ExtremA, class-A strikes back?

PMA · 2014-11-13 8:58 am

Federmann said:
For good sound sufficient SR 1,6V/µs.

High audio frequency never have large amplitudes.

Plain nonsense.

SSassen · 2014-11-13 9:52 am

I couldn't resist whipping up Pavel's idea in the simulator, made it into a full bridge for better efficiency and set the output for 80W/8R. The numbers I got aren't in the least bit disappointing, here's the THD20K at 80W/8R:

Fourier components of I(r38)
DC component:1.88071e-006

Harmonic Frequency Fourier Normalized Phase Normalized
Number [Hz] Component Component [degree] Phase [deg]
1 2.000e+04 3.171e+00 1.000e+00 -0.28° 0.00°
2 4.000e+04 3.375e-06 1.064e-06 -55.67° -55.39°
3 6.000e+04 5.025e-06 1.585e-06 -118.35° -118.07°
4 8.000e+04 2.325e-06 7.333e-07 157.09° 157.37°
5 1.000e+05 1.683e-06 5.308e-07 77.54° 77.82°
6 1.200e+05 9.958e-07 3.141e-07 0.57° 0.85°
7 1.400e+05 4.393e-07 1.386e-07 -107.32° -107.04°
8 1.600e+05 4.818e-07 1.520e-07 101.26° 101.54°
9 1.800e+05 8.989e-07 2.835e-07 7.13° 7.41°
Total Harmonic Distortion: 0.000216%(0.008551%)

SSassen · 2014-11-13 9:59 am

Alright, simulated efficiency is 186W in, 80W out, that's 80/186 > 43%, not bad considering this is class-A.

BV · 2014-11-13 10:15 am

Try to run asc file
http://www.diyaudio.com/forums/solid-state/96853-extrema-class-strikes-back-30.html#post4120514
Fourier components of I(r40)
DC component:0.000407211

Harmonic Frequency Fourier Normalized Phase Normalized
Number [Hz] Component Component [degree] Phase [deg]
1 1.000e+04 2.669e+00 1.000e+00 178.12° 0.00°
2 2.000e+04 5.812e-08 2.178e-08 3.00° -175.12°
3 3.000e+04 5.652e-06 2.118e-06 -124.58° -302.70°
4 4.000e+04 3.135e-08 1.174e-08 -0.20° -178.32°
5 5.000e+04 3.423e-07 1.282e-07 67.10° -111.03°
6 6.000e+04 2.038e-08 7.636e-09 0.03° -178.09°
7 7.000e+04 2.216e-08 8.303e-09 -60.47° -238.59°
8 8.000e+04 1.510e-08 5.659e-09 -0.11° -178.23°
9 9.000e+04 1.385e-08 5.188e-09 5.15° -172.97°
Total Harmonic Distortion: 0.000212%

jmd01 · 2014-11-13 10:19 am

SSassen,
Power supply can be lower (V2 and V3 = 18-20V).

SSassen · 2014-11-13 10:33 am

BV said:
Try to run asc file
http://www.diyaudio.com/forums/solid-state/96853-extrema-class-strikes-back-30.html#post4120514
Fourier components of I(r40)
DC component:0.000407211

Harmonic Frequency Fourier Normalized Phase Normalized
Number [Hz] Component Component [degree] Phase [deg]
1 1.000e+04 2.669e+00 1.000e+00 178.12° 0.00°
2 2.000e+04 5.812e-08 2.178e-08 3.00° -175.12°
3 3.000e+04 5.652e-06 2.118e-06 -124.58° -302.70°
4 4.000e+04 3.135e-08 1.174e-08 -0.20° -178.32°
5 5.000e+04 3.423e-07 1.282e-07 67.10° -111.03°
6 6.000e+04 2.038e-08 7.636e-09 0.03° -178.09°
7 7.000e+04 2.216e-08 8.303e-09 -60.47° -238.59°
8 8.000e+04 1.510e-08 5.659e-09 -0.11° -178.23°
9 9.000e+04 1.385e-08 5.188e-09 5.15° -172.97°
Total Harmonic Distortion: 0.000212%

Are you referring to the discrete design, or the opamp design? I'm assuming the former? In that case, .four 20k I(R40) is what you'd like to use.

SSassen · 2014-11-13 10:39 am

jmd01 said:
SSassen,
Power supply can be lower (V2 and V3 = 18-20V).

Correct, with 20V per rail for the output stage there's a bit of headroom left still and efficiency goes up. Simulated efficiency is 152W in 80W out, which is 80/152 > 53%.

BV · 2014-11-13 10:40 am

.four 20k I(R40) is what you'd like to use.

