1. This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
    By continuing to use this site, you are consenting to our use of cookies.

    Dismiss Notice

Interconnect Impact on Electrical Waveform Lab with Results

Discussion in 'Sound Science' started by myinitialsaredac, Oct 2, 2008.
First
 
Back
1 2 3 4 5
7 8
Next
 
Last
  1. helicopter34234
    Quote:

    however this is not the place to discuss this.



    I would argue that in general, in setting up any well run experiment, you should have an idea of the mechanisms contributing to the differences in your experimental measurements. Otherwise you won't have any idea what parameters you need to specifically control for and vary. You don't have to be able to fully explain exactly how or how much each factor affects the results, but you should be aware of all potential influences so as to control for them in the experiment. I am not implying that you personally did not fully account for all variables in your experiment, I am just responding to your comment.

    For instance, lets say (for the sake of arguement) that the only difference in cable performance (relevant to a typical audio user) is the ability to shield RFI in a certain frequency range typically experienced when routed near audio equiptment (maybe particular RFI ranges generated by amplifier power supplies). If you were setting up the experiment without considering what could possibly contribute to observed performance in the field (once again I am not implying that you did not), one may not specifically consider testing the cable response to the exact RFI intensity and frequency that it would experience in real usage. In such a case, the experiment did not accurately reflect real life usage conditions and therefore less conclusions can be drawn from the experiment.

    Coming from a scientific background, I know that very often the initial round of experiments do not fully control for all relevant variables. Then after discussion of the actual model to explain the results, it becomes apparent that there were relevant variables that were not fully controlled for. Therefore experimentalists and theorists should not operate independently.
     
  2. myinitialsaredac
    Quote:

    Originally Posted by helicopter34234 /img/forum/go_quote.gif
    I would argue that in general, in setting up any well run experiment, you should have an idea of the mechanisms contributing to the differences in your experimental measurements. Otherwise you won't have any idea what parameters you need to specifically control for and vary. You don't have to be able to fully explain exactly how or how much each factor affects the results, but you should be aware of all potential influences so as to control for them in the experiment. I am not implying that you personally did not fully account for all variables in your experiment, I am just responding to your comment.

    For instance, lets say (for the sake of arguement) that the only difference in cable performance (relevant to a typical audio user) is the ability to shield RFI in a certain frequency range typically experienced when routed near audio equiptment (maybe particular RFI ranges generated by amplifier power supplies). If you were setting up the experiment without considering what could possibly contribute to observed performance in the field (once again I am not implying that you did not), one may not specifically consider testing the cable response to the exact RFI intensity and frequency that it would experience in real usage. In such a case, the experiment did not accurately reflect real life usage conditions and therefore less conclusions can be drawn from the experiment.

    Coming from a scientific background, I know that very often the initial round of experiments do not fully control for all relevant variables. Then after discussion of the actual model to explain the results, it becomes apparent that there were relevant variables that were not fully controlled for. Therefore experimentalists and theorists should not operate independently.




    Indeed this is true when talking about variables outside of the original build such as RFI. In this experiment the variables discussed in your question were the individual construction of the cable. This is not important to real life usage in the sense that running waveforms through different cables the cables need to be different in construction and was never meant to be tested in this experiment. This experiment was simply to test different construction of cables and see how waveforms pass through them in the same environment to see if they move waveforms differently.
    When discussing how the individual construction of the cable changes the waveforms one must be able to have many cables built the same way with only one variation per cable from the original built in many ways.

    In your RFI experiment discussion it is pertinent to know your RFI sources, intensity, frequency, etc. however that is an outside variable that must be controlled. In that experiment knowing how the cabling construction differences impact the RFI susceptibility is unimportant.

    Dave
     
  3. myinitialsaredac
    Quote:

    Originally Posted by Omega /img/forum/go_quote.gif
    Sure you can, the oscilloscope is a tool like any other, with the given limitations and inaccuracies. I agree with the thrust of your praise; objective measurement with appropriate instruments and positive/negative controls is very appropriate! Good going OP! I'm just saying that 100% belief in instrumentation is not so different than the 100% belief in human senses that you were knocking [​IMG]. Skepticism is the first thing! If you're a fan of Feynman, "The first principle is that you must not fool yourself--and you are the easiest person to fool. So you have to be very careful about that. After you've not fooled yourself, it's easy not to fool other(s)."



    Agreed, and one can argue that human hearing is far beyond conventional oscilloscopes bitdepths etc.etc. It is important that we do not fully rely on machines to tell us what we can hear, because our hearing may be more accurate. These tests when they do show differences on the scope offer a quantitative measure that when applied to our higher bit-depth hearing if you will experiences even more drastic results. So at a lower bit depth differences are become more emphasized at a higher accuracy measurement.

    Dave
     
  4. edstrelow Contributor
    Quote:

    Originally Posted by Caution /img/forum/go_quote.gif
    Seems like quite a lot of trouble for a high school student [​IMG]

    I must say though, to me this experiment seems pointless unless you actually want to see what type of output wave a specific cable produces kind of like the frequency charts for headphones on headroom. Hasn't it been proved time and time again that cables will produce different waves due to the material used? The results seem pretty obvious to me...




