Will this work as audio amplifier?

[MikeB]

Originally posted by thanh

OOOH! Are you sure? Which circuit did you simulated? I'm using Orcad 9.2.My sim says that single ended Jfet produces 2nd ,3rd very much and a little high order harmonics .
I think Jfet with no source resistor only produces 2nd but in the real world Jfet alway is wired with a resistor
http://www.its.caltech.edu/~musiclab...per-crobat.pdf

???
I think i use some 9.1 ...
Which jFet did you use ? I mainly use the sk170.
I use the jFets without sourceresistor, works fine. (In real world)
With singleended, i had nearly no 3rd harmonic, about <0.5% of
2nd, and for high order i have to zoom far in.
A diffamp with jFets showed 3rd Harmonic, 2nd completely canceled
out, and other barely existent.
And all harmonics were way below a BJT, really extreme with
large voltage/current-swings.
When using a ccs, the BJTs seems better, but this does not work
for voltageamplification. (But perfect for buffer)

If only jFets would be cheaper, matching these gets really expensive !
(i paid 49cent per sk170, only 4cent per bc546b/556b)

Mike

[lumanauw]

Hi, Dad,
I think you forgot to put feedback resistor to base of Q1.
Please write the SIMS result that you find. Especially the harmonics figures, and if there is anything strange with this cct.

[Justcallmedad]

I will send you tomorrow simulation results, but it's mean nothing for sound quality.

Try it ! In real world !

I think that you should better keep the R8 resistor that sets the gain, and avoid feedback, but it's something you can test (listen).

Did you bilt it ?

[Justcallmedad]

FFT plot for 1v AC input, 20 V peak to peak Output.
With no feedback, and R8=1k, if you need more gain increase R8.

Build it, and keep me inform about listening tests.

[lumanauw]

Hi, Dad,

Try it ! In real world !

Yes, I've built it look at post#22. In my prototype, the servo type doesnt work. The offset is about 7V. But when I replace the servo with manual adjustment VR, it works.

Build it, and keep me inform about listening tests

Look at post #28

With no feedback, and R8=1k, if you need more gain increase R8.

If we make it without R8 (you said its for sim only), and put output stage instead, it will need feedback resistor, wont it?

You just indicating one interesting point. Is this design can be made non-feedback? I think the gain is low enough for non-feedback, and needs only other R adjustment. Loading it with R8 maybe a good candidate for non-feedback type.

In your drawings, what is the purpose of R3,R4,R5?

Thank you for simulating it and generate FFT. But what is the interpretation of lumanov.cir? Is it like ordinary power amp, or differs?

Also, do you find problem with stability in SIMS?

[lumanauw]

This is my prototype. Q1+Q3+25K is replaced by only Q1, so VAS is formed folded cascode.

[thanh]

Mike! I use 2SJ109
You are using orcad 9.1,aren't you? Can you post your circuit?
Indeed,I'm not good at designing FET.
Look at Dad's simulation, 3rd harmonic is quite high. Ok! high order harmonics are still alive
A ideal Jfet has no juntion resistors .But in the real world and in simulating sofware Jfet has a juntion resistor

[thanh]

But this means that
simple THD-measurements are useless to describe the soundquality
of an amp. (Not a new discovery at all...)

And IMD ?
I like zero distortion.I think HIFI is high fidelity .This means you can't regconize the difference of sound of amp and singing

[MikeB]

Hi thanh !

Originally posted by thanh
Mike! I use 2SJ109
You are using orcad 9.1,aren't you? Can you post your circuit?
Indeed,I'm not good at designing FET.
Look at Dad's simulation, 3rd harmonic is quite high. Ok! high order harmonics are still alive
A ideal Jfet has no juntion resistors .But in the real world and in simulating sofware Jfet has a juntion resistor

At moment i'm at work, but as soon as possible i can post some
schematic. But maybe we should use a separate thread ?
With junctionresistor you mean a virtual resistance inside the
device ? I think in Dad's simulation the harmonics are typical for
an asymetrical bjt-design with moderate feedback. But still 3rd
harmonic is only ~10% of 2nd. Not bad, i think !

Originally posted by thanh

And IMD ?
I like zero distortion.I think HIFI is high fidelity .This means you can't regconize the difference of sound of amp and singing

It seems, that low TIM is important, but as representive as THD.
Of course, an Amp with no THD/TIM would be ideal, but unrealistic.
So, we need compromises...
Take a look at http://www.gedlee.com/distortion_perception.htm
If i understand "masking" correct (damned psychoacoustic), it
means that if you have enough 2nd harmonic, the 3rd gets
simply inaudible ! lumanauw should really try out his circuit.
This would mean, a good sounding amp might be a circuit that
simply "masks" 3rd harmonic by generating enough 2nd. But sadly
2nd harmonic is to easy cancelled out. (With nearly any symetrical
circuit, like pushpull)
Do you have a spicemodel of tubes ? I heard that tubes don't
generate 3rd harmonic like transistors do.
Something interesting: in my latest amp (asymetrical, but pushpull)
i simulated fuses by adding 0.05ohm to powersupply. And suddenly
the amp produced nearly even harmonics only...
Symetrical design does not react on these resistors in this way.
Okay, i will continue my research on transfercurves.

Mike

[thanh]

But maybe we should use a separate thread ?

Let you do as you can! Mike!

With junctionresistor you mean a virtual resistance inside the device ?

Yes
Have you read a document at http://www.its.caltech.edu/~musicla...aper-crobat.pdf ?