Experimenting with tripath output filter

[robyonekenoby]

Hello folks,

just playing around a TA2020 board to get rid of some residual harshness in top end and annoying sibilance

Swapping output filter caps didn't give me big results, so I tried something apparently weird: removed caps at all and leave only inductors

Well, surprise surprise, everything seems ok now, top end very extended, defined, and crystal clear, without any trace of sibilance, and also bass seems to have gained tightness and power

The idea comes from this paper http://www.tij.co.jp/jp/lit/an/slyt198/slyt198.pdf

Any considerations?

 

[Saturnus]

Don't expect it to last long as the police might knock down your door at any moment. You changed the filter from a 2nd order to a 1st order filter, so you've just made a fairly powerful RF noise transmitter.

 

[robyonekenoby]

very funny

Yes it's a first order filter, but with the same fcut frequency

Anyway, listening right now, the SQ if by far the best ever heard from a tripath, and I have also the trends TA 10.1 and DIY TA3020

 

[Saturnus]

very funny

Yes it's a first order filter, but with the same fcut frequency

No. You moved the Fcut down by 1.414 so it now starts well within the audio frequency band and basically cuts away much of the top end.

You think it's funny? TA2020 is spread spectrum from 100KHz-1Mhz. That means you're sending out RF noise at aircraft navigation and emergency frequencies as well and AM radio frequencies.

 

[theAnonymous1]

With the capacitor....

Without....

 

[robyonekenoby]

Yes I use the same web tool; for a given R load the fc is same

 

[nigelwright7557]

Its a trade off between higher audio frequencies getting through but not the carrier signal.

 

[theAnonymous1]

From what I've seen with most designs, you want to shoot for around -40dB at the carrier.

 

[MusicSeeker]

I had suggested the same cap removal in another thread one yar ago and it finally revealed a very natural sound with no more treble peak due to the interaction of the 2nd order filter and the real life speakers impedance.
The interference threat is wrong as if you consider the huge speaker impedance at the tripath operating frequency, the radiated power is insignificant to create harm to air traffic communications. I have an old valve AM receiver from 1957 in the same room, and you can not hear anything whatever the tuned fequency that covers easily the tripath ones