    Hardly pointless, there have been tons of thread under tweaks in which numerous jackasses have insisted that cabling makes no difference whatsoever, cable is cable etc.

    However I am a little unclear on what I am supposed to be seeing in the photos. Some verbal description of the salient aspects of each trace is called for.
     
  5. myinitialsaredac
    Quote:

    Originally Posted by edstrelow /img/forum/go_quote.gif
    Hardly pointless, there have been tons of thread under tweaks in which numerous jackasses have insisted that cabling makes no difference whatsoever, cable is cable etc.

    However I am a little unclear on what I am supposed to be seeing in the photos. Some verbal description of the salient aspects of each trace is called for.




    "Let me take some time to explain how this worked. I used a bnc to rca adapter coupled with an rca to dual rca adapter out of the waveform generator in order to send the exact same signal to each channel. I tested the ALO A against ALO A in order to ensure the splitter was not causing any differences. The first channel is yellow and the second channel is green. When the wave appears completely white that means that each channel is equal, so no yellow or green is showing. When there is a difference you will be able to see yellow and green separately meaning the waves are not equal. I was unable to calculate a magnitude due to the nature of white noise being almost completely random and because it changes so quickly. In these tests the upper window is a 50ns clip of the wave and the bottom window is a 10ns clip of the wave."

    Basically to sum that up if the wave is white that means the cables are equal at those points. If there is distinct yellow and green it means there are differences between the cabling. In some of the shots it appears that there are 4 waves (or 2 white waves) because my camera caught some ghosting; simply look to the wave with the highest contrast and that is the wave that is present on the screen at the time.

    Dave
     
  6. nick_charles Contributor
    Quote:

    Originally Posted by edstrelow /img/forum/go_quote.gif
    Hardly pointless, there have been tons of thread under tweaks in which numerous jackasses have insisted that cabling makes no difference whatsoever, cable is cable etc.



    and just as many jackasses have insisted on huge audible differences between cables with no evidence apart from seriously flawed sighted listening.

    The tests in this thread have cerrtainly shown electrical differences, not alas any further proof whether they are audible or not.
     
  7. myinitialsaredac
    Quote:

    Originally Posted by nick_charles /img/forum/go_quote.gif
    and just as many jackasses have insisted on huge audible differences between cables with no evidence apart from seriously flawed sighted listening.

    The tests in this thread have cerrtainly shown electrical differences, not alas any further proof whether they are audible or not.




    Agreed,
    I feel that this argument is about to get out of hand and would kindly ask you to continue this discussion else where.

    I will reconfirm that this experiment can not show what cable is best or even if the differences are audible. What it does show is that there is the propensity for the electrical differences shown here to be audible.

    I will also reconfirm that I am working on a much more refined experiment to show whether or not the differences are audible. However upon doing that I am still unable to show what cable is "best" because if the differences are audible it comes down to personal taste, just like headphones, speakers, amps, dacs, etc.etc.

    Dave
     
  8. nick_charles Contributor
    Quote:

    Originally Posted by myinitialsaredac /img/forum/go_quote.gif
    I will also reconfirm that I am working on a much more refined experiment to show whether or not the differences are audible



    Cool !
     
  9. hawat
    Thank you very much for your hard work, this is a very interesting thread.
     
  10. Pio2001
    Thank you for the results.

    I can see written in picture 30 "2.00 GSa/s" in the top left had corner of the screen. I suppose that it means 2 billions samples per second.
    So the oscilloscope gets 20 samples in 10 ns and 100 samples in 50 ns.

    Since in the lowest part of the screen, the waveforms oscillate very quickly, I think that 50 ns and 10 ns are not the lenght of the windows, but of the divisions. Otherwise, there would not be enough samples in each window in order to draw such fast oscillations.

    There are ten divisions per window. Then the upper window is rather 500 ns wide, and the lower one rather 100 ns wide.

    We can see, on the pictures where the waveform differ, that the green and yellow signals cross each other about once per division at least in the lower picture. The period of this difference is then at most 20 ns.
    Which means that the frequency at which the cables differ is at least 50 MHz.

    They may differ in lower frequencies too, but the pictures can't show it, because the time widow is too short, and if a lower frequency difference exists, which would lead to parallel, but distinct, waveform, the effect is here masked by the higher frequency differences.

    The 50 MHz differences are probably caused by the bandwidth of the cables. In order to test this hypothesis, you can measure the amplitude of a sinewave at different frequencies (or directly draw the frequency response of the cables if you have got an analyzer). The cables that behave differently should have a different bandwidth.
    you can also generate a wav file with white noise, apply a parametric equalizer in order to remove its upper frequencies, then compare the waveforms before and after the lowpass, and see if they look the same as the ones on the oscilloscope, frequency apart.

    The RFI measurments can show differences between cables, but their amplitude should heavily depend on the output impedance of the generator, that short-circuits them. So measuring their amplitude would only be relevant when measured from the output of a CD player or a preamplifier.
     
