H2

I am not Nelson, but I would say - certainly this can be done. For instance take a class A-amplifier and make it asymmetric....

Hi, voltwide.

Thanks, for taking the time to respond. Yes, generically speaking, waveform asymmetry is what produces 2nd order distortion. However, I'm asking Nelson something different from that.

As Nelson addresses in his H2 paper, the residual distortion waveform of the H2 product can be either positvely or negatively phased, relative to the fundamental waveform. I'm wondering how to predict such relative phase between the fundamental and H2 residual from only the amplifier's schematic. In other words, before building and measuring the working circuit.
 
Member
Joined 2011
Paid Member
Ken, if you remember how to calculate Fourier series from school, it's a pretty straightforward exercise.

Just define "waveform 1" as (sinewave of amplitude 10 on the positive half cycle) plus (sinewave of amplitude 5 on the negative half cycle). Calculate its Fourier series.

Then define "waveform 2" as the inverse: (sinewave of amplitude 5 on the positive half cycle) plus (sinewave of amplitude 10 on the negative half cycle). Calculate its Fourier series.

Waveform1 is the output of a circuit whose puller upper gain is 2X greater than its puller downer gain.

Waveform2 is the output of a circuit whose puller downer gain is 2X greater than its puller upper gain.

Because they are (wildly!) asymmetric, waveforms 1 and 2 will have a lot of 2nd harmonic distortion "H2". One of them will be positive phase H2, the other will be negative phase H2. Fourier series will tell you which is which.

Then you can open up to a brand new page in your engineering notebook and write:

HOW TO GENERATE NEGATIVE PHASE H2:

Design a circuit whose puller upper gain is (GREATER, SMALLER) than its puller downer gain.

{Circle either the word GREATER or the word SMALLER according to your Fourier results}.

###########################################

Here's an illustrative example. I created two waveforms in the LTSPICE circuit simulator. One of them has 2X greater amplitude on the positive half cycle, the other has 2X greater amplitude on the negative half cycle. Then I told LTSPICE to calculate the Fourier decomposition of those two waveforms. The results are clear: one of them ("FOXTROT") has positive phase H2, see blue box. The other of them ("GOLF") has negative phase H2, see green box. Which is which? I won't say. I don't want to rob you of the joy of discovery. Do the calculations, have some fun, and reap the rewards of diligent effort!

_
 

Attachments

  • dist_X.png
    dist_X.png
    14.5 KB · Views: 660
  • dist_Y.png
    dist_Y.png
    14.3 KB · Views: 651
  • HD2.png
    HD2.png
    48.6 KB · Views: 648
Last edited:
Ken, if you remember how to calculate Fourier series from school, it's a pretty straightforward exercise.

_

Mark, thanks for producing such an extensive response! You saved Nelson the trouble. :D

The key design parameter: "Design a circuit whose puller upper gain is (GREATER, SMALLER) than its puller downer gain." Intuitively, I would think that whichever half of the fundamental waveform showed the greater asymmetry would alone dictate whether the H2 is positive or negative, without resort to Fourier analysis.
 
Last edited:
Ken, if you remember how to calculate Fourier series from school, it's a pretty straightforward exercise.

Just define "waveform 1" as (sinewave of amplitude 10 on the positive half cycle) plus (sinewave of amplitude 5 on the negative half cycle). Calculate its Fourier series.

Then define "waveform 2" as the inverse: (sinewave of amplitude 5 on the positive half cycle) plus (sinewave of amplitude 10 on the negative half cycle). Calculate its Fourier series.

Waveform1 is the output of a circuit whose puller upper gain is 2X greater than its puller downer gain.

Waveform2 is the output of a circuit whose puller downer gain is 2X greater than its puller upper gain.

Because they are (wildly!) asymmetric, waveforms 1 and 2 will have a lot of 2nd harmonic distortion "H2". One of them will be positive phase H2, the other will be negative phase H2. Fourier series will tell you which is which.

Then you can open up to a brand new page in your engineering notebook and write:

HOW TO GENERATE NEGATIVE PHASE H2:

Design a circuit whose puller upper gain is (GREATER, SMALLER) than its puller downer gain.

{Circle either the word GREATER or the word SMALLER according to your Fourier results}.

###########################################

Here's an illustrative example. I created two waveforms in the LTSPICE circuit simulator. One of them has 2X greater amplitude on the positive half cycle, the other has 2X greater amplitude on the negative half cycle. Then I told LTSPICE to calculate the Fourier decomposition of those two waveforms. The results are clear: one of them ("FOXTROT") has positive phase H2, see blue box. The other of them ("GOLF") has negative phase H2, see green box. Which is which? I won't say. I don't want to rob you of the joy of discovery. Do the calculations, have some fun, and reap the rewards of diligent effort!

