• State
  • Locked Locked
  • Replies 3 replies
  • Subscribers 47 subscribers
  • Views 2201 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.

OPA548 Non-linearity?

Chris Doughty
Intellectual 295 points

Other Parts Discussed in Thread: OPA548, OPA551

Hi,

I'm using a OPA548 to generate a 100kHz sinewave at variable voltage from 100mVp up to about 11.5Vp.  The output current ranges from 0 - 200mA and will be current limited (although it isn't currently).  I am feeding the amp a (nearly) pristine sine wave and am asking it to give me a gain of two.  E/S is left floating (measured at ~ -11.5V) and ILIM is being driven by a DAC but is set to limit at 3.6A during my testing.  At very low voltage (<1V) and at no-load, the output looks fine.  However, when I start to add load, a glitch appears on the falling edge (and to a lesser extent on the rising edge). 

I have tried a number of things to mitigate the problem but, thus far have had little luck.

  1. Add feedback capacitance (up to 47pF) across R52.  This seems to soften the edges on the glitch slightly but also starts to attenuate my output waveform.
  2. Add snubber (10ohms + 0.01uF) to ground at R43.  This seemed to make the problem slightly worse.
  3. Increase decoupling caps (10uF).  No improvement.
  4. Remove the current sense transformer from the output.  No improvement.
  5. Add 0.01uF to ILIM.  No improvement.
  6. Increased R47 to 5ohm (max allowed by my application) and added 47pF between U4:6 and U4:2.  No improvement.

I'm running out of ideas pretty quickly.  Anything you geniuses out there can come up would be greatly appreciated.

Thanks,

Chris

  • 0
    TI__Guru* 93105 points

    Hi Chris,

    It appears that you are pushing on the upper useable end of the OPA548's large signal operating range. The OPA548 unity gain bandwidth is around 1 MHz and it only has about 20 dB (10 V/V) of open-loop gain left at 100 kHz. I am having a little difficulty seeing all the details of your schematic, but it appears you have it set up for a non-inverting gain of +3V/V. At 100kHz the loop gain is low and the negative feedback is insufficient to correct for the amplifier distortion. I suspect what you are seeing is the cross-over distortion of the output stage. Loading the stage increases the output current demand and I expect doing so worsens the cross-over distortion.

    There isn't any way around the bandwidth limitations of a particular amplifier. I suggest considering an amplifier with higher unity gain bandwidth and slew rate specifications. If your maximum output current requirement is 200 mA, the OPA551 (DDPAK package) could be a better choice.

    Regards, Thomas

    PA - Linear Applications Engineering

  • 0 in reply to Thomas Kuehl
    Intellectual 295 points

    Thomas,

    Thank you for your quick response!

    Your read of the schematic was pretty close; I've been working with a gain of 2x (although have tried 1x as well).  

    What is a good rule for the amount of open-loop gain you need to keep amplifiers in this category behaving well?  I'd always assumed you could run them right up to their limits.

    I'll look into the faster part.  For now, I'm working with a legacy PCB and may not be able to do a redesign.  Would you have any recommendations for things to try to get a little more out of the part?   E.g. negative feedback instead of positive, additional feedback/feed forward gain by adding another op-amp, etc.?

    Thanks, again.

    Chris

  • 0 in reply to Chris Doughty
    TI__Guru* 93105 points

    Hi Chris,

    Usually, I aim for at least 40dB more open-loop gain (Aol) than the closed-loop gain (Acl) at the frequency of interest and that only insures the gain error will be less than 1%. The higher Aol is with respect to the Acl the greater the negative feedback applied resulting in ever lower gain error and distortion.

     Your application really requires a faster, wider bandwidth amplifier than the OPA548. Any circuit variations to increases the bandwidth which still involve the OPA548 likely won't get the performance up to the level you need.  When I search the power operational amplifiers and high output current, high-speed operational amplifiers relative to all the application requirements the OPA551 comes closest to satifying them.

    Regards, Thomas

    PA - Linear Applications Engineering