• State
  • Locked Locked
  • Replies 4 replies
  • Subscribers 49 subscribers
  • Views 605 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.

Current limiting an op-amp

ILYA LISENKER
Prodigy 60 points
Other Parts Discussed in Thread: OPA2277, LM6172, BUF634, THS4052, OPA2613, LM6171, OPA2677, OPA2674

Hello everyone,

first a disclaimer, I am a mechanical engineer. As a grad student, I need to build an AC amplifier with current limiting (and it needs to function for my experiments for the next year or two). There is a nice Burr-Brown application note on this subject here  http://www.ti.com/lit/an/sboa042/sboa042.pdf      figure 2.

My challenge is that I'd like this beast to operate at  up to 10MHz (higher if possible), +/-5V. The signal is a pure sinusoidal tone. There is a note about stability on page 1, column 2 and I am having a heck of a time finding an op-amp to use in place of OPA2277 that has a slew rate >100V/microsec and bandwidth that is greater than 20MHz. Does anyone know, can I use a faster op-amp, like an LM6172 and add an RC "filter" to the feedback loop around A2 and A3? Do I need to worry about stability if my signal is just a single frequency sinusoid. Alternatively, could I just add gain to the A2 part to slow it down? My signal comes from a function generator so I am pretty flexible on the input side. 

While I'm at it, I need a higher current output than a single BUF634 can provide. I'm planning to try driving 3 of them in parallel and I know that voltage mismatch at the outputs can cause problems for regular op-amps. Can I tie all 3 buffer outputs together before the sense resistor or should I use a resistor for each one and tie them together afterwards? Or do I have to build 3 entirely separate circuits altogether and tie them at the output? There is not much power going through the first two stages but I am not sure how to handle the feedback circuitry.

Someone, help out a Mech E!  Thanks!

  • 0
    TI__Genius 15975 points

    Hi Ilya,

    The LM6172 could be used for your application and add an RC filter across feedback loop around A2 and A3 to compensate the loop. I would be concerned about stability because for higher output current drive, the device could get significantly hot if unstable and result in burn-out issues. Also, would THS4052 be the best fit for your application? I am referring to THS4052 because it has lesser bandwidth than LM6172 and satisfies slew rate and bandwidth requirement for your application. Also, the A2 bandwidth needs to be approximately four times lesser than A3 bandwidth for stability purpose (ofcourse you can compensate the loop with LM6172) which with THS4052 might be an easier constraint on the buffer amplifier?

    Best Regards,

    Rohit

     

  • 0 in reply to Rohit Bhat
    TI__Genius 15975 points

    Hi Ilya,

    Also, what is the output current requirement from the buffer for your application? One other thing was that is the input signal bandwidth for your application 20MHz or is it that you are looking for the op-amp f-3dB bandwidth to be 20MHz? Inorder to achieve higher output current drive than the BUF634 which is a fixed gain buffer, it might be advisable to use a discrete op-amp as unity gain buffer which has higher output current drive such as OPA2613.

     Regarding your question about tying outputs of 3 buffers in parallel together, I would not recommend doing that because of the current source & sink issue at the outputs due to output voltage mismatch. Please refer to this blog, if you are interested in paralleling three output buffers together. http://e2e.ti.com/blogs_/archives/b/thesignal/archive/2013/03/26/paralleling-op-amps-is-it-possible.aspx 

    Best Regards,

    Rohit

     

  • 0 in reply to Rohit Bhat
    Prodigy 60 points

    Hi Rohit,

    first, apologies for not seeing figure 3 in the BB tech note, showing the RC compensation circuit around A2 and A3.

    Second, isn't the OPA2613 slew rate 70V/microsec? I need 100 if I'm going to use a buffer output stage. Also, I am hesitant to go to surface mount devices. It's one more thing to learn that doesn't contribute to my thesis. The same is the reason for using lm6171 instead of something like THS4052 although I see your point. Also, thermal issue drove me toward the TO-220 package of the BUF634. I'll be running this thing continuously for hours so I need something I can attach to a (very) large heatsink.

    Regarding the bandwidth question, I don't care about the gain flatness as I am running a single frequency. Really, if I had a high power function generator with a current limit, that would have sufficed. So some roll-off at 20MHz is just fine since I can always turn up the gain on the input signal.

    Regards,

    Ilya

  • 0 in reply to ILYA LISENKER
    TI__Genius 15975 points

    Hi Ilya,

    I apologize for my overlook of OPA2613 slew rate being less than 100V/usec. In that case I would suggest using the OPA2674/OPA2677 as unity gain buffer at the output of THS4052. I understand that surface mount devices could be hard to learn but you could still use a large heatsink on the top of OPA2674 or OPA2677 to limit the junction temperature of the device at extreme output currents. One of the advantages for surface mount devices is contact to external interface circuitry is smooth and depending upon the package, thermal conductivity can be achieved on the back of the part through the pcb like the powePAD package.

    I was concerned about the bandwidth because usually the harmonics might degarde significantly depending upon how much you gain up in the high power function generator.

    Best Regards,

    Rohit