Please clarify. What data? I'm very interested cos the AVVD mod made a huge difference in SQ.
My Philips 9710 fullranger can't cope with more than 10W anyway and who needs 195W at home?
This statement makes sense) Because it's just a matter of knowing what distortion you get at X watts.
DrMordor & cdsgames working together, that would be an all star team I guess.
Dani I love you! Are you married?
LOL 2 kids bro !
You mean a dream team....
the ugly the bad and the good is missing
By the way listen to this, with good amp please )))
The Good, the Bad and the Ugly - The Danish National Symphony Orchestra (Live) - YouTube
Last edited:
ok just to clear things up , TNT has seen a message from me posting my FFT...on drmordor GB. Honestly I did not realized (poor excuse)and posted without thinking. Its below my standard(or anyones here) and understandable that TNT lost patience. I have asked a moderator to remove my post and it was done .
Drmordor , I apologize ! I will be extra careful from now on.
Regarding Avdd , since the beginning we understood that's a very important rail. I chose dedicated LDO to feed power to that rail. From what I remember , I read the noise on Cap pin (forgot what pin number) and experimented with different values of caps(tpa datasheet is substandard with ceramic small cap).220uF , 470 and 1000uF. I found a decrease on that rail (that most likely has the comparator and wave generator) from 23uV to 8uV. 220uF and 470uF were not enough I think that at 470uF we seen only 3-6uV of noise reduction.
(all this while playing a 1Khz file)
Further I crosschecked the reduction of noise with FFT and THD+N data. We see a small correlation. However I am convinced (by gut ), based on many subjective feedback's and the fact that analog to PWM conversion takes place on this voltage , that reduction of noise on that pin is VERY important for SQ. It just makes sense ...
In the end , I have used the LT3042 (or lt3045) LDO to power that rail along with 1000uF polymer cap(on decoupling pin) that gave us the best noise reduction..
Hence I was suggesting that 100 to 470uF might not be enough , but some caps are better than others in reducing noise.
Drmordor , I apologize ! I will be extra careful from now on.
Regarding Avdd , since the beginning we understood that's a very important rail. I chose dedicated LDO to feed power to that rail. From what I remember , I read the noise on Cap pin (forgot what pin number) and experimented with different values of caps(tpa datasheet is substandard with ceramic small cap).220uF , 470 and 1000uF. I found a decrease on that rail (that most likely has the comparator and wave generator) from 23uV to 8uV. 220uF and 470uF were not enough I think that at 470uF we seen only 3-6uV of noise reduction.
(all this while playing a 1Khz file)
Further I crosschecked the reduction of noise with FFT and THD+N data. We see a small correlation. However I am convinced (by gut ), based on many subjective feedback's and the fact that analog to PWM conversion takes place on this voltage , that reduction of noise on that pin is VERY important for SQ. It just makes sense ...
In the end , I have used the LT3042 (or lt3045) LDO to power that rail along with 1000uF polymer cap(on decoupling pin) that gave us the best noise reduction..
Hence I was suggesting that 100 to 470uF might not be enough , but some caps are better than others in reducing noise.
While I agree , need for power is part of us , all.We rather have more ,than not enough.
Yes our TPA gain is 10db but we made a front end (nothing like datasheet) using opa2156 (incredible opamp with great RFI rejection) and FDAs. The gain is selectable in 3 stages giving you (on max) 150W (like 0 , 5db 10db)
Further my LLC has also 3 voltages . 50V 42V and 37V. Gain of TPA3255 is in function of V therefor you can further tweak the power delivered .
On Monday I will show my unit. It looks very ugly since I have done many tests on it. As we spoke privately Dani , I will also show publicly my last great "trick" and hopefully convince you all , that it was SQ and not THD my main goal. Because of that pursuit I was delayed..
You will clearly understand ! In one pic.
My last PCB was sent to manufacturer for samples in 2 flavors. In one flavor , I used parallel planes (by the way I am using 8 planes as per EMI best practice , please refer to Otto book ) capacitance to try to reduce the noise on PWM. Not sure if it will work..second flavor is a small tweak of what we already have.
I would say 2 months to market , but I have been wrong before.