Cable Distortion Measurements: Part Deux

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Sy, I will tell you this:
I am making a series of fairly difficult measurements with the test equipment that I have. My ST was designed to work to approximiately -100dB, at measurement levels about .3V or more. With added FFT processing, I can separate the harmonics from the null residue and reduce the noise floor.
I CAN produce measurement artifacts, if I am not careful. For example, if I load the oscillator buffer excessively, I can begin to produce extra 2'nd and 3'rd harmonic distortion. However, when I find a 'clear' area of measurement, then I can make measurement comparisons.
Once I have a measurement set-up that seems OK with my reference cables, I make NO changes to the measurement equipment. I just change the cable by removing it from the connector and replacing it with another.
At this point, I can see differences between different cables, both of different lengths and the same length. I can also see differences with similar cables, but with different use patterns, such as the amount of signal passed through them over time.
I can also see minor distortion 'bumps' build up, if my reference cable's connectors and the the whole external assembly are not periodically cleaned with industrial purity isopropyl alcohol The 'bumps' go down or away, after cleaning.
I monitor the input of the analyzer with a 350MHz Tek 485 scope, running maximum bandwidth, given the test conditions. I see no oscillation on the scope.
At this point I can do no more to show anything.
If Bruno's results are again a null, then the comparison is ended.
 
John, thanks for the additional details.

This may be asking a lot, and I won't at all take offense at a "no" answer (nor draw any inferences; lots of people just like their privacy), but would it be possible for you to demonstrate for me how the measurement is done? I'd like to do an informal gauge R&R by connecting and disconnecting the same cable, playing around with the dress of the cable, and so forth.

As I said, my own professional expertise is the physics of conductive materials, and coincidently, fourier transform techniques (I was the #2 guy in R&D at Nicolet back when they were the Big Guys in FFT instrumentation, which certainly dates me!). So this really seems to be up my alley. I've got no axes to grind, I just want to understand your measurements and what they really mean.

As recompense for your time, I'll happily cart along a bottle of something rare and vinous.
 
One person I like to see participate here is Siegfried Linkwitz he was the BIG GUY who constructed FFT monsters at HP a very nice technician that can sweeping the fog away.

Another good techie Nelson Pass I will remember that he made a measure on cables years ago maybe this can be used as a reference here.
 
kamskoma said:
About test equipment, can I use a simple THD analyzer like the HP 332A to null out the fundamentals and put this signal into a soundcard like the DAL CardDeLuxe togetherwith a good FFT software, is this a good measurement device?.

Well, the graphs of the measurements John did some time ago were done by feeding the residual from his 1700B into the built in sound card in his Macintosh using a demo version of Mac The Scope software.

What I think is, it may be very simple to make a reference measurement that will go for all comming measurements or?

Just measure the nulling device with a good soundcard Card DeLuxe then you have a very good an cheap instrument.

If not! where is the fault, and can I do it with even cheaper nulling device and soundcards.

The issue here isn't determining whether John is measuring distortion or not. He clearly is. The issue is whether the distortion he's measuring is in fact being produced by the cables or by his 1700B.

So far, using a system known to have lower distortion and greater resolution than what John's using, none of the high order distortion products that John routinely measures have shown up. John says he measures distortion with virtually every cable he's used, but even when measuring a good 20dB below where John's measuring, nothing has turned up.

The measurements that Bruno did had been dismissed by someone else because Bruno wasn't measuring the same cables as John. So this is an attempt to repeat the measurements with both John and Bruno measuring the same cables.

se
 
kamskoma said:
One person I like to see participate here is Siegfried Linkwitz he was the BIG GUY who constructed FFT monsters at HP a very nice technician that can sweeping the fog away.

Another good techie Nelson Pass I will remember that he made a measure on cables years ago maybe this can be used as a reference here.

If you think you can interest either of these gentlemen in doing low level distortion measurements on cables, that would be a plus. I know Nelson has an AP rig.

se
 
Disabled Account
Joined 2003
john curl said:
Frank, please quit putting your foot into this.

ouch!

SY said:
(I was the #2 guy in R&D at Nicolet back when they were the Big Guys in FFT instrumentation, which certainly dates me!).

Wow! we have some seriesly knowledgeable folks here.

