One JFET poweramp looking for THE FET SOUND

[lineup]

JFET has a different sort of distortion than BJT.
Most poweramps have BJT input + Error correction.
And even those poweramps having JFET inputs
has got so much gain and negative feedback,
that they should not sound at all or sound just like BJT.


This idea of mine uses only 1 JFET to amplifiy and take care of feedback.
It will have some distortion, but of the type of dist that JFET has.
See diagram.
It is a very simplified presentation of my idea.

The signal comes across R3.
You need no input cap, but we usually attach one anyway.
2N4416 JFET was chosen because of VGS.
2SK170 is not very good, because here we need VGS 1-2 Volt for producing some milliAmperes.

On top is current mirror.
That mirror will not change or amplifiy. Just mirror.
On the bottom is corresponding CCS, current source,
to balance current coming from mirror.
This stage drives the output stage follower.

The feedback and gain is handled by R2/R4.
And feedback is taken to the JFET Source.

What is important is to add a good output stage.
One JFET (or other small signal device) will give more distortion
the heavier the load is.
So, to give suitable light load here I would say we need a
Triple Darlington or similar.
With very high input impdance.

I have a couple good working protypes in SPICE.
Where distortion is at very respectable level.
I will post one.
And, yes!
It gives what I was after: mostly 2nd order distortion.
THE FET SOUND is here ...

 

[Globulator]

JFET has a different sort of distortion than BJT.
Most poweramps have BJT input + Error correction.
And even those poweramps having JFET inputs
has got so much gain and negative feedback,
that they should not sound at all or sound just like BJT.


This idea of mine uses only 1 JFET to amplifiy and take care of feedback.
It will have some distortion, but of the type of dist that JFET has.
See diagram.
It is a very simplified presentation of my idea.

The signal comes across R3.
You need no input cap, but we usually attach one anyway.
2N4416 JFET was chosen because of VGS.
2SK170 is not very good, because here we need VGS 1-2 Volt for producing some milliAmperes.

On top is current mirror.
That mirror will not change or amplifiy. Just mirror.
On the bottom is corresponding CCS, current source,
to balance current coming from mirror.
This stage drives the output stage follower.

The feedback and gain is handled by R2/R4.
And feedback is taken to the JFET Source.

What is important is to add a good output stage.
One JFET (or other small signal device) will give more distortion
the heavier the load is.
So, to give suitable light load here I would say we need a
Triple Darlington or similar.
With very high input impdance.

I have a couple good working protypes in SPICE.
Where distortion is at very respectable level.
I will post one.
And, yes!
It gives what I was after: mostly 2nd order distortion.
THE FET SOUND is here ...

Link p12 has FETs on it.

Distortion depends a bit on the accuracy of the SPICE model which may not match reality for that degree of analysis - but I do like FETs.

If you really just want 2nd and very little else, you need a triode front end like a nice 6N2P or 12AX7

 

[wakibaki]

What was the point of quoting the entire post? It added nothing and wastes bandwidth.

w

 

[AndrewT]

not only bandwidth but had he included the sketch he would have broken Forum rules, which prohibit quoting entire posts.

 

[lineup]

I am sorry I forgot this topic.
I will post a working circuit

 

[lineup]

JFET with triple darlington

Triple darlington output works the best.
- Class AB
- Max 25 Watt into 8 OHM
- Less than 0.020 % THD at 25 Watt
- It is fast, 2 MHz -3dB at gain >20

Second Image shows 25 Watt into 8 ohm

 

[MiiB]

would it be possible to adapt it for +- operation and fit a pair of lateral fet's for output devices...the triple seems to complex for my liking...

M

 

[Bonsai]

Lineup, nice simple design - but I agree a lateral FET output is the way to go. Otherwise, what about a nice Locanthi 'T'?

 

[Melon Head]

My vote also goes with lateral mos output

 

[MiiB]

A little thinking...on your concept...Sure it'll sound good.. distortion is app -85 dB second order...then dropping to app -100 dB third and even lover 4.th...

Dual rail lateral FET output..here with 600 mA..app 30 W dissipation each...a little hot maybe..


