• State Resolved
  • Locked Locked
  • Replies 4 replies
  • Subscribers 31 subscribers
  • Views 724 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.

TPA3244: High-Frequency Operation

Evan Cornell
Expert 1410 points
Part Number: TPA3244
Other Parts Discussed in Thread: TINA-TI, TPA3255

I see on the first page of the TPA3244 datasheet that the device supports "Signal Bandwidth up to 100 kHz...". 

I have an application that requires passing more sinusoidal-type signals close to that 100kHz frequency. However, I don't see any characterization of performance at these high frequencies in the datasheet. Is there any such data, to see if there is any rolloff at high frequency?

When using the LC Filter Calculator excel tool, I can say I want a cutoff freq. of 200kHz, which allows for acceptable filter behavior at 100kHz, but the inductance required is ~2.25uH, which appears to be below the recommended minimum inductance of 5uH, per TPA3244 datasheet. Is there some way to safely operate TPA32xx with lower than recommended inductance values?

My load does not necessarily have a flat frequency response, thus the "peakiness" that I see at various load impedances with the chosen LC filter values are of concern. Would Post-Filter Feedback help to mitigate this issue, and if so, can it be simulated in TINA-TI, along with a model of my actual load?

Since I would like to operate at a much higher than typical audio frequency, the LC filter cutoff frequency necessarily has to be higher than typical with TPA32xx applications. Would it be beneficial to operate TPA32xx in AM2 mode (highest switching frequency), in order to get more attenuation of the switching frequency components on the output?

Thanks!

  • 0
    TI__Mastermind 24835 points

    Hi Even,

    We are lookin into this, and Robert will get you back soon later.

    Best regards,

    Shawn Zheng

  • 0 in reply to Shawn Zheng
    TI__Mastermind 27980 points
    Hi Even,

    Is there any such data, to see if there is any rolloff at high frequency?
    -We currently don't have test data out to 100 kHz. There will be some rolloff before 100kHz, but I would have to get some more information from design to be specific.

    Is there some way to safely operate TPA32xx with lower than recommended inductance values?
    -The recommended 5 uH is to ensure the robustness of our integrated protection. If the inductance falls below 5uH during operation the over current protection feature will be at risk of failing.

    Would Post-Filter Feedback help to mitigate this issue, and if so, can it be simulated in TINA-TI, along with a model of my actual load?
    -We currently do not have a model of our TPA32xx family of devices in TINA, we are working with our modeling team to create it. Once we have this it could be modeled in TINA, and the PFFB circuit could be modified and simulated. That being said, PFFB can help or hurt the peak near the FC of the circuit. Our current passive PFFB reference design actually adds to the peak at 40 kHz cutoff (see page 18 of www.ti.com/.../slaa788.pdf).

    Would it be beneficial to operate TPA32xx in AM2 mode (highest switching frequency), in order to get more attenuation of the switching frequency components on the output?
    -Yes.

    Justin
  • 0 in reply to Justin Bohr
    Expert 1410 points

    Justin,

    Great, thanks for the information. 

    A few follow-up questions. 

    1. If I were to test the TPA32xx EVMs for high frequency response at say, 5-10W, would a TPA3255 operated at 24V give similar results to TPA3244/45/50/51 at the same test conditions? I am sort of assuming that the control loop / high frequency response are all similar between this family of parts. 

    2. For simulation, I will look forward to TPA32xx model. Is there any timeline on this?

    3. For simulation of the LC filter response and the load, is it necessary to simulate both legs of the LC filter, or just one L + grounded capacitor (half the value of the differential filter cap.)?

    Thanks,

    Evan

  • 0 in reply to Evan Cornell
    TI__Mastermind 27980 points
    Hi Evan,

    1. They would be similar but their design s are not 100% identical, so there might be slight variations in response.

    2. We hope to have this completed by the end of February.

    3. You should be able to just simulate one half bridge, the difference between should be minimal.

    Justin