Steve Eddy said:


Well, the graphs of the measurements John did some time ago were done by feeding the residual from his 1700B into the built in sound card in his Macintosh using a demo version of Mac The Scope software.


se

I didn't follow this thread as closely but two things here:

1) did John repeat the tests? maybe the sound card is faulty the time John was testing it?
2) did the distortion come from the source? what if John used a very short or better cable? Would John have seen anything materially different?

I think we are certainly making progress here, from anecdotes to concrete measurements. Keep up the good work, folks.
 
Wow! we have some seriesly knowledgeable folks here.

Not compared to the #1 guy there at the time, a spooky-smart Cajun named Warren Vidrine. I was seriously lucky to be able to get a job working for him, and I've even forgiven him for introducing me to my first wife. We're still close friends, even after all these years (damn, was that really 20 years ago?!).

It's sobering for me to realize how sophisticated the stuff we were doing then seemed to be, yet it's all trivially easy these days with PCs and modern software.
 
Hello -

I haven't been following this whole thing very well, but if I understand correctly, you are trying to have somebody in Europe verify John Curl's cable measurements with an Audio Precision.

If that is really what's going on, this is a colossal waste of time.

There seems to be some attitude here that the AP is in a whole other league from the Sound Technology equipment that JC uses. Nothing could be further from the truth.

The Sound Technology gear was outstanding in its day, as is the Audio Precision. But here is the deal -- the AP is basically the same thing as the ST, except implemented with a digital *interface* so that it can be controlled via a computer rather than front panel pushbuttons.

(I'm simplifying slightly here, as the AP does offer a digital-domain measurement system that is like having another piece of test equipment inside the same box, but that is not relevant to this discussion.)

What I'm saying is that there has been *no* massive breakthrough in distortion measurement by AP since the ST units were designed. The fundamental performance between the two is virtually identical.

Now, when John Curl uses the ST by itself, there is no way in the world he can see what he sees. He has to also uses an external FFT on the residual output from the ST to look at harmonics that would otherwise be buried in the noise. The same is absolutely true of the AP. You won't be able to see what John is seeing without using an external FFT.

If the European guy with the AP is doing so, then please ignore this post.

Charles Hansen
 
Charles Hansen said:
Now, when John Curl uses the ST by itself, there is no way in the world he can see what he sees. He has to also uses an external FFT on the residual output from the ST to look at harmonics that would otherwise be buried in the noise. The same is absolutely true of the AP. You won't be able to see what John is seeing without using an external FFT.

If the European guy with the AP is doing so, then please ignore this post.

Post ignored.

The European guy with the AP is measuring the residual output of the AP's generator after the fundamental has been notched out.

With statistical averaging, he's measuring a good 20dB below where John is measuring yet so far, even when measuring the most bog standard A/V cable, none of the high order harmonics that John has been routinely measuring are evident.

So it's beginning to look as if the distortion John has been measuring is being produced by his 1700B rather than the cables as John has claimed.

However to make absolutely sure, because John and the European guy with the AP didn't measure exactly the same cables, I made up two sets of the same cables and sent one set to John (which he has already measured and reported on) and the other set to the European guy with the AP.

I'll report the results as soon as I receive them.

The European guy with the AP by the way is Bruno Putzeys, a chief engineer at Philips' Digital Systems Labs in Belgium.

se
 
Steve -

You don't have an AP. I've owned and used one for over 10 years. I know quite a bit more about its capabilities than do you.

My Dual-Domain System One cannot duplicate the method used by Curl. I seriously doubt that Putzeys' AP can either.

Please show that Putzeys' method (in at least a general way) duplicates Curl's method of sending a sine wave through the DUT, notching out the fundamental with an analog filter, making 100 FFTs of the resultant signal, and then averaging them together to reduce the noise contribution.

Charles Hansen
 
Hello SY,

Yes, you are correct.

The point that I was making was that my Dual-Domain System One won't allow for averaged digital measurements. Furthermore, it cannot do a digital measurement at the same time it does an analog measurement. Therefore it is fundamentally incapable of reproducing Curl's method without the use of an external FFT that can average many measurements.

Is Putzeys' AP different than mine in this regard? It's possible, but I doubt it. However, since Eddy is trying to use Putzeys' measurement to refute Curl's results, it is incumbent upon Eddy to show that the two results can be compared.

Charles Hansen
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.