|
Vinyl Asylum Welcome Licorice Pizza (LP) lovers! Setup guides and Vinyl FAQ. |
|
|
|
Vinyl Asylum Welcome Licorice Pizza (LP) lovers! Setup guides and Vinyl FAQ. |
|
In Reply to: Does this test really measure IMD? posted by sser2 on May 24, 2004 at 14:56:37:
Hi,I disagree with most of your analysis above (if I understand your points correctly). IMD is caused simply by the effects of non-linearity on a multi-tone signal, and it results in sidebands being produced with frequencies which are combinations of integer multiples of both tones.
2nd order intermodaulation distortion is an effect of 2nd harmonic generation, because if one analyses the spectrum of a signal comprising of two tones f1 and f2, which is subject to 2nd order harmonic distortion, the spectrum will comprise of the following peaks: f1, 2xf1, f2-f1, f2, f2+f1, 2xf2.
For example, the spurious frequencies resulting from 2nd harmonic distortion of a two tone signal comprising of a low tone of 200Hz and a high tone of 4000Hz are as follows:
200Hz (fundamental of low tone);
400Hz (2nd harmonic of low tone)
3800Hz (2nd order IM product of 200Hz and 4000Hz)
4000Hz (fundamental)
4200Hz (another 2nd order IM product of 200Hz and 4000Hz)
8000Hz (2nd harmonic of high tone)So the two second order peaks in John Elison’s graphs at 3800Hz and 4200 Hz are certainly intermodulation products – by definition!
Amplitude modulation could be viewed as beihg similar to 2nd order intermodulation distortion because, for example, the two sidebands resulting from modulating a 4000Hz carrier signal with a 200Hz signal are 3800Hz and 4200Hz (same as above). However, amplitude modulation does not produce 2nd harmonics of the fundamentals (eg no peaks at 400Hz and 8000Hz in the example above).
Amplitude modulation is completely different from the periodic oscillation in the envelope function produced when two periodic signals with different frequencies are added together.
For a linear system, when two tones are added, the spectrum will only show the two original frequencies. So if there were no distortion in the example above, there would not be any sidebands around the 4,000 Hz signal, even though the envelope function of the combined signal oscillates in the time domain.
For John Elison’s tests with the Dynavector cartridge, I expect that the 2nd order IM products (at 3800 Hz and 4200 Hz) will be even larger than those produced when the experiment was carried out with the Shure cartridge. This is because John has already demonstrated that the 2nd order harmonic generation of the Dynavector is also larger than that of the Shure. And, as pointed out above, 2nd order IM products are produced by 2nd order non-linear effects.
The critical point in John’s measurements, from a sonic point-of-view, is probably the higher order harmonic generation, and the corresponding IM products. For example, in John’s graphs for the Shure cartridge, side bands can be seen at 3600Hz and 4400Hz, and also at 3400Hz and 4600Hz. These are IM products resulting from the following combinations of the two tones:
f2 +/- 2 x f1 (third order IM products);
f2 +/- 3xf1 (forth order IM products).Similarly there are higher order effects at higher frequencies in the spectrum, such as: 7800Hz and 8200Hz (third order IM products given by 2 x f2 +/- f1);
7600Hz and 8400Hz (forth order IM products given by 2 x f2 +/- 2 x f1)It is often stated that the higher order odd harmonics produce an unpleasant sound to the ear, so the critical comparison might be carried out by looking at the amplitudes of the peaks at 3600Hz and 4400Hz, and similarly 7800Hz and 8200Hz (all of which are third order IM products).
Best regards,
BK
This post is made possible by the generous support of people like you and our sponsors:
Follow Ups
- Yes - most difinitely! - bkearns 05/25/0407:03:21 05/25/04 (0)
Top of Page ]
[ Contact Us ]
[ Support/Wish List ]
[ Copyright © 2025, Audio Asylum, all rights reserved ]
