Interconnect Impact on Electrical Waveform Lab with Results

[powertoold]

Did anyone explain why the PAPER CLIP INTERCONNECT tested the same as the ALO cable?

Doesn't this mean the test is somehow flawed?

A paper clip IC measuring the same as the ALO cable, but some other cables not measuring the same = WHAT!?

 

[myinitialsaredac]

Quote:

Originally Posted by powertoold /img/forum/go_quote.gif Did anyone explain why the PAPER CLIP INTERCONNECT tested the same as the ALO cable? Doesn't this mean the test is somehow flawed? A paper clip IC measuring the same as the ALO cable, but some other cables not measuring the same = WHAT!?

The paper clip IC did not measure the same; it tested closely, yet offered plenty of observable differences. This serves as a bit of food for thought as to what makes a cable behave a certain way. I will say that the paper clip IC was very susceptible to RFI whereas the ALO cables remained fairly unaffected - however not on par with the tube types I tested.

Dave

 

[nick_charles]

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. The hz frequency is in the audible spectrum as it is white noise.

As Pio2001 says any signal that can be expresed as a whole cycle in the 50ns window cannot contain audio frequencies (white or otherwise) since the freqency must by definition be in the Mhz range.

You set your signal generator to create white noise, but that just means a random distribution of all freqencies it can generate, it does not limit it to the audible spectrum unless you can specify the frequency range.

And if you did this, your sampling window becomes too small to represent a single (audible) wave so you are back to seeing noise.

 

[myinitialsaredac]

Quote:

Originally Posted by nick_charles /img/forum/go_quote.gif As Pio2001 says any signal that can be expresed as a whole cycle in the 50ns window cannot contain audio frequencies (white or otherwise) since the freqency must by definition be in the Mhz range. You set your signal generator to create white noise, but that just means a random distribution of all freqencies it can generate, it does not limit it to the audible spectrum unless you can specify the frequency range. And if you did this, your sampling window becomes too small to represent a single (audible) wave so you are back to seeing noise.

I hate to be quoting wikipedia but "White noise is a random signal (or process) with a flat power spectral density. In other words, the signal contains equal power within a fixed bandwidth at any center frequency."

As is the nature in our world that it cannot be completely random, it was in fact limited to the audioband.
At 100MS/s I am under the impression that I am seeing parts of the original wave even at the 50ns window.

Before you did say that "at 100mhz each sample is 10ns and you have a 50ns window so you get ~ 5."
Which makes sense because you have 100 million samples per second and you are looking at 50ns of that.
Either way if you are looking at the noise of the cable or parts of the audio waveforms the electrical differences still exist, however as is the nature of the inquisitive mind I would like to understand a lot more of what I am seeing.

Dave

 

[nick_charles]

Quote:

Originally Posted by myinitialsaredac /img/forum/go_quote.gif I hate to be quoting wikipedia but "White noise is a random signal (or process) with a flat power spectral density. In other words, the signal contains equal power within a fixed bandwidth at any center frequency."

Yes, I know this. The point being the bandwidth (range of frequencies) used as I mentioned before.



Quote:

As is the nature in our world that it cannot be completely random, it was in fact limited to the audioband.

So you set the generator to generate white noise with a bandwidth of ~ 20khz (the conservative audio spectrum) , that is good, thanks.

Quote:

At 100MS/s I am under the impression that I am seeing parts of the original wave even at the 50ns window.

Ah, sorry, some confusion here, I thought you were referring to the frequency of the test signal generated by the signal generator not the frequency that the scope uses for sampling, sorry if I did not make this distinction.

Technically yes, but here comes the but, but you created a signal with a bandwidth of 20khz. No single frequency in that signal exceeds 20Khz. No addition of the the frequencies will give you a waveform whose frequency exceeds 20khz. What you get is a very very complex waveform but still not greater than 20khz.

here is a simple example of wave addition

So take this "20K complex wave" and stretch it out and in terms of your scope it becomes practically invisible, i.e your 50ns window shows precisely 0.001 waves of the original waveform. It does not matter how frequently your scope samples the 20khz input other than it will give you a finer display.

Freq........Window........waves visible
1...........1S............1
20,000......1s............20,000
20,000......0.5s..........10,000
20,000......50ms..........1,000
20,000......50ns..........0.001


Quote:

Before you did say that "at 100mhz each sample is 10ns and you have a 50ns window so you get ~ 5."

See above

Quote:

Which makes sense because you have 100 million samples per second and you are looking at 50ns of that.

See above , you scope sampling frequency does not alter how much of the wave you see in a 50ns window, it does give you a more fine display of it.

Quote:

Either way if you are looking at the noise of the cable or parts of the audio waveforms the electrical differences still exist, however as is the nature of the inquisitive mind I would like to understand a lot more of what I am seeing. Dave

Yes, agreed the images do show a difference in noise.