DC component:0.00494159

Harmonic Frequency Fourier Normalized Phase Normalized
Number [Hz] Component Component [degree] Phase [deg]
1 1.000e+03 7.672e+00 1.000e+00 -0.18° 0.00°
2 2.000e+03 4.651e-04 6.062e-05 85.67° 85.85°
3 3.000e+03 4.593e-05 5.987e-06 49.85° 50.03°
4 4.000e+03 1.180e-06 1.537e-07 -117.73° -117.54°
5 5.000e+03 3.060e-07 3.988e-08 -80.47° -80.28°
6 6.000e+03 4.605e-08 6.002e-09 51.90° 52.09°
7 7.000e+03 5.001e-08 6.519e-09 -2.38° -2.19°
8 8.000e+03 3.390e-08 4.418e-09 -46.01° -45.83°
9 9.000e+03 7.551e-08 9.842e-09 167.27° 167.46°
Total Harmonic Distortion: 0.006091%

 

[Bonsai]

How does this amp perform if you overdrive it?

 

[MiiB]

if you talk about my take on the nice Lineup concept..

Then I think it will perform quite well.. when driven harder distortion increases but maintains the same kind of distribution...with predominantly 2. order..
The lateral fets slide gently into class B operation..and the feedback currents are smooth with no trace of switching spikes...

I believe that it will be a good music performer..

Harmonic Frequency Fourier Normalized Phase Normalized
Number [Hz] Component Component [degree] Phase [deg]
1 1.000e+03 2.301e+01 1.000e+00 -0.19° 0.00°
2 2.000e+03 4.672e-03 2.030e-04 83.92° 84.12°
3 3.000e+03 6.458e-04 2.806e-05 -26.08° -25.88°
4 4.000e+03 1.800e-04 7.823e-06 179.05° 179.24°
5 5.000e+03 3.477e-04 1.511e-05 80.23° 80.43°
6 6.000e+03 1.901e-04 8.262e-06 171.35° 171.54°
7 7.000e+03 3.969e-04 1.725e-05 81.88° 82.07°
8 8.000e+03 7.447e-05 3.236e-06 178.11° 178.30°
9 9.000e+03 1.487e-04 6.460e-06 82.14° 82.34°
Total Harmonic Distortion: 0.020665%

 

[CopperTop]

Is it not possible to synthesise this sort of 'musicality' digitally, even if you have the misfortune to own a more-or-less distortionless amp?

 

[MiiB]

CopperTop.....What is the point..???

This is not by a long shot the most well specified amplifier...but the pure simplicity of the circuit and the nature of distortion does hold some promise.... Think lineup has found a very good way to use the Jfet..to keep it simple and at the same time keep it scalable..and flexible.. gain can be adjusted to fit the needs...here as a Power amp...but by inserting a different buffer also as a nice preamp..with really good specs for such a simple circuit.. could also pose as a nice I-V converter for DAC's.
To me a good circuit building block with plenty of applications...and variations

 

[lineup]

THD 1 Watt 8 Ohm = 0.005%

As shown for 8 Ohm is 25W at THD 0.020%
Max out into 4 Ohm is 49W at THD 0.019%
see image.

Bias Class AB with 50mA idle in output in MJL3281A/MJL1302A
VGain is x25

 

[MiiB]

app same distortion as the lateral fet version....think the feedback injection below the Jfet is the decisive factor...

I have decreased feed back a little and increased gain to 30 dB....bandwith is app 300 Khz at -3 dB

I have worked a little more on the concept and fitted a non switching output stage, separate supplies with +-60 V for the front end and +- 50V for the output stage....I plan to build it...thanks

 

[lineup]

Thanks for tips, MiiB
I may try your concept, sort of.
Interesting

 

[lineup]

How does this amp perform if you overdrive it?

It will distort when too high voltage input.Clip.
For 4 ohm it looks good.

 

[Bonsai]

Well, thanks for that Lineup!

What I meant was is there any rail sticking or parasitic oscillation?

 

[cbdb]

What about stability? (gain/phase margins).