• State TI Thinks Resolved
  • Locked Locked
  • Replies 6 replies
  • Answers 1 answer
  • Subscribers 29 subscribers
  • Views 591 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.

LME49720: Capacitive load tolerance for audio application.

Jim Adams1
Prodigy 50 points
Part Number: LME49720
Other Parts Discussed in Thread: LM4562

I see strong recommendations for a “rule-of-thumb” 50-100 ohm R be placed at the op amp output to isolate a CL of >100pf or the square wave step response will ring, oscillate and/or overshoot badly. Is this still a worry when the op amp signal is not a fast-rising edge but instead is a sine wave audio music signal which could be 20KHz or something above that? There might be some music elements that are fairly fast-rising transients but nothing like a 10V/uS step. Question: I have a 1,200pF cable load; the application is music audio; the op amp gain is -1; the op amp is a “… specifically for high-fidelity audio” type (as above or the like) and am using a 100Ω isolation R at the output pin. The specific application is a state-variable realization of the Linkwitz-Riley 4th-order, 2-way electronic cross-over where the high-pass and low-pass outputs are driven directly by op amp sections inside the active state-variable loop. In other words the low and high-pass cable outputs are taps off of active filter elements without output buffers. Those two op amp sections are doing double-duty as in being part of the filter loop and driving 1200pF cables via 100Ω isolation Rs.

Obviously this filter design cannot tolerate ANY kind of misbehavior of any kind at those two op amp output pins so for audio spectrum sine wave duty will the op amp stability be unaffected?

  • 0
    TI__Guru* 84915 points

    Hi, Jim,

    Welcome to E2E and thank you for your interest in our products!

    Using a capacitive load above 100pF would be always a risk of having instability. So, it is important to select a correct series resistor to be placed before the capacitive load. As you mentioned, this resistor helps to isolate the capacitive loads of high values.

    In order to deal with the high capacitance value, it is necessary to use external compensation techniques. Usually, a series resistor from 5ohms to 50ohms should be enough to avoid instability. If the capacitive load and the PCB capacitance is high, a snubber (R and C array from output to GND) can be connected for low voltage operations. This method depends of the internal structure of the op amp, but in general, you would need to try with different resistors (from 60 to 200ohms) to see how the peaking is reduced. The capacitor is then calculated as C = (3/2) * pi * fpeak * R; where fpeak is the frequency at which the frequency peaking occurs.

    I hope this helps you. Please let me know if you have additional questions or comments on this.

    Best regards,
    Luis Fernando Rodríguez S.

  • 0 in reply to Luis Fernando Rodriguez S.
    Prodigy 50 points

    Thanks Luis.  My specific question had to do with relatively slow-rising/falling op amp signals driving high-C loads vs. what the datasheet scope-shots show as examples of very high frequency step responses which is not the case here.  I posted the question because this particular LME49720 is billed as a "High Performance, High Fidelity Audio Operational Amplifier" on the one hand and on the other hand watch out, this device would be in trouble driving just 4 - 5 feet of audio interconnect cable.  But the example of being "in trouble" is characterized by misbehavior resulting from a step input into a CL of 100pF but a step doesn't ever happen in the audio spectrum.  The example is not germaine - or is it?

    Maybe you or another could think in terms of a 20KHz sine wave edge as the signal source which may or may not over-excite the internals of the LME49720.  That would be great info.

  • 0 in reply to Jim Adams1
    TI__Guru* 84915 points

    Hi, Jim,

    Unfortunately, we don't have much details on the internal structure of this particular family of devices. However, the LME49720 has been designed to work with signals from DC level to high frequency content. Even the device offers an excellent DC performance in case it is being used as single DC amplifier. So, it shouldn't have issues for your particular application.

    Please let us know if you have additional questions or comments on this.

    Best regards
    Luis Fernando Rodríguez S.

  • 0 in reply to Luis Fernando Rodriguez S.
    Prodigy 50 points

    The LME49720 is a TI part but maybe you or others know it better by the original PN of LM4562.  They are identical op amps.  Chatter is that TI wanted to include it into the LME family with new PN but big customer refused to change their inventory PN designation.  Thus LME49720 = LM4562

    Please try again. 

  • 0 in reply to Jim Adams1
    Prodigy 50 points

    I did some testing on this subject of resistive isolation for op amps with capacitive loads.

     

    Equipment was:

    Dual-trace 100MHz storage scope

    Sine/Triangle/Square-wave function generator.

    Linkwitz-Riley 4th-order, 24dB/octave 2-channel cross-over circuit powered by ±15V.

     

    TEST #1:

    1) Set the function gen to 40KHz sine-wave, 6Vp-p and connected that into the cross-over Right channel input.

    2) Made up a RCA male plug with a parallel combination of 1500pF and 20KΩ as a load and plugged that into the Right channel high-pass output.

    3) Attached Channel 1 scope across the RCA male output loading plug, #2.

    4) Attached Channel 2 scope across cross-over Right channel input.

     

    I then adjusted the scope’s Channel 1 vertical and horizontal such that the 6Vp-p 40KHz zero crossing was expanded and filled much of the screen. I then used the time and voltage cursors to determine the slope or slew of that smaller portion of the full 6Vp-p sine-wave high-pass signal output

     

    What I measured was 760mV/µS as the edge slope or the rate of slew or what have you. This measurement was taken from the high-pass output, across the R + C load and driven by a LME49720 (aka LM4562) cross-over element with 100Ω isolation in-series with the RCA output load. Comparing to the input I saw no aberrations of any kind.

     

    TEST #2

    1) Set the function gen to 100KHz square-wave, 20Vp-p and connected that into the cross-over Right channel input.

    2) I used the same 1500pF // 20KΩ RCA loading plug on the high-pass output.

    3) Attached Channel 1 scope across the RCA male output loading plug, #2 as before.

    4) Attached Channel 2 scope across cross-over input as before.

     

    What I saw was:

    1) The function gen’s input traversed 18V of its 20Vp-p leading edge in 72nS or the equivalent of 250V/µS.

    2) The C-R loaded and 100Ω isolated LME49720 high-pass output traversed 18V of the 20Vp-p leading edge in 1.06µS or ~18V/µS with visible roll-off slanting and rounding of leading and falling edges but no ringing or overshoot or other misbehavior.

     

    I was pleased with these outcomes and especially pleased and surprised with the unusual, non-audio challenge of TEST #2.

  • 0 in reply to Jim Adams1
    Prodigy 50 points

    Since I posted the above, I now have learned that the TI LME49720 and its twin the LM4562 both suffer from "pop corn" impulse noise.  At least one potential high-volume buyer designed the LME49729 out of his 500,000pc/year project b/c the % of bad parts he was seeing made screening and over-buying a non-solution.  He was forced to find an alternative source.

    Very interesting.  A good read.

    http://e2e.ti.com/support/audio/f/6/t/725471