So try it, simply edit 10kHz to 20kHz..
Fourier components of I(r40)
DC component:0.000265904

Harmonic Frequency Fourier Normalized Phase Normalized
Number [Hz] Component Component [degree] Phase [deg]
1 2.000e+04 3.465e+00 1.000e+00 176.23° 0.00°
2 4.000e+04 8.835e-09 2.550e-09 28.75° -147.48°
3 6.000e+04 4.338e-06 1.252e-06 -138.55° -314.78°
4 8.000e+04 1.114e-08 3.216e-09 -8.34° -184.57°
5 1.000e+05 3.114e-06 8.989e-07 63.60° -112.63°
6 1.200e+05 5.183e-09 1.496e-09 -7.20° -183.43°
7 1.400e+05 3.595e-07 1.037e-07 -119.38° -295.61°
8 1.600e+05 3.816e-09 1.101e-09 -12.84° -189.07°
9 1.800e+05 4.946e-08 1.427e-08 61.28° -114.96°
Total Harmonic Distortion: 0.000154%

SSassen · 2014-11-13 10:45 am

BV said:
So try it, simply edit 10kHz to 20kHz..

I'm sorry, I'm not entirely sure what you'd like me to do? You also need to change the input to SINE(0 1.4 20k) if you want to look at THD@20K.

SSassen · 2014-11-13 10:57 am

@BV,

You're referring to your post #298 correct? If you can forward me the .asc I can have a closer look, I can be reached at ssassen(at)hardwareanalysis(dot)com, feel free to email the .asc to that address.

BV · 2014-11-13 10:58 am

Sine source is defined this way, "SINE(0 1.3 {freq} 0 0 0 10)", and parameter {freq} is defined in schematic as spice directive line ".param Freq=20k". So it is enough to edit 10kHz to 20kHz in this line...
And .asc file is in post 298 .

SSassen · 2014-11-13 11:03 am

BV said:
Sine source is defined this way, "SINE(0 1.3 {freq} 0 0 0 10)", and parameter {freq} is defined in schematic as spice directive line ".param Freq=20k". So it is enough to edit 10kHz to 20kHz in this line...
And .asc file is in post 298 .

I'm sorry, but I don't need tutoring on how to use LTspice

PMA · 2014-11-13 11:10 am

jmd01 said:
SSassen,
Power supply can be lower (V2 and V3 = 18-20V).

With 2 x 18V for the opamp, you need at least 2 x 22V for the output stage (ccs and triple junction margin), otherwise distortion gets higher).

31697B · 2014-11-13 11:16 am

Nice idea.
Kinda reminds me of driving a OPS bias pair with ref to the output. Conceptually, wasnt that done in a former life with a diode-resistor from output to bias (spreader)? (NPass). But this is a nice balanced circuit refinement.... like a comparitor, too.
Opamp input now then one of the many good discrete front end, later?

THx-RNMarsh

PMA · 2014-11-13 11:30 am

This OPS input current is, hopefully, only in microamperes, as long as the output stage remains in class A. That's all the magic - it almost does not load opamp output.

31697B · 2014-11-13 11:37 am

No load on the opamp is great as that was the last area of concern (OPS loading) for using opamps. How much current does the sim say the opamp must supply (max) when the amp goes out of class A?

THx-RNMarsh

PMA · 2014-11-13 11:52 am

RNMarsh said:
No load on the opamp is great as that was the last area of concern (OPS loading) for using opamps. How much current does the sim say the opamp must supply (max) when the amp goes out of class A?

THx-RNMarsh

Not nice, then. Starts quite mild, as attached, but continues to milliamps and tens of milliamps.

The image is for stepped load 8, 6, 4 ohms. You can see input current and ouput current through one output device.

added an image for 2 ohm load

BV · 2014-11-13 12:07 pm

2 SSassen

I'm sorry, but I don't need tutoring on how to use LTspice

So why You ask this?

I'm sorry, I'm not entirely sure what you'd like me to do? You also need to change the input to SINE(0 1.4 20k) if you want to look at THD@20K

franzm · 2014-11-13 12:36 pm

SSassen said:
...Simulated efficiency is 152W in 80W out, which is 80/152 > 53%.

How much is the (theoretical) maximal efficiency of this OPS?

PMA · 2014-11-13 12:42 pm

You have to specify idle current first. Otherwise there is no good answer.

Class A for 8 ohms only, or for 4 ohms as well, or for 2 ohms as well? Idle current is half peak current to stay in class A.

ExtremA, class-A strikes back?

Member

Member

Member

Member

Member

Member

Member

Member

Member

Member

Member

Member

Member

Disabled Account

Member

Attachments

Disabled Account

Member

Attachments

Member

Member

Member