Part Number: TPA3255
Tool/software:
Hello,
I’m planning a to build a hi-fi audio design based on the TPA3255, and I’m aiming to squeeze the absolute best performance out of this chip!
I'm currently focusing on the input coupling capacitor distortion. After some research, here’s what I’ve found regarding the tradeoffs:
- Aluminum Electrolytics: Sensitive to aging and temperature variations, and the distortion becomes higher if the cutoff frequency isn’t at least two decades below the lower audio bandwidth (0.2 Hz).
- MLCC (Class II): Suffer from the microphonic effect.
- Class I ceramics (C0G/NP0): Excellent, but not available in the required µF range.
- Film capacitors: Great for audio, but bulky, expensive, and becoming harder to find (e.g. WIMA has discontinued many of their series).
From what I’ve read in older E2E threads, an input capacitor is mandatory on the TPA3255 due to an internal DC bias (which I assume is around 0.5 × AVDD?). So even if I use a dual-rail preamp, this internal bias would still couple into the preamp output if not blocked.
My question is:
Is there any clever analog approach to avoid placing a large electrolytic directly in the signal path? I’m exploring ideas like using a DC servo or a bootstrapped/floating resistor, basically, any technique that allows replacing the large coupling cap with something smaller, maybe placed in a feedback path of an op-amp stage idk.
I came across a few documents related to this concept:
TI Application Report: Slyt226
Patent US5568561
THAT Corporation Application Note: AES6261
Also, does PFFB (Post Filter Feedback) help with this issue in any way?
I would be grateful for any ideas, especially from those who worked on the original TPA3255 IC design, as they'd have the best imagination of what's happening inside the black box! A very detailed response would be awesome :)
Thank you very much!
