• State
  • Locked Locked
  • Replies 3 replies
  • Subscribers 31 subscribers
  • Views 886 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.

TPA3251: Questions on power supplies and decoupling capacitors

Carlos Moreno1
Intellectual 550 points
Part Number: TPA3251

Hi,

Question 1:  Are the two output power supplies (PVDD_AB and PVDD_CD) completely independent?  For example, can I use two different voltages  (say, I want AB with 35V for the woofer and CD at 30V for the mid/tweeter that has lower power capability)?

Corollary:  are the three pins for each of the above two (three pins for PVDD_AB and three pins for PVDD_CD) also independent?  For example, can I connect three power supplies of the same voltage, one to each pin, to add the power capabilities of the power supplies?

Question 2:  What are the recommended types of decoupling capacitors?  I'm highly reluctant to use X7R (due to microphony) in audio applications --- even for power supply decoupling.  However, the NP0 caps only get to 0.1uF in 1206 size  (which still fits, but it's a bit uncomfortable to place).  I'm inclined to believe that it is better to place 0.01uF NP0 rather than 0.1uF or 1uF X7R --- what do you think?  For the output stage (the PVDDs), would it be better a combo of 0.01uF NP0 next to the pin and a 1 or 2 or 10uF tantalum, say, 1cm away from the chip?

For the input side (the DVDD, AVDD, GVDD_xx, VBG pins), I'm inclined to think that 0.01uF NP0 is ok --- what do you think?

Notice that the assumption is:  for decoupling capacitors, surface mount is better (avoid making an extra hole on the ground planes), and smaller size is better (thus, placeable very close to the pin).

What's your advice?

Thanks,
Carlos
--

  • 0
    Intellectual 860 points

    Hi Carlos,

    I can't speak to your power supply independence question, but in regards to X7R vs NP0 decoupling caps, my experience has been that the microphonic characteristics of X7R (or X5R) are not an issue, even in high performance amps. I certainly wouldn't use them in the audio signal path due to well-known distortion mechanisms of that application, but in a power supply, but I don't think the tradeoff of eliminated microphony would be worthwhile versus the impact on your power delivery network design. A closed-loop part like the TPA3251 does have decent PSRR (65dB in this case), and given that any microphonically-induced anomalous voltage in the supply is likely to be induced by audio bandwidth vibration mechanically coupled into the caps, it will be attenuated to some degree by the PSRR. As you noted, NP0 caps have much larger packages, and depending on the form factor, your bypass caps will have significantly higher mounting and loop inductance, thus potentially compromising EMI performance. Personally, I would not use something as small as .01uF to decouple the supply of an amp like this. Such a value is likely (depending on what you have it paralleled with) to produce a strong antiresonant peak in the power supply's impedance versus frequency plot.  You could also apply mechanical damping to your PCB if you want to extract the last iota of performance.

    Just my $.02.

    Josh

  • 0 in reply to Josh Heiner
    TI__Mastermind 27990 points

    Hi Carlos,

    Question 1: I have spoken with our team and determined that there are internal references that use PVDD as a supply, and then assume both outputs to be at the same voltage. This will cause unknown behavior on the output, therefore supplying two different voltages is not recommended.

    Question 2: I agree with Josh in that using X7Rs for supply decoupling will not have an effect on audio performance, and with their size are easily placed near the pins, which is a requirement, as too much distance from this pins can cause ringing/overshoot of the supply voltage after large power draws occur, and will damage the part. We use X7Rs on our boards for supply decoupling and get the data sheet rated performance.

    Justin

  • 0 in reply to Justin Bohr
    Intellectual 550 points
    Thanks Justin,

    Follow-up on Q1 --- is it that the PVDD_AB and PVDD_CD pins are internally connected, or is it that the internal circuits use the signals and rely on the two voltages being the same? The distinction is relevant if, for example, I want to use two power supplies of 100W each to get a total power input of 200W. Physically connecting the outputs of the two power supplies may be a problem for the power supplies.

    However, I wonder what happens if I do that (connect two independent power supplies with identical voltages), but then one amplifier is drawing more power and it makes its corresponding PVDD_xx go down a bit (maybe a volt or a few volts?). Would this difference now cause any trouble?

    Thanks,
    Carlos
    --