Rerurn to Romy the Cat's Site


In the Forum: Melquiades Amplifier
In the Thread: Building Melquiades: Chronicle of full-range
Post Subject: Re: Sined or sinnedPosted by hagtech on: 7/28/2006
 romy wrote:
I have quite rudimentary and not very well performing generator


That's what I thought.  I see a point at the tips of the sinewave, a dead giveaway to a ICL8038 style waveform generator.  I wanted to see the input because much of the distortion could be caused by it.

There is nothing wrong with seeing the 2nd order distortion - in fact it appears to be quite small for such a high power level.  It's just nice to be able to see it.  I understand how you like to FEEL the amplifier as you sweep the frequency.  I do the same thing.  You watch the distortions and glitches and look for patterns and trends.  It is easy to spot irregularities that cannot be seen in a static snapshot.  Try also to do the same thing with a triangle or sawtooth input. 

Guy suggests the square wave.  I'm absolutely anal about transient response.  I spend more time on step response and nice looking square waves than anything.  It gives you a very good understanding of phase and temporal relationship with the amplitude.  Now if we can only get more loudspeaker designers to do this.

Oh, I just thought of something.  The phase (polarity) switch needs to be ahead of the Melq.  And here's my theory:  The 2nd order distortion is polarity sensitive.  It must match the input polarity so that it adds properly.  Otherwise it subtracts.  Vocals are nonlinear and mimic a heavy 2nd order distortion.  If you amplify it with a nonlinearity of the wrong polarity then you partially cancel, or worse, turn it into 3rd order.  Think of it this way.  The transfer function of a PP amplifier is an "S".  For an SE amp it is a "J".  If you add two SE amps in series you either get a stronger "J" or it turns into an "S", depending on polarity.  I think it is important to match absolute polarity in this fashion.  Sound in a gas is similar to SE (1/P function).  A kick drum transient that pushes air (compress) before it pulls (rarefaction) will sound wrong if played backwards.  I think it also makes a difference if you add 2nd order distortion in the wrong way.  Sound in a gas has a curved function, so does your SE amp.  My theory is that it is better to have them sum rather than cancel.

jh

Rerurn to Romy the Cat's Site