[audioengr]

I finally found some time to make some USB and S/PDIF cable measurements.  These are using my 6GHz Scope with jitter software.  The scope input is terminated with 75 ohms.  No probe used.  The histograms show the variation of cycle-time over thousands of samples.  These were all made using the Off-Ramp 5 driven from PC laptop running XP and Foobar.  Not the best, but okay for comparisons.  The test file was music in .wav format at 96kHz.

The first set are S/PDIF cables.

Starting with the 1.5m Bitmeister S/PDIF cable ($650) that I used to manufacture.  75 ohm BNC to RCA adapters at both ends:



Next, my new BNC-BNC 4 foot cable that sells for $250:



Quite a bit better than the older more expensive cable.

Now for some USB cable measurements.  Note the change in axis time.  41 psec per division.  This is automatically selected by the software.

First a 5m Belkin Gold:



Next a 5m Polestar:



This has less "lumpiness" and "bulges" in the base of the distribution.  The spires are less separated.

It should have a smooth ramp up and down and very steep spires.  A single spire would be best, but not critical.  This indicates some assymmetry in the rise and fall times.

Steve N.

[brj]

Very interesting, Steve - thanks for sharing it!

Do the USB cables show jitter differences with respect to length similar to S/PDIF cables?

[audioengr]

Very interesting, Steve - thanks for sharing it!

Do the USB cables show jitter differences with respect to length similar to S/PDIF cables?

I tried a short USB cable (5 feet) and it seemed to be very close to the same length Polestar in performance, so I didn't post it.   I actually have some 2 foot cables, so I'll try this soon.

The problem with short cables with my setup is that there are virtually no reflections from the scope.  With real audio electronics there are always reflections, so the too short cable (0.5m) will be bad. Also, my drivers have risetimes in the 3 nsec range so even 0.5m may not be a problem.  Other vendors drivers will be slower, so better to use the longer cable.

Steve N.

[Ern Dog]

 From what I can tell by looking at the graphs, the new BNC cable offers the least amount of jitter and the Bitmeister has the most jitter.  It is more difficult to tell between the Polestar and the Belkin, but it looks like the Belkin offers lower jitter than the Polestar.  Am I reading this right?

[acg]

Steve, are you likely to test one of the RAL USB cables that you sell?  That would be interesting.

Anthony

[audioengr]

From what I can tell by looking at the graphs, the new BNC cable offers the least amount of jitter and the Bitmeister has the most jitter.  It is more difficult to tell between the Polestar and the Belkin, but it looks like the Belkin offers lower jitter than the Polestar.  Am I reading this right?

The max jitter with both USB cables is 25sec.  The shape of the curve is better with the Polestar.

Steve N.

[audioengr]

Steve, are you likely to test one of the RAL USB cables that you sell?  That would be interesting.

Anthony

It would be .  I dont currently have one or I would test it

Steve N.

[DaveC113]

Thanks for posting this test, it is very interesting.

Do you feel the Polestar USB cable is sonically superior to the Belkin Gold in listening tests?

[audioengr]

Thanks for posting this test, it is very interesting.

Do you feel the Polestar USB cable is sonically superior to the Belkin Gold in listening tests?

Absolutely.

Steve N.

[audioengr]

Here are some jitter plots of OR5 with OTL mod with and without the Short-Block.  The effect of the SB will be minimal with the scope and the small ground-loop with the OR5.

Without Short-Block:




With Short-Block:




Slightly better distribution shape.  Both with worst-case P-P jitter of 80psec.  Both with the majority of the jitter in the 15-18psec range.

This is a case where the system measurement without the scope would better show the difference.

Steve N.

[tpaxadpom]

Looks like it measures better without the short block. Mean is the same with the STD Dev being smaller without the short block.

[audioengr]

Looks like it measures better without the short block. Mean is the same with the STD Dev being smaller without the short block.

Probably not better.  Its within the measurement variation.  Just moving a cable a bit can make this much difference, or just clearing it and making a second measurement.

I would say they are virtually identical, proving that the common-mode noise between my PC laptop and the OR5 is extremely low in my bench setup.  My system setup is probably a lot different.

Steve N.