My MOSFET amplifier designed for music

Founder of XSA-Labs
Joined 2012
Paid Member
Hugh,

The snubbers I removed were C-RC snubbers across each secondary of the mains transformers. The philisophy behind this can be found in the Quasimodo thread:
Simple, no-math transformer snubber using Quasimodo test-jig

It's funny how the mind is pleased by tweaking a pot and observing a 'scope trace for the "perfect" resistor value which will kill all the switching noise. I was convinced that this had to result in better sound so I went through my entire system adding snubbers... but after a while I realised that some female vocals had some sort of colouration. This was most noticeable with for e.g. Alison Krauss and Patty Griffin, when the vocals soar up they also tended to harden up. It's a similar sonic effect to lens flare with cameras.

Anyway... long story short... removing each snubber improved this and also made a distict improvement in clarity and tonal balance - this is why I originally thought the Mooly amp was a bit bright.:ashamed:

Just another cul-de-sac amongst the hundreds I've bumbled around in during the course of this daft but mostly harmless hobby :D

This is an interesting, and at the same time, very puzzling observation. How does a snubber at the front end of a CRC, affect the perception of the sonic signature being "too bright"? It just makes the power supply smoother.

I have also installed snubbers in many places on my PSU's - now wonder if they are altering the mids and highs negatively.
 
Aha like 2SK1058/2SJ162, yes i have some VSSA.
Shaan pcb early designed. :)

Why don't you take a risk and build this one.... its my best pcb design yet...
and also this design is specifically for music.... aren't we all in this hobby for music?
BTW, I have shipped pcbs I had promised (some cap-mx and jlh amp pcbs)
 
Last edited:
This is an interesting, and at the same time, very puzzling observation. How does a snubber at the front end of a CRC, affect the perception of the sonic signature being "too bright"? It just makes the power supply smoother.

I have also installed snubbers in many places on my PSU's - now wonder if they are altering the mids and highs negatively.

It could be as well be perception of music heard earlier with old hardware/ system. I would believe diode ringing removed from music would only improve the experience. Sorry Dave, these are just my thoughts.....
P.S.
I seem to have a habit of getting involved in many controversies, may be its best left to guys with experience.....
 
Last edited:
Founder of XSA-Labs
Joined 2012
Paid Member
Why don't you take a risk and build this one.... its my best pcb design yet...
and also this design is specifically for music.... aren't we all in this hobby for music?
BTW, I have shipped pcbs I had promised (some cap-mx and jlh amp pcbs)

I admit it Prasi, I am having wandering amp eyes again, and that is is a fine looking board - bravo!

Makes me want to build it too after seeing the 3d renders! Using 4 latFETs (especially, discontinued ones) makes this a tough sell for me, personally though.

Beautiful to look at nonetheless...

706880d1538490629-mosfet-amplifier-designed-music-3d3-compact-png


Congrats on an excellent looking layout.
 
I admit it Prasi, I am having wandering amp eyes again, and that is is a fine looking board - bravo!

Makes me want to build it too after seeing the 3d renders! Using 4 latFETs (especially, discontinued ones) makes this a tough sell for me, personally though.

Beautiful to look at nonetheless...

Congrats on an excellent looking layout.

Thanks x,
yes i understand its a tough sell, to use 4 of the costly mosfets.
 
This is an interesting, and at the same time, very puzzling observation. How does a snubber at the front end of a CRC, affect the perception of the sonic signature being "too bright"? It just makes the power supply smoother.

I have also installed snubbers in many places on my PSU's - now wonder if they are altering the mids and highs negatively.

To avoid confusion:
I don't use CRC power supplies in my amps. By that, I mean e.g. 10KuF - 0.22R (or an inductor) - 10KuF on the supply rails AFTER the rectifiers. I find the series element reduces dynamics.

What I was writing about was a C + RC snubber BEFORE the rectifiers.

I agree that the Quasimodo setup does give a nicely damped waveform when there is NO LOAD CONNECTED. But maybe this is not the case when supplying some (not so) nice spiky currents into the supply caps??
in the end, the Quasimodo tweaks only an isolated part of the system.

Once again, sorry of OT. Maybe this should be in the Quasimodo thread?
 
Last edited:
Ah, BB, very droll........
I think it is a reference to 90% of the SS amplifiers in the marketplace which were not necessarily designed for music, but rather for low THD.

Study the design. It is different to most, and there are tricks in it which are interesting. The amp is designed for a monotonic declining harmonic.

HD
 
I have now reverted to the original configuration (albeit with reduced gain using 22K and 1K feedback resistors).


The high gain is supposed to be part of the "designed for music". You can make two options, one with 1K the other one with may be 680, and listen for these variations. The optimum result will depend also on your speaker design.


And the OPA604 preamp is also part of the "designed for music" :D
 
Ah, BB, very droll........
I think it is a reference to 90% of the SS amplifiers in the marketplace which were not necessarily designed for music, but rather for low THD.

Study the design. It is different to most, and there are tricks in it which are interesting. The amp is designed for a monotonic declining harmonic.

HD

Like this?
 

Attachments

  • Untitled.png
    Untitled.png
    193.1 KB · Views: 663
very impressive, Bimo! Very low THD, congratulations........

BUT, it is not the THD which is remarkable. It is the monotonic decrease which impresses. Straight line, linear. An accomplished amplifier!

HTF did you do this? I think John Curl would be very impressed with this too. Not to mention NP. Perhaps you should ask Pass Labs for a job!

HD (high distortion)