Amplifiers with low output impedance produce very little distortion in the voltage when i measure with the Klippel Analyser. When i measure distortion in the current (the Klippel has a sensor in series with the speaker, so can measure current) it can be very high.
The loudspeaker is a current driven transducer ( conventional dynamic type, that is).
Interestingly loudspeakers with Alnico magnets produce much less distortion in the current and magnet systems can be optimised to lower that effect. This is such a long story about the whys and hows that i spend a good part of my life to get a handle on that problems.
So a certain speaker connected to a certain amp produces a certain distortion profile.
Is a combination of amp and speaker that produces the effect. Think about it, without the amp the speaker makes no sound and without the spaeker the amp makes no sound.
Ohhhh... this is to primitive for this thread, sorry...
The loudspeaker is a current driven transducer ( conventional dynamic type, that is).
Interestingly loudspeakers with Alnico magnets produce much less distortion in the current and magnet systems can be optimised to lower that effect. This is such a long story about the whys and hows that i spend a good part of my life to get a handle on that problems.
So a certain speaker connected to a certain amp produces a certain distortion profile.
Is a combination of amp and speaker that produces the effect. Think about it, without the amp the speaker makes no sound and without the spaeker the amp makes no sound.
Ohhhh... this is to primitive for this thread, sorry...
That's very interesting.
Is the degree of current distortion related to the amount of GNFB?
A virtually perfect way to separate the back EMF of the loudspeaker from the drive voltage/current, would be to make a large, linear, current output amplifier, perhaps 20A capability, and drive it into a .1 ohm power resistor. This would simulate a normal power amp with a damping factor of 80 or so, not bad, over a large bandwidth. Then the input voltage could be compared to the voltage generated across the power resistor and we could easily measure the reverse voltage/current due to the speaker.
Finally, we drive the current amp with steps, pulses, sweeps, etc. It would be VERY revealing, especially what the typical power amp has to absorb through its output stage, and we will probably find real problem speakers.
Finally, we drive the current amp with steps, pulses, sweeps, etc. It would be VERY revealing, especially what the typical power amp has to absorb through its output stage, and we will probably find real problem speakers.
The loudspeaker driver is a more or less non linear load.
The distortion seen in the current of the amplifier is due to being loaded by a nonlinear load.
The amount of distortion have a direct relationship to the output impedance of the amplifier.
Setting a one ohm resistor in the return from the speaker and measuring the voltage drop shows once again that the distortion is back to where it is nominally for the amplifier in use.
This is a complete nonissue to me and the observed distortion in the current do not show in the achoustical output from the loudspeaker driver.
BUT the output impedance as such will quite markedly effect the measured frequency response of the loudspeaker in use.
I am not speaking about the observed measured changes in the actual output of the amplifier instead I am observing the actual driver voltage as seen after the passive crossover.
Many such passive crossover design will show resonant phenomena giving much greater change to each driver than can be seen at the output of the amplifier.
The Spendor BC1 is such a speaker where a 7 dB difference in the voltage appearing at the terminals of the supertweeter the STC4001 could be seen using either the ARC D76A or the Crown DC300A due to the formers 1.5 ohm at 10 kHz output impedance compared to the DC300A´s 0.3 ohm at 10 kHz output impedance.
The DC 300A sounded like it was broken emphasising clicks and scratches from 78 shellac disks out of all proportion while the D76A sounded very good with very subdued clicks and scratches.
Mimicking the D76A´s output impedance on the DC300A made it sound almost the same as the D76A.
The effect of making the output impedance almost the same makes the speaker behave almost the exact same in both cases thus the sound becomes very close indeed.
No wonder that Bob Carver on one of his amps has a "Tube" output in addition to the normal one and the only difference is a 1.5 ohm resistor in series with the "Tube" output terminals.
There are so many pitfalls trying to pin down audible differences and most of the things one can measure as "Distortion" has zilch effect on the perceived result in many cases as whats being heard is due something entirely else in my experience.
Only when you connect a speaker directly to a Vin x voltage gain = Vout amp ( an amp with high damping factor and global feedback) is it true that distortion in the current is simply short circuited and does not return to the speaker and does not not circulate in the feedback or groundloop. Add any impedance ( cables, lossy parts in the crossover etc.) and the situation changes from that over simplified model. One good argument for active spaekers.
John, your proposed test is very interesting. You are simply a g.......
John, your proposed test is very interesting. You are simply a g.......
Well, ever onward. Yesterday, I was told that a competitive peer in audio design was getting his own custom transistors made, and this is why his audio equipment was that much better than anyone else. Demian knows this designer. What about it Demian?
This other designer is also called a 'g', even by me, but he allegedly tells his customers that I am not a real engineer. Sound familiar? ;-)
This other designer is also called a 'g', even by me, but he allegedly tells his customers that I am not a real engineer. Sound familiar? ;-)
Jürgen Ultee, that designed the new Inductive Phonostage for Van den Hul got some input transistors made by That to his specs. He uses a parallel symmetric common base stage at the input i think.
This is not a big deal and less expensive then you might think.
I can remember that he had to order a thousand pieces and the price was around 3000,-€
For the whole sheebang. I mean, this is not ceap but for 3,-€ a piece i think it´s ok.
Unfortunately the standard THAT arrays are not so well matched and low noise as it should be.
This is not a big deal and less expensive then you might think.
I can remember that he had to order a thousand pieces and the price was around 3000,-€
For the whole sheebang. I mean, this is not ceap but for 3,-€ a piece i think it´s ok.
Unfortunately the standard THAT arrays are not so well matched and low noise as it should be.
Joachim,
What is the relation of this current distortion to the distortion of the radiated signal, if any?
Edit: You already answered this in post 3727. Sorry.
jd
Last edited:
- Status
- Not open for further replies.