_
I have a question.In the LTSpice distortion readings is the phase of the second harmonic referred to the first harmonic? ie: Does 90 degrees 2nd harmonic phase mean that it starts at the 90 degree point of the 1st harmonic?
 
Member
Joined 2011
Paid Member
The quickest way to find out is to make a little simulation circuit which sums a 1kHz ideal sinewave voltage source with a 2kHz ideal sinewave voltage source. Then run LTSPICE .FOURIER analysis on the summation. When the two ideal sources are displaced by +60 degrees (according to your definition), what does LTSPICE call that? +60 degrees? -60 degrees? Or some other number?
 
The quickest way to find out is to make a little simulation circuit which sums a 1kHz ideal sinewave voltage source with a 2kHz ideal sinewave voltage source. Then run LTSPICE .FOURIER analysis on the summation. When the two ideal sources are displaced by +60 degrees (according to your definition), what does LTSPICE call that? +60 degrees? -60 degrees? Or some other number?

Thanks for the tip. +90 degrees 2nd harmonic is perfect -ve phase second harmonic according to Spice. The degrees are referenced to the harmonic ie: 2nd at +90 degrees=1st harmonic at +45 degrees.
 
Thanks for the tip. +90 degrees 2nd harmonic is perfect -ve phase second harmonic according to Spice. The degrees are referenced to the harmonic ie: 2nd at +90 degrees=1st harmonic at +45 degrees.

I should clarify one thing.My reference to the 1st harmonic at +45 degrees does not mean is was at +45 degrees while the 2nd was at +90 degrees in the spice log.The 1st was at 0 degrees.I just meant that 45 degrees 1st equals 90 degrees second time wise.
 
Ken, if you remember how to calculate Fourier series from school, it's a pretty straightforward exercise.

Just define "waveform 1" as (sinewave of amplitude 10 on the positive half cycle) plus (sinewave of amplitude 5 on the negative half cycle). Calculate its Fourier series.

Then define "waveform 2" as the inverse: (sinewave of amplitude 5 on the positive half cycle) plus (sinewave of amplitude 10 on the negative half cycle). Calculate its Fourier series.

Waveform1 is the output of a circuit whose puller upper gain is 2X greater than its puller downer gain.

Waveform2 is the output of a circuit whose puller downer gain is 2X greater than its puller upper gain.

Because they are (wildly!) asymmetric, waveforms 1 and 2 will have a lot of 2nd harmonic distortion "H2". One of them will be positive phase H2, the other will be negative phase H2. Fourier series will tell you which is which.

Then you can open up to a brand new page in your engineering notebook and write:

HOW TO GENERATE NEGATIVE PHASE H2:

Design a circuit whose puller upper gain is (GREATER, SMALLER) than its puller downer gain.

{Circle either the word GREATER or the word SMALLER according to your Fourier results}.

###########################################

Here's an illustrative example. I created two waveforms in the LTSPICE circuit simulator. One of them has 2X greater amplitude on the positive half cycle, the other has 2X greater amplitude on the negative half cycle. Then I told LTSPICE to calculate the Fourier decomposition of those two waveforms. The results are clear: one of them ("FOXTROT") has positive phase H2, see blue box. The other of them ("GOLF") has negative phase H2, see green box. Which is which? I won't say. I don't want to rob you of the joy of discovery. Do the calculations, have some fun, and reap the rewards of diligent effort!

_

Hi Mark Johnson. Your above suggestions can be readily implemented by using Class aP amplification as follows:

1. Precision rectify the AC signal to get separate positive and negative pulses.

2. Pulses of either polarity can now be processed independently by DC amplifiers.

3. For example; the positive going pulse is amplified by a gain factor of two and the resultant summed with the intact negative pulse.

In your earlier post# ~190, you also discussed the subject regarding push-pull amplification. Class aP amplification is actually Class B;, but with a broader flexibility [and needed complexity] to fuss at will with distortion as you taught.

I will dust-off my one-of-kind stereo "precision aP generator" and explore experiments. A picture of it and its schematic are therein the thread.

best
Anton
 
Listening impressions?

From the picture of the table at Burning Amp, a large number of attendees received the H2 Generator board supplied gratis by Nelson.


That being said, I can't say that I have seen many posts with regard to people's experiences with the H2 in their systems.


FWIW, putting the H2 into my system comprised of XA60.8 amps and Martin Logan CLC Art speakers fed by a PC-based music server and Mytek Manhattan II dac, I was surprised by the affect even tho' Nelson had indicated that the XA60.8s had a fair amount of 2nd harmonic designed in.



The system sans H2 is detailed but not brittle or hard with a deep and wide soundstage which is quite immersive. With the H2, the trebles have more rounded edges without loss of detail. MIdrange voices are more natural and lifelike. Some soundstage width and depth is lost but not enough to be an issue.