 

[Pio2001]

Quote:

Originally Posted by nick_charles /img/forum/go_quote.gif As Pio2001 says any signal that can be expresed as a whole cycle in the 50ns window cannot contain audio frequencies (white or otherwise) since the freqency must by definition be in the Mhz range.

Right, but I did not say that the difference was cyclic. Maybe there are parts of it that are not included in the visible cycles, and that we can't see on the pictures, because they are too large.

Quote:

Originally Posted by myinitialsaredac /img/forum/go_quote.gif I hate to be quoting wikipedia but "White noise is a random signal (or process) with a flat power spectral density. In other words, the signal contains equal power within a fixed bandwidth at any center frequency." As is the nature in our world that it cannot be completely random, it was in fact limited to the audioband.

The fact that the signal is not completely random does not mean that it must stop at 20 kHz.
It means that the "power within a fixed bandwidth" does not converge towards the statistical average value as smoothly as in a true white noise.

 

[myinitialsaredac]

Quote:

Originally Posted by Pio2001 /img/forum/go_quote.gif The fact that the signal is not completely random does not mean that it must stop at 20 kHz. It means that the "power within a fixed bandwidth" does not converge towards the statistical average value as smoothly as in a true white noise.

You misunderstand my words. I meant that it was not completely random in our world, and also that the white noise was fixed to the audio band.

Dave

 

[digger945]

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.

Care to give us some details or hints maybe. Don't need a dissertation, just a teaser. I must say I admire your work and would gladly support any future project you may be working on, or thinking about.
Thanks,
Scott

 

[Anaxilus]

 

[Prog Rock Man]

Stumbled on this thread whilst looking for other stuff on the internet. Thought it was worth a bump to see if there was any progress.
 
The big hole in the thread is that it is not clear if or how audible the measured differences are.

 

[myinitialsaredac]

Quote:

Stumbled on this thread whilst looking for other stuff on the internet. Thought it was worth a bump to see if there was any progress.
 
The big hole in the thread is that it is not clear if or how audible the measured differences are.

The purpose of this was to view quantitative measurable differences using an oscilloscope. The next part of this was to qualify these differences to see if they are audible.


Quote:

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.

Care to give us some details or hints maybe. Don't need a dissertation, just a teaser. I must say I admire your work and would gladly support any future project you may be working on, or thinking about.
Thanks,
Scott

 
Originally my plan was to use a head with a binaural microphone in an isolating chambers and simply record the waveforms using a constant set of cans and amps and dacs and just swapping the interconnects and measure what showed up. The problem with this is finding the equipment, to date I know of only one facility already set up to do it, and was originally planning to do it with but my time got eaten up as I started college.
 
Maybe at some point in the future I will have the time and resources to further this work or atleast some derivation of it.
 
dac
 
 

 

[kwkarth]

Quote:

The purpose of this was to view quantitative measurable differences using an oscilloscope. The next part of this was to qualify these differences to see if they are audible.

Originally my plan was to use a head with a binaural microphone in an isolating chambers and simply record the waveforms using a constant set of cans and amps and dacs and just swapping the interconnects and measure what showed up. The problem with this is finding the equipment, to date I know of only one facility already set up to do it, and was originally planning to do it with but my time got eaten up as I started college.
 
Maybe at some point in the future I will have the time and resources to further this work or atleast some derivation of it.
 
dac

You've got to figure out "what" to measure before you can figure out "how" to measure it, and both, before you can actually measure anything.
 
 

 

[myinitialsaredac]

Quote:

You've got to figure out "what" to measure before you can figure out "how" to measure it, and both, before you can actually measure anything.
 
 

 
Im sorry but im unsure of what the purpose of this comment is? 
 
Regarding this lab it was to test the electrical differences that manifested when the signal from the generator was sent through two different cables and compared graphically on an oscilloscope. Theres a lot more detail to it that I realize now but that doesnt make the results any different or the test parameters, simply how the data would be explained with regards to audio. 
 
The next experiment to qualify whether these high frequency shifts caused by the cables impact audio would be measuring amplitude shifts of different frequencies in the 20-20k range which is again not a terribly involved process but requires more sensitive tools than I have access to.
 
dac

 

[Mambosenior]

“...hence the mega-buck cable manufacturers and their victims (moi?) await your results with terror.”
 
(Sorry, do carry on.)

 

[kwkarth]

Quote:

 
Im sorry but im unsure of what the purpose of this comment is? 
 
Regarding this lab it was to test the electrical differences that manifested when the signal from the generator was sent through two different cables and compared graphically on an oscilloscope. Theres a lot more detail to it that I realize now but that doesnt make the results any different or the test parameters, simply how the data would be explained with regards to audio. 
 
The next experiment to qualify whether these high frequency shifts caused by the cables impact audio would be measuring amplitude shifts of different frequencies in the 20-20k range which is again not a terribly involved process but requires more sensitive tools than I have access to.
 
dac

The purpose of the comment was to indicate that there is no guarantee that you are currently measuring the factors actually responsible for attributed sonic differences.
What is this "high frequency shift" to which you refer?