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.

Why the circuir can't working normally when the input sine wave's operation frequency fs>8Khz

Other Parts Discussed in Thread: LM2902, LM324, TL3474, TL084, LM358, LM6584, OPA4322, OPA322

Dears,

I'm using LM2902 to realize the presicion rectifier circuit . But found when the input sine wave's operation frequency is higher than 8KHz, there has a distortion at the output side (point A).

And also the enevlope tracking circiut can not track the peak value, the voltage at point B is getting lowerr and lower when the operating frequency is getting higher. ? Could you recommended an Amplifier that has better performance than LM2902, and they are should pin to pin and easy to use, cost can be higher than LM2902, but not so expensive.    

And by the way, could you help to review the circuit and point out some problem during the usage? 

Best regards,

Helen

  • I check the schematic in detail, found the main problem is the usage of U1, during the simulation, if I use LT1358 instead of the LM324, then the envelope tracking circuit works quite well till 30KHz, as the  input signal is a 20Hz-20Khz sinewave, so it is OK for the use. 

    Here my question is  "which parameter" affect the working of this envelop tracking circuit?  SR or some other parameter?

    Could you please recommend one "TI amplifier " which is  fit for use?  And it can worked with single power supply.  And it is not so expensive.

    then I'll redraw the schematic and PCB for the customer, hope to get your reply as soon as possible, this is really urgent. Thanks for the kindly support!

    BR,

    Helen

  • Helen,

    At low frequency the amplitude is limited by the VOH and VOL data sheet specifications. At some frequency the large signal response will be limited by the device’s slew rate. If signal distortion was no concern, the output could provide a triangle wave output amplitude versus frequency of A=SR/(2*F), where “A” is peak to peak output amplitude, “SR” is slew rate (V/S), and F is frequency (Hz). If distortion is a concern, the formula can use sine wave’s maximum slope at the ‘zero crossing’ of the input sine wave. This changes the formula to A=SR/(F*pi), where pi=3.14.

    Typical slew rate is 0.5V/us = 500,000V/s, so low distortion 20kHz maximum Vpp = 500,000V/s / (20,000Hz * 3.14) = 0.8Vpp

    To complicate matters the diodes in the first stages will cause a response delay also due to slew rate.

    A faster op-amp is needed. The LT1358 is a very fast device, much more than you need.

    10V/uS  slew rate will be fast enough. Try TL3474, it has a model too. www.ti.com/.../toolssoftware

    If you use a quad device all four op-amp will need to share a single +VCC and -VCC

  • Hi, 

    Thanks for the quick reply. 

    So I'll choose TL3474 in my application. By the way, as we have TL084 on hand, can we also use TL084? since its slew rate is the same to TL3474. What is the main difference between TL084 and TL3474?

    Best regards,

    Helen

  • Hi,

    I simulate the circuit with TL3474 found that the minimum output voltage is 550mV with 5V single power supply, the circuit seems can't working normally.
    The result is even more bad than with LM2902.Why? Which parameter shows its minimum output voltage ability?
    The maximum sine waveform is change from 20Hz to 20Khz, its amplitude is change from 0-0.3V. Is the 0-0.3V input amplitude too low for TL3474?

    The result of TL084 is also not good.

    BR,
    Helen
  • So it seems that we need an amplifier with (1)very low output voltage VoL ability, 20mV like LM358 and LM2902, (2)but also with high slew rate. Could you help to recommend one?
    Thanks!
  • Helen,

    Yes you are correct. I thought I saw a negative power supply in circuit.
    For a single 5V supply there are many op-amp available.

    www.ti.com/.../precision-amplifier-products.page
    Note link might not work unless copy and pasted manually.

    LM6584 and OPA4322 are lowest 1k price for 5V, rail to rail input and output with >=10V/uS slew rate.
  • Helen,

    TL3474 has a emitter output so VOL and VOH do not go to supply rails. Same with TL084.
    TL084 also does not allow input near VCC- rail.

    LM2902 does have low VOL, but it is very weak. It can only sink 30uA typically.
  • Ronald,

    Thanks for the reply.I'll try OPA4322. 

    As I'm also want to use dual power supply if possible, so I simulate the TL3474 ith +-5V dual supply. The maximum voltage comparison circuit works well, but the enevlop tracking circuit seems can't tracking the peak voltage.

    The maximum voltage is 600mV, but the envelop tracking results is only 535mV, as you know, I use LT1518 in the simulation circuit before, it can tracking the 600mV peak very well. So whats wrong with my circuit? Or I have to use a much quicker circuit than LT3474? 

    Do you have some better dual or quad amplifier to recommended to me with dual power supply? What's your email adress, there's something wrong on the web page, I can't insert picture into the file. I will send the picture to you by email.

    BR,

    Helen

  • Thanks,

    OPA4322 can't be used in the LTSpice, so could you help to simulate the circuit for me? What about its performance of envelop tracking with 20KHz sinewave?

    BR,

    Helen

  • Hi Ronald,

    OPA322's Spice model works well in my circuit. Its performance is also very good!

    Thanks for your kindly support!

    But I still can't unstand why LT3474 works not well with dual power supply? It can't tack the peak voltage, its strange? Is it because of the Bandwidth? For sine waveform, is bandwidth important?

    BR,

    Helen

     

     

    BR,

    Helen

  • Helen,

    I think the issue is caused by the output going below ground. The lower the output voltage, the farther the voltage must go to turn the diode back on.

    A single supply won't reverse bias the diode as much so the output voltage can turn on the diode faster.
  • Thanks Ronald.

    So, does it means my circuit is not fit for the dual power supply application? Because I'm worried my customer will alo use a very cheap amplifier (with +- power supply), then the result goes wrong, I'll explain this to the customer and persuade them to use OPA4322.

    BR,

    Helen

  • Hi Ronald,

    Whats a pity, the maximum supply voltage in some corner case would be 6V, but OPA4322's maximum supply voltage is only 5.5V, could you help to recommended one amplifier with higher than 6V supply voltage? And its VoL should be as low as 0-50mV. But its performance is similar to OPA4322/2322.

    LM6584 is not fit, because its output voltage swing low value is 0.5V typical.

    Thanks for your kindly support!

    BR,

    Helen

  • Helen,

    I finally found time to simulate the LM324 myself to see exactly where the problem is located.

    It seems the problem only occurs when the diode duty cycle is low. Notice the that buffer peak output goes to 600mV on the first pulse then it finally settles out near the filtered output at 552mV.  

    This week I will test in lab to see if this is real or a simulation error.

    Do we really get matching in phase and inverted input signals? What is the amplitude range for these signals?

  • Thanks Ronald,

    The maximum amplitude is 300mV sine wave, like what you simulated with. The range is a from 0-300mV sine waveform input, and we hope to get a 50mV(0)-600mV envelop output signal.

    The phase of the sine wave is can be at any angle, we just need to get its envelop signal, so the phase is doesn't matter.

    The experimental result is much poor than what we simulated, I can send the waveform to you by email, because I can't attach the waveform in the forum  thes days.

    BR,

    Helen

  • Is your email address: <removed>, I have send the experimental result to you, the resullt is much much poor than what we simulated. It can't work normally even at the first stage, the Vmax signal is totally distorted at 20KHz, and it begin to distorted at >5KHz.

  • Helen,

    I will look at your results and reply about them tomorrow.

    My simulation shows that the peak detector will need to be a faster op-amp than the half wave rectifier op amps.