  11. Pio2001
    It would be very interesting to know the output impedance of the generator, and the input impedance of the oscilloscope, as the effect of the cable depends of them.
    The input impedance of an amplifier is between 10 and 50 kOhms, and the output impedance of a CD player should be between 100 and 500 Ohms.
    If the input impedance of the receiver is lower, the cable has more effet. If it is higher, the cable has less effect.
     
  12. myinitialsaredac
    Quote:

    Originally Posted by Pio2001 /img/forum/go_quote.gif
    Thank you for the results.

    I can see written in picture 30 "2.00 GSa/s" in the top left had corner of the screen. I suppose that it means 2 billions samples per second.
    So the oscilloscope gets 20 samples in 10 ns and 100 samples in 50 ns.




    Quote:

    Originally Posted by Pio2001 /img/forum/go_quote.gif
    The 2.00 GSa/s is referring to how fast the oscilloscope is sampling the samples from the pulse generator, think of it like processor speed. The pulse generator was supplying pulses at 100MS/s (100 million) so it gets 5 samples in 50 and 1 sample in 10.

    Since in the lowest part of the screen, the waveforms oscillate very quickly, I think that 50 ns and 10 ns are not the lenght of the windows, but of the divisions. Otherwise, there would not be enough samples in each window in order to draw such fast oscillations.

    There are ten divisions per window. Then the upper window is rather 500 ns wide, and the lower one rather 100 ns wide.




    Actually the entirety of the window is 50ns so that gap is only 5ns and 1ns. The hz frequency is in the audible spectrum as it is white noise.

    Quote:

    Originally Posted by Pio2001 /img/forum/go_quote.gif
    The RFI measurments can show differences between cables, but their amplitude should heavily depend on the output impedance of the generator, that short-circuits them. So measuring their amplitude would only be relevant when measured from the output of a CD player or a preamplifier.



    Indeed that was just a bit of a side experiment. Again measuring the cables as a circuit with a null signal should verify that some are subject to rfi more than others as a circuit with a signal and a source that they are transferring between.

    Thanks for all the great questions!

    Dave
     
  13. myinitialsaredac
    Quote:

    Originally Posted by Pio2001 /img/forum/go_quote.gif
    It would be very interesting to know the output impedance of the generator, and the input impedance of the oscilloscope, as the effect of the cable depends of them.
    The input impedance of an amplifier is between 10 and 50 kOhms, and the output impedance of a CD player should be between 100 and 500 Ohms.
    If the input impedance of the receiver is lower, the cable has more effet. If it is higher, the cable has less effect.




    The input impedance of the oscilloscope was set to a whopping 1mOhm, so basically any input impedance of an amplifier would be accounted for. I do not know the output of the pulse generator.

    For point of reference I also tried the oscilloscope at 50 ohms input resistance and the results were of a higher amplitude, however I photographed the 1mOhm.

    Dave
     
  14. Pio2001
    Quote:

    Originally Posted by myinitialsaredac /img/forum/go_quote.gif
    Actually the entirety of the window is 50ns so that gap is only 5ns and 1ns.



    Picture 11 seems to be one of the most detailed.
    So the small irregualities on the curves must not be taken into account. the oscilloscope measures 2 samples per division in the lower picture, and draws an interpolated line between them.

    I actually see that the green and yellow curves change sides only three times in this picture. Three times in 10 ns means that the complete period is 2 x 10/3 = 6 ns. Which gives a frequency somewhere between 150 and 200 MHz.
    I'm not talking about the frequency of the white noise, but of the difference between the cables that we can see on the pictures.

    We can say that they are different at least at this frequency.

    Quote:

    Originally Posted by myinitialsaredac /img/forum/go_quote.gif
    For point of reference I also tried the oscilloscope at 50 ohms input resistance and the results were of a higher amplitude, however I photographed the 1mOhm.



    Then the differences between the cables when they are connected in a hifi system should be between the two.
     
  15. myinitialsaredac
    Quote:

    Originally Posted by Pio2001 /img/forum/go_quote.gif
    Picture 11 seems to be one of the most detailed.
    So the small irregualities on the curves must not be taken into account. the oscilloscope measures 2 samples per division in the lower picture, and draws an interpolated line between them.

    I actually see that the green and yellow curves change sides only three times in this picture. Three times in 10 ns means that the complete period is 2 x 10/3 = 6 ns. Which gives a frequency somewhere between 150 and 200 MHz.
    I'm not talking about the frequency of the white noise, but of the difference between the cables that we can see on the pictures.

    We can say that they are different at least at this frequency.



    Then the differences between the cables when they are connected in a hifi system should be between the two.





    Or not necessarily different at a given frequency but the frequency of differences between the cable if I am understanding you correctly?

    Yes the electrical differences between the cable should fall between the 50 ohm and 1mOhm viewed differences.

    Dave
     
First
 
Back
1 2 3 4 5
7 8
Next
 
Last

Share This Page