• State
  • Locked Locked
  • Replies 4 replies
  • Subscribers 30 subscribers
  • Views 1136 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.

Original question:

TPA3251: Post filter feed back issue

TPA3251: tpa3251, PFFB and pll mode

Florent charles
Prodigy 100 points
Part Number: TPA3251
Other Parts Discussed in Thread: TPA3255

Dear again I'm  not sure that the 10uH give more power than the 7uH, on my measurement it seems almost equal.

Could you just explain me this double pole just to be sure to understand?

When you decrease a capacitance in a filter, the cutoff frequency increase, right?

Then why on the PFFB application note the capacitance C_FBout is higher when the switching frequency is @ 600kHz than 450kHz.

I try to understand how this system work with this T filter.

Another thing I try is the different mode MST,AM1,AM2 without PFFB first, every circuit seems to work properly in his own base frequency.

When i try to use the AM2 i get some issue on my results.

TPA3251 with 7uH 680nF or 10uH 1uF --> Master mode perfect (600kHz)

TPA3255 with 7uH 680nF or 10uH 1uF --> Master mode perfect (450kHz)

TPA3251 and TPA3255 with 7uH 680nF or 10uH 1uF --> AM2 some THD+N issue vs output power, spikes

Is it normal? The LC filter with this value work only for the master frequency? is it needed to adapt the output filter with different mode?

  • 0
    Prodigy 100 points

    I just see that I make a mistake. I didn't see that the resistor value for the pwm mode is the same for both 3251 and 3255.

    It means that i never managed to make it working completely on a 600kHz with either the 7uh or the 10uh

    What i have is that when i get more than 10W rms i started to have some spike on the THD

    Nornally it's around 0.002%, i get some spike to the triple

    Is there any thing to change to make it work with the 600khz pwm ?

    Regarding the PFBB, do you have a calculator for the settings, because all measurement I have that works with the PFFB are all with 450kHz (Mater mode for TPA3255 and AM2 for TPA3251)

    Iget this instability on the THD way more with the PFFB. Also using the settings in the application notes work only for me with the 450kHz

  • 0
    TI__Mastermind 19080 points

    Hi,

    The C_FBout should have nothing to do with power inductor here. The LC double pole comes from L_out and Cout. I think C_FBin,C_FBout, and R_fb_gnd together built a T-type R-C network. The calculation could be complex, but I believe larger C_FBout decreases the feedback path gain, thus makes the closed-loop gain and also the bandwidth higher. With 600kHz switching frequency, you can use higher bandwidth.

    Do you mean that you've tried the values as the app note suggested, but only 450kHz values works for you? Did you try keep the 450kHz values for 600kHz working condition?

    PFFB could be quite complex that it's better to run some simulation. However, the problem is I don't have a model for these devices either...

    Thanks!

    Regards,

    Hao

  • 0 in reply to Sam(MPAA) Zhang
    Prodigy 100 points

    Do you mean that you've tried the values as the app note suggested, but only 450kHz values works for you? Did you try keep the 450kHz values for 600kHz working condition?

    Yes exactly, I used the value in the application note. And also I was thinking at the beginning that I was using the nominal mode 600khz on the TPA3251 because i let my jumper in the same position than with the 3255.

    But after looking at the datasheet i saw that i was still on the 450khz.

    Then the value in the application notes 3251 with cfb_out 1nF and rfb_gnd 2K is not working with 600khz. Only with 450khz.

    If I reduced the cfb_out to 670pf or 330pf I can kee a good thd  against output power. (without pffb limit @ 90W RMS, pffb 1nF @ 40W RMS, pffb 670nF @ 50W RMS, pffb 1nF @ 70W RMS)

    Let me know i f you dont understand.

    I don't understand the ratio on the settings between the 450khz and 600khz everytime the 600khz is used a 1nF is recommended but this seems to reduce the available bandwith no? like less output power linearity

    Do you know how TI get this value in the application notes ?

    Is there a way to write to the guy :-)

    I can also send you some measurement and graph if you are interested

    Thanks for you help and time

  • 0 in reply to Florent charles
    TI__Mastermind 19080 points

    Hi Florent,

    I can't find the authors either, they are not in our audio team any more... Is it possible for you to build a simple circuit to look into the feedback network in simulation tools like TINA? Simple amplifier model can be used rather than class D audio amplifier. I think basic idea here is you need to limit bandwidth to lower level for lower switching frequency.

    I'm curious about it too but really can't spare enough time on it... Any way, I will try to do this from my side at the same time. Let me know if you have any progress. Thanks!

    Regards,

    Hao