a wonderfull sziklai

[catalin]

Dadod ,as I said the cob of the power bjt+drivers are important for thd and sound also .If you don't have models for 2sc5200,SA1943 and 2SC5171,1930 I will share later .

And an important aspect is that human ear is making difference to the low thd at low power 1-3 watts .For 100 Watts the ear or brain is making the difference impossible to hear.
Also for listening at home 1-3 watts is often used .

[dadod]

Don't think if I ask some questions that I don't like your amp, I think it is very good and I like your "twisted" ideas about NFB.
I used some of it here bootstrapsCCS+T-TMC and later I designed this amp Little GEM.
dado

[catalin]

Originally Posted by dadod

Don't think if I ask some questions that I don't like your amp, I think it is very good and I like your "twisted" ideas about NFB.
I used some of it here bootstrapsCCS+T-TMC and later I designed this amp Little GEM.
dado

No Dadod ,I didn't say this .I just want to help to get the same results .Excuse if I am a little "hard" speaking .
I am happy that you have better results with my tweaks .I call this cascode miller compensation .
Personal I am a fan of the thd at 20khz ..
The results are more important in stability terms .
On some layout's this ideea is help a lot to get rid of some instability .Also the phase reserve is better .
And another advantage by using this kind of compensation is that the DC potentials on the cap are almost equal to 0 .May be you know that every cap has a internal noise directly dependent with the dc voltage applied .

[Jay]

Originally Posted by catalin

Jay ,I don't remember to have a schematic with cascode ...
Always pay attention to the simulation ,idle currents and models .

Ups, only current mirror. I mixed up with Wahab's amp for Gaborbela.

[catalin]

Here is my bjt models+transient ltspice sim .
Dadod ,my model of temperature compensation bjt was corupted .Is fixed now .

Jay ,I will post a version of no current mirror for you .I am also adept to simplier things .By removing the current mirror and the ef before vas we can put a miller cap of a half from what is now ,6pf .
What I don't like in that version is the dc offsets with the temperature though the olg at 20khz is almost the same with one with current mirror so we can achieve almost the same thd at 20khz .

[catalin]

For Jay :Joy_res :the no current mirror in attach .And I think is one of the vfb amp's with the lowest phase in OLG and also miller cap .
For both I am interested to design a load for the vas to be current non-dependent with temperature ,can be a resistor without bootsrap or a current mirror .

What encyclopedia do you have on your pc !

Also I think that both versions can work with +/-10mV at output without the 22uF input cap

[Jay]

Hi Catalin, thanks.

I would say that the FFT of your first circuit is equal with P3A, except that your circuit running much higher bias current, and one special characteristics of the P3A is the beautiful distortion profile.

I haven't check the AC analysis of your circuit, but P3A is wonderful there.

The second circuit is interestingly good, especially very good for low SPL listening. The THD for high SPL is very bad tho.

I think I will use both P3A and your circuit to benchmark my SSA-CFP.

[catalin]

Originally Posted by Jay

Hi Catalin, thanks.

The second circuit is interestingly good, especially very good for low SPL listening. The THD for high SPL is very bad tho.

I think I will use both P3A and your circuit to benchmark my SSA-CFP.

I don't remeber my first circuit .Can you direct to the schematic ?
Why do you say that the thd is very bad for high spl ?

[Jay]

Originally Posted by catalin

I don't remeber my first circuit .Can you direct to the schematic ?

60-80W Power Amplifier

ADD: Or may be you're asking the joy_sin.asc, the one with current mirror.

Originally Posted by catalin

Why do you say that the thd is very bad for high spl ?

Sorry, I used to see amp behavior from the very first seconds the signal is applied. It is probably not valid to say so, but whatever, I like to see a homogenous FFT for input signal from 0.1 to 0.7 at short stop time:

[catalin]

Yes Jay ,Thank you .The resistive load version is modded from the current mirror for your interest .I will continue for the moment with the current mirror version but if you are interested we can improve also the resistive one decreasing the emittor ltp if you are a "searcher for best thd" .But doing this the miller cap must be increased which I don't prefer .
As I said my interest is for the first 1-10watts ,enough in a 20-30 mp room .
More than this can't be detected related to thd . Also I prefer exactly the sound of the one with current mirror and small current feedback .

I see that the R21 resistor which is setting the current of the ltp is 560 ohms by my hurry .It should be 220ohms to have almost 3 mA in Ltp