Now that the pcb is in a build-box with external meter and pot for adjusting voltage, I am looking forward to experimenting with voltage +/- 0.3 to see if either are readily discernible and where my personal preference lies.


Thanks to Nelson for making the H2 Generator available!
 
From the picture of the table at Burning Amp, a large number of attendees received the H2 Generator board supplied gratis by Nelson.


That being said, I can't say that I have seen many posts with regard to people's experiences with the H2 in their systems.


FWIW, putting the H2 into my system comprised of XA60.8 amps and Martin Logan CLC Art speakers fed by a PC-based music server and Mytek Manhattan II dac, I was surprised by the affect even tho' Nelson had indicated that the XA60.8s had a fair amount of 2nd harmonic designed in.



The system sans H2 is detailed but not brittle or hard with a deep and wide soundstage which is quite immersive. With the H2, the trebles have more rounded edges without loss of detail. MIdrange voices are more natural and lifelike. Some soundstage width and depth is lost but not enough to be an issue.


Now that the pcb is in a build-box with external meter and pot for adjusting voltage, I am looking forward to experimenting with voltage +/- 0.3 to see if either are readily discernible and where my personal preference lies.


Thanks to Nelson for making the H2 Generator available!

Hello Big Guy. Your audio gear is impressive and your brain's ability to resolve the above subjective differences bears on amazing. I'd love to see a picture of your H2 build.

The H2 Harmonic Generator or its PCB by Mr. Pass are not available at the diyAudio store. I'd love to get my hands on one.

As built, it has a great potential to be the simplest headphones amp. Thus it will allow an up close assessment of H2 performance at this earliest/first stage of signal processing.

The ~400 Ohm output impedance of H2 suggests it maybe used as a current source amp. If needed, a buffer at its output will convert it to a voltage source amp to suit ones hearing/headphones.

Best
Anton
 
The one and only
Joined 2001
Paid Member
FWIW, putting the H2 into my system comprised of XA60.8 amps and Martin Logan CLC Art speakers fed by a PC-based music server and Mytek Manhattan II dac, I was surprised by the affect even tho' Nelson had indicated that the XA60.8s had a fair amount of 2nd harmonic designed in.

The amount of 2nd in the XA60.8's is miniscule compared to the amount
coming out of the Korg or H2.

BTW I have 200 of these little boards, and I am trying to find the time
to stuff some up for BAF.

:cheers:
 
Hello Big Guy. Your audio gear is impressive and your brain's ability to resolve the above subjective differences bears on amazing. I'd love to see a picture of your H2 build.

The H2 Harmonic Generator or its PCB by Mr. Pass are not available at the diyAudio store. I'd love to get my hands on one.

As built, it has a great potential to be the simplest headphones amp. Thus it will allow an up close assessment of H2 performance at this earliest/first stage of signal processing.

The ~400 Ohm output impedance of H2 suggests it maybe used as a current source amp. If needed, a buffer at its output will convert it to a voltage source amp to suit ones hearing/headphones.

Best
Anton


Thanks, Anton.


As far as getting an H2 is concerned, see Nelson's post above. I would send him a PM.


I started by replacing the board mounted RCA jacks with chassis mount jacks. I also hard wired the wall wart power supply.
My build started out pretty vanilla but as it progressed I decided to go steampunk. I added a meter powered by 9V Lithium (since powering with the internal supply resulted in noise), then I put in a switch for the meter, and finally a pot so that I can experiment with +/- voltage settings all without needing to "pop the hood". I fabricated a couple short RCA interconnects to go from my preamp to the H2.
 

Attachments

  • H2 OFF 191005.jpg
    H2 OFF 191005.jpg
    549.2 KB · Views: 409
  • H2 ON 191005.jpg
    H2 ON 191005.jpg
    486.4 KB · Views: 414
Thanks, Anton.


As far as getting an H2 is concerned, see Nelson's post above. I would send him a PM.


I started by replacing the board mounted RCA jacks with chassis mount jacks. I also hard wired the wall wart power supply.
My build started out pretty vanilla but as it progressed I decided to go steampunk. I added a meter powered by 9V Lithium (since powering with the internal supply resulted in noise), then I put in a switch for the meter, and finally a pot so that I can experiment with +/- voltage settings all without needing to "pop the hood". I fabricated a couple short RCA interconnects to go from my preamp to the H2.

Hello Big Guy.

Thank you for explaining and showing your H2 build. The external digital VM and the tuning pot are an enhancement [as you said above] for facile experimentation. I hope that your experiments will cover both negative and positive phase as the sole variables. I look forward to your upcoming impressions.

Thanks for your advice to acquire an H2. H2 is valuable. I'll PM Mr. Pass.

N.B. I do not have access to PM Mr. Pass.

Mr. Pass: I'd love to acquire an H2. Thanks.

Best
Anton
 
Last edited: