[myinitialsaredac]

UPDATE: I received an email back from George Cardas this morning explaining a bit about his oscilloscope for the experiment he conducted and a bit of theory in general about why there are differences. His information was quite helpful in determining the type of oscilloscope I will need.
Using this information, I have revised the experiment a small amount in the original post.

I will be asking for an oscilloscope that can do calculations of the differences between two input waves (i.e. an M-Line) , has RCA inputs or BNC-RCA converter plug, and has a greater dynamic range than 60db with at least 20MS/s (million samples per second). Optimally 100db dynamic range with greater than 80MS/s.

Dave

 

[myinitialsaredac]

UPDATE: Sent out e-mails seeking knowledge and input on the experimental procedures to TheCablePro and Meier Audio.

Regards,
Dave

 

[myinitialsaredac]

No one has input on the procedures/variables/constants?

Dave

 

[myinitialsaredac]

UPDATE: Another revision of the OP to include the equipment needed for this test.

Regards,
Dave

 

[nikongod]

Quote:

Originally Posted by myinitialsaredac /img/forum/go_quote.gif No one has input on the procedures/variables/constants? Dave

Dont worry about sine or triangle waves. a 10K hz square wave tells you more about what something does to the signal (without a spectrum annaliser) than a sine wave.

As a VERY significant variable: try a few different effective "source impedances" by putting resistors in series with the source. Compare the output at the end of the cable to the input BEFORE the resistor. Subtract the resistor "sound" (you can measure this easily) and there you go.

 

[myinitialsaredac]

Nikongod,

That is a very helpful idea, I will definitely implement the resistor test, which, if I'm understanding correctly, will show how much the cable is influenced by different impedance equipment correct and simulate different impedances?

Dave

 

[myinitialsaredac]

UPDATE: Mike from Headroom responded to my plea for help, and gave me some very useful information regarding the setup.

I refined the procedures and equipment list in the OP.

Dave

 

[myinitialsaredac]

UPDATE: Another revision of the experimental procedures in the OP adding specifics to the setup of the oscilloscope and pulse generator.

Regards,
Dave

 

[myinitialsaredac]

Still looking for input, comments, critiques, on experimental procedures/variables/control.

Dave

 

[myinitialsaredac]

UPDATE: Recieved more knowledge from George Cardas, very informative and helpful.

Dave

 

[kevin gilmore]

You are going to have significant problems on the equipment side
of this. All of the digital oscilliscopes you are likely to get your hands
on are likely 8 bit scopes with very limited dynamic range. The units
with 10 or 12 bit digitizers are very expensive. Sort of like these things
Agilent | U1070A Acqiris 12-bit High-Speed PCI Digitizers

but even this is not really enough. You will get far more information
by using the right tools. You will need a real pulse generator with
very fast rise and fall times, not a signal generator. You will need
a TDR (time domain reflectometer) and you will need a baseband
spectrum analyzer with 100+ db dynamic range.

Then you will find huge differences between the cables you are about to
test. Which you can largely mitigate by using the proper source and
termination resistors.

 

[myinitialsaredac]

Kevin,
Thank you for the input.
It is very helpful.

If I am understanding correctly, I can use a TDR to measure the reflections caused by impedance differences in the cable right? I will ask to see if the have one.

As for a baseband spectrum analyzer, is that a fancy name for an oscilloscope? I will be asking for the highest resolution scope with the largest dynamic range, hopefully they can supply something in the 12-14 bit depth with 100+db dynamic range.

If you could explain a bit more I would be greatly appreciative.

Regards,
Dave

 

[myinitialsaredac]

UPDATE: Edited equipment list and experimental procedures in OP.

Regards,
Dave

 

[Pio2001]

Quote:

Originally Posted by Caution /img/forum/go_quote.gif 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...

Measurments on cables are rare, except in radiofrequencies. Do you know of any measurment made in audio frequencies on interconnects ?

Quote:

Originally Posted by MoodySteve /img/forum/go_quote.gif Some would say that testing cables using clean signals such as sine waves doesn't accurately simulate their response to the considerably more complex waveforms that music produces.

It does, because cables behave linearly from an algebric point of view (their THD and IMD are strictly null, and their noise is negligible). Therefore it is relevant to assume that a complex signal is a sum of simple ones.

Quote:

Originally Posted by myinitialsaredac /img/forum/go_quote.gif No one has input on the procedures/variables/constants? Dave

You can find here some measurments done on a set of various interconnects :
homecinema-fr.com • Afficher le sujet - Résultats du test en aveugle - câbles de modulation

Get the file at the bottom of the first post. This can help you figuring out the magnitude order of the expected differences.

You plan to use triangle and square waves. This raises the question of the frequency band studied. Should the study be restricted to audio frequencies (20 to 20 kHz) ? In this case, triangle and square wave, whose frequency range are infinite, should not apply.

Beware that a square wave of 10 kHz is nothing else than a sine wave of 10 kHz, plus a sine wave of 30 kHz, plus a sine wave of 50 kHz, then 70 kHz, 90 kHz etc.
Restrict the square wave into the audio band, and all that remains is a pure 10 kHz sine wave !

If you restrict yourself in audio frequencies, then the time accuracy of the oscilloscope becomes a minor factor, and you can get much more accurate recordings from a computer soundcard than from an oscilloscope (around 17 bits of accuracy, given their signal-to-noise ratio).

The attenuation of the interconnects are strongly dependant on the impedance of the source and of the load. The normalized source impedance in hifi is 470 Ohms, and the normalized load impedance in 47 kOhms. Oscilloscopes should have a higher impedance, resulting in less distorsion.

Measuring time-domain reflexions seems irrelevant, since they will be defined by your RCA/BNC adapters, impulse and signal generators, and oscilloscope, and will have nothing to do with the reflexions in a normal setup.

Anyway, the expected difference are extremely small. The highest chances to get a measurable difference is using very long cables (get a 5 or 10 meters extension cord for interconnects, for example), or analysing very high frequencies (more than a MHz), where the attenuation should be more important.
Given your setup, the square wave meets this condition, as it features high frequecies harmonics, far outside the audio band.

Good luck, and keep us informed.

 

[myinitialsaredac]

Pio,
Thank you for the input, I have spoken with George Cardas and Mike Olsen and they both recommend using the square waves, but I will also include the sine wave and triangle wave for validity. I am still working on a set frequency, and will likely do multiple frequencies of each wave.
I plan to ask for a 14 bit resolution oscilloscope with a large dynamic range, and this should yield at least some results.

The time-domain reflection test is almost a separate deal, I will be using the TDR to look at reflections of the wave in the cable caused by various impedance differences. I will be asking George Cardas and Mike about their input on this test.

Thanks again for the input,
Regards,
Dave