Building an ultimate low power class A amplifier – my way

Member
Joined 2019
Paid Member
giphy.gif
 
  • Like
Reactions: 3 users
Member
Joined 2019
Paid Member
LM329 is fixed shunt voltage reference at about 6.9 V, so not suitable without additional components. Noise at OPA828 supply rails is not a problem as capacitors (electrolytic + ceramic) would help there. Simple RC filter is enough to drop excess voltage and provide constant PSRR up to very high frequency. OPA828 has a very good PSRR up to 100 Hz – 1 kHz, then it starts to decline at standard 6 dB/octave. RC filter at supply rails helps to keep good PSRR performance. This amplifier has about 100 dB PSRR in the whole audio band.

Regulated supply for an opamp wouldn’t bring better distortion performance or sound, only convenience with various supply voltages. My goal was to keep amplifier design simple while achieving desired performance (sound and sound stage of this amplifier are incredible).
 
  • Like
Reactions: 1 user
I think JHL will still sound better and can be done with Lateral Mosfet output devices, while being simpler and operates near single ended. The opamp should be removed, it is the weakest link in sound quality. Just increase your pre-amp gain, and you would see a remarkable difference, because the op amp is not class A and produces all odd harmonics. That is the last thing you want in "high-end" listening experience, and would make the treble much more pronounced or harsh. It would simplify the PCB and save some components, as well as helping the layout be simpler and less prone to noise and rubbish floating around. Always keep in mind the simpler a design the better it sounds.
 
“My goal was to keep amplifier design simple while achieving desired performance (sound and sound stage of this amplifier are incredible).”

IMO, you met your goals Tombo :yes:.
I would echo @Vunce words as I am loving the amp and its going head to head with my Sissy SIT R3 in a smaller format

@tombo56 you have an incredible amplifier design
 
  • Like
Reactions: 1 user
Always keep in mind the simpler a design the better it sounds.
To a point , someone slammed me with this yesterday over extras like cap multipliers and cascodes. So , this statement is at best ...subjective.
As far as "how it sounds" ? At 20W , you would fail to hear the difference between this design or any similar diamond buffer (fast EF3-AB discrete) design.
The op-amp or EF3/lateral FET differences/choices might be marginally audible .... but only to "golden ears".
PS- I have a DB/EF3 AB design like this. They do sound quite good , you can hear fast....
OS
 
  • Like
Reactions: 1 user
LM329 is fixed shunt voltage reference at about 6.9 V, so not suitable without additional components. Noise at OPA828 supply rails is not a problem as capacitors (electrolytic + ceramic) would help there. Simple RC filter is enough to drop excess voltage and provide constant PSRR up to very high frequency. OPA828 has a very good PSRR up to 100 Hz – 1 kHz, then it starts to decline at standard 6 dB/octave. RC filter at supply rails helps to keep good PSRR performance. This amplifier has about 100 dB PSRR in the whole audio band.

Regulated supply for an opamp wouldn’t bring better distortion performance or sound, only convenience with various supply voltages. My goal was to keep amplifier design simple while achieving desired performance (sound and sound stage of this amplifier are incredible).
Thanks Tombo

it is just an idea to do easy because most of amp builde rwith Class A have the "normal" 23-25V rail.
chris
 
i can confirm that the OPA828 is very good. !!!!
we swapped opamps.
i had a hearing session with a BA3- Toshiba and M2OPS with famous Toshiba mosfet.
Source Auralic Vega G2, compare amp was AVM 6.3, LS KEF 11.

if i can say that Toshiba are similar to Lateral the sound is incredible good.
Tombos amp is on my todo list..

chris
 
  • Like
Reactions: 1 user
Member
Joined 2019
Paid Member
Uh, Nico, you missed the point on all accounts as you evidently didn’t look at the amplifier details and measured results. :cool:
Personally, the op-amp at the input is the weakest point.
I think that, opamp is the strongest point of this low power design.

Just increase your pre-amp gain, and you would see a remarkable difference, because the op amp is not class A and produces all odd harmonics.
That could be so many years ago, but modern opamps are far ahead of what you describe. It is of no consequence but opamp works in class A up to the clipping. Its output current is in the uA range and never exceeds opamp output stage bias current. Check post #11 and point out where are those odd harmonics. ;)

It would simplify the PCB and save some components, as well as helping the layout be simpler and less prone to noise and rubbish floating around.
This PCB design is specifically not prone to EMI as it uses outer copper layers as static shields.
Check post #4. Amp output noise peaks are below 1 uV each. Check build by manniraj at post #208. No hum at all with 100 dB/1W loudspeakers.
 
  • Like
Reactions: 6 users
Member
Joined 2019
Paid Member
As far as "how it sounds" ? At 20W , you would fail to hear the difference between this design or any similar diamond buffer (fast EF3-AB discrete) design.
The op-amp or EF3/lateral FET differences/choices might be marginally audible .... but only to "golden ears".
I’m not in the “golden ears club” and I leave them to be (self proclaimed or real). For the rest of us, there is one reliable method to tell how good is the sound and that is an overall impression after long listening session, details in sound reproduction and perceived sound stage. There we can find differences between amplifiers, beyond distortion numbers.

They do sound quite good , you can hear fast....
Yes, I would describe sound as fast or very transient and detailed, but it is important that it doesn’t produce any listening fatigue. Probably, due low intermodulation distortion and high slew rate.
 
  • Like
Reactions: 1 user
Member
Joined 2019
Paid Member
i had a hearing session with a BA3- Toshiba and M2OPS with famous Toshiba mosfet.
Source Auralic Vega G2, compare amp was AVM 6.3, LS KEF 11.
That are really outstanding components available to you for listening sessions. With extremely low distortion amplifiers as is this one, I would use direct preamplifier output from Auralic Vega.

KEF R11 can provide wide and deep sound stage, full of details and completely detached from the loudspeakers, and aside from power requirements are a perfect match. If you manage to build this amplifier, I would be very interested in your findings.
 
  • Like
Reactions: 1 users
There we can find differences between amplifiers, beyond distortion numbers.
Like Valve amps ? Or on SS amps , how the VAS reacts at clip ? Again , like Valve amps... You can definitely hear a saturated BJT VAS.
A Hawksford VAS sounds different than a CCS one near the rails (and definitely at them).
CFA = I suspect the interaction of the FB loop with the EMP/Xover of the drivers is at play.
OS
 
Member
Joined 2019
Paid Member
Like Valve amps ? Or on SS amps , how the VAS reacts at clip ?
I don’t mean differences between very different designs (valve, SS, class A, AB or D) or differences at near clipping levels. My listening habits provide that clipping is an unknown experience.

For me, there is a difference in overall reproduction impression, as I enjoy music as a whole reproduction performed on the stage far behind loudspeakers, and there unmasked details in reproduction, sound balance and sound stage details play dominant role. There are differences, especially when some sounds disappear behind your back (OK that is only with artificially created sound stage as is with “Too much rope” by Roger Waters). It is not the same with different amplifiers.

Your amplifier designs are very accomplished and probably all sound extremely good, but differ at tiny sound details or near max. power.
I’m just returning to this hobby after a very long pause.
 
  • Like
Reactions: 2 users