero of the second stage slightly before the AOL dominant pole kicks in then the response will stay mostly flat until the second stage gain begins to decrease again.

    Thanks,
    Collin Wells

  • 0 in reply to Collin Wells
    TI__Guru 62745 points

    Hi Wayne,

    I missed a few of your questions.  No, the INA111 nor any other 3-amplifier instrumentation amplifier can be used for a transimpedance amplifier.  The INA111 does have a bandwidth product, but it usually shows it only in closed-gain configurations which is why you will likely find mutiple numbers on the datasheet for the closed-loop bandwidths depending on the specified gain. 

    Amplifiers can have a variety of input stages from CMOS, Bipolar, JFET, and some interesting hybrid varieties.  The different topologies are typically targeted towards applications where they provide benefit such as low-noise, low input bias current, or low cost/size.

    Regards,
    Collin Wells

  • 0 in reply to Neil Albaugh
    Prodigy 120 points

    Hi Neil,

    Really appreciate your insight.  I will try the OPA827 for the TZA (1st stage), as well as what Collin recommended (OPA124, OPA129, OPA627).  One thing did jump out at me concerning the OPA827 is its max input bias current of 50 pA (typical input bias current at 15 pA).  The source current I will be dealing with would be around the range of pA up to tens of nA.  Of the four aforementioned op-amps, The OPA129 seems to have the lowest input bias current at 0.1 pA.

    Also, I am just wondering do OPA827 (or any of the above precision opamps) have a dual op-amp package.  The intention is to stay with the composite op-amp configuration and still maintain a small footprint for the 2nd stage of the amplification?

    Thanks,

    Wayne

  • 0 in reply to Collin Wells
    Prodigy 120 points

    Thanks Collin,

    I am looking at requesting the samples for these aforementioned op-amps.  Looking at the available parts, I see many options for the same design.  For example:

    OPA827 has OPA827AID, OPA827AIDG4, OPA827AIDGKR, OPA827AIDGKT.  I understand the first two and the latter two have different package types (SOIC vs MSOP), but are there any other differences between the parts?

    The same goes for OPA124.  The two part I am looking at are OPA124U and OPA124U/2K5.

    Thanks again!

    Wayne

  • 0 in reply to Wayne Chu
    TI__Guru 62745 points

    Hi Wayne,

    Some of our precision parts have High-Grade and Low-Grade versions.  The OPA124 is an example of this type of product, where the OPA124UA is the High-Grade and the OPA124U is the Low-Grade. 

    Besides the High/Low Grade distinctions the other characters represent the package type and whether it is Lead Free and/or Green. 

    The "2K5" and/or "R" characters mean it is in a full reel of several thousand units, this is not what you're looking for if you want a few samples.

    Regards,

    Collin Wells

  • 0 in reply to Wayne Chu
    Genius 5430 points

    Wayne;

    You're probably right-- I see that you did not need much BW in your first stage so a lower input bias current op amp would be a better choice; the OPA129 has extremely low Ib so it should do nicely.

    Another one to keep in mind for this type of application is an OPA320. It has low Ib and higher GBW so it can fit in applications where you need more BW but still need low current noise.

  • 0 in reply to Wayne Chu
    Genius 5430 points

    Wayne;

    One further thought-- if you need even lower input bias current, try cooling the transimpedance amplifier with a small Peltier cooler. The Ib and current noise of an OPA129 will improve down to about -40C. While you're at it, cool the 1G feedback resistor, too-- its thermal noise will decrease as well.

  • 0 in reply to Neil Albaugh
    Prodigy 120 points

    Collin and Neil, thank you both for your help and advice!

  • 0 in reply to Wayne Chu
    Guru 21975 points

    Wayne,

    I'll add a bit more information to this long thread...

    Regarding the OPA827, we will soon be improving its input bias current spec to 10pA. A change in wafer processing allows this improvement. Earlier devices had input bias current that varied with common-mode voltage. Furthermore the bias current took many seconds to settle to a new value when the common-mode voltage changed. This effect has been eliminated. The OPA827 is a truly exceptional op amp and it's even better now. The new data sheet should appear on TI site in a week, or so. Though we are purging stock of older devices, it's difficult to assure that all stock is cleared from distributor inventory. Contact us if you need details on the transition.

    I did not see the OPA140 mentioned as a possibility. It also is a great new FET input op amp with lower noise and higher speed than the OPA111 with excellent speed/power ratio. It also has a 10pA spec for input bias current.  Single, dual and quad versions are available.

    Regards, Bruce.

  • 0 in reply to Bruce Trump
    Prodigy 120 points

    Hi Bruce,

    I requested some samples over the weekend, but did not see OPA827 being available for sampling.  I will check out OPA140 as well, the dual version will be very helpful for the second stage of amplification.

    Thanks again for everyone's help!

    Wayne

  • 0 in reply to Wayne Chu
    Genius 5430 points

    Wayne;

    Since the second stage is a gain equalization stage with gain rising with frequency, choose an op amp that has very low voltage noise.

  • 0 in reply to Neil Albaugh
    Prodigy 120 points

    Hi Neil, thank you for the advice regarding the second stage.  Originally, I had intended to try the following combination for 1st and 2nd stage:

    1.  OPA124 (1st stage) and OPA2320 (2nd stage)
    2.  OPA129 (1st stage) and OPA2320 (2nd stage)

    I just found out our power supply VDC is +-15V, so this will change my second stage op-amp selection.  With the OPA2320's Vs_max at +-6VDC, I will have to reconsider a different opamp for 2nd stage.  Creating a voltage divider may also not be suitable for the small footprint we want to achieve.

    Searching through the TI's Low Input Bias Current/FET Input offering with my power supply requirement (+- 15VDC_max) in mind, I found OPA2140 offer a 5.1 nV/sqrt(Hz) at 1kHz. It also offers dual channel, suitable for the small footprint I wanted for 2nd stage.  What's your take on it?

    I was also pretty amazed the LMP7721 by National Semiconductor has a input bias current of 3fA (max at 900fA or 0.9pA) and 6.5 nV/sqrt(Hz) at 1kHz.  It takes in a max supply VDC of +- 6V, so too bad my power supply is not at that range, it would be nice to see it in action.

    I am still learning by the day, so really appreciate your support and suggestions!

  • 0 in reply to Wayne Chu
    Genius 5430 points

    Wayne;

    An OPA2140 should be a good choice.

    We never stop learning, Wayne.