Recreating Nelson Pass's Science or Snake Oil Tests

There is an ongoing debate in the audio community about cables. About 40 years of debate at this point. Some think they make a world of difference, others think they make no difference at all (assuming you have a well-made cable of appropriate gauge and, in some cases, shielding).

For this post, I want to focus only on speaker cables. Since, at least in my system, they are the longest cable, and they are the last component in the signal chain before the speaker. Arguably, the sonic impact of cables higher up in the chain can have a bigger impact on sound (the sonic "signature" of the cable between source and preamp receives the most gain). But I have pick a place to test first, and speaker cables is it. 

Given that I'm not the first person, and certainly won't be the last, to test speaker cables. I thought I would start by recreating the "classic" testing done by Nelson Pass in 1980. 

https://www.passlabs.com/press/speaker-cables-science-or-snake-oil

To do so, there are a few things ill need:

1. A good voltmeter (Fluke 179)

2. A signal generator (function and sweep tone) (https://www.amazon.com/gp/product/B01HM70CMY)

3. An Oscilloscope (https://www.amazon.com/gp/product/B0195ZIURK)

4. Deoxit

5. A variety of speaker cables

6. A variety of speaker and amplifier pairs

7. People for the listening tests

8. A selection of music everyone can agree on that "flexes" the frequency range
9. An LCR meter (https://www.amazon.com/gp/product/B00S298KJO/ref=oh_aui_detailpage_o05_s00?ie=UTF8&psc=)
10. A tape measure (we used a bosch laser measuring device)

Measuring resistance and capacitance is easy using just the voltmeter. Measuring inductance takes either an LCR meter or a bit more effort with a function generator, an oscilloscope, and a bit of math. Rather than go through all the details, here is a quick tutorial:

https://www.wikihow.com/Measure-Inductance

It is worth noting that inductance and capacitance can vary significantly at different frequencies. For example, a 10uF capacitor looks an aweful lot like a high value resistor at very low frequencies (read DC).

Once you have Resistance, Capacitance, and Inductance for each speaker wire. And you know the impedance curve of the speaker, and the output impedance of the amplifier... and you choose a desired damping factor... It's just a matter of a bit of math to calculate the maximum speaker wire length. 

Here is a quick tutorial on how to calculate using a pure resistive model:

http://soundsphere.mseaudio.com/wp-content/uploads/Soundsphere-WireSizeSelection.pdf?x14417

Hmm... Now that I'm thinking about it. It would be really nice if there was a web-based (or app-based) tool that let you just select your amp, speaker, and wire length... and it calculated what speaker wires would be the best electrical matches to your system... Another project for another time. 

Also, it's worth calling out that I'm using Resistance and Impedance interchangeably here. But if we want to be correct about it, Resistance is for DC and Impedance is for AC... Music is AC. 

As Nelson Pass pointed out in his 1980 article. Having the parts around for a "damper network" wouldn't be a terrible idea. Either a couple 0.047uF capacitors and 6Ohm resistors, or 0.1uf and 5Ohms should do the trick... Its worth noting that a few cable vendors (such as Transparent, MIT, and DNM) include (or sell separately) "networks" with their cables. Perhaps we do a bit of testing with the network in, and network out of the circuit as well. 

I just happen to have a bag of 0.047uF 5% 250VDC\160VAC Vishay BC Metalized Polypropaline Film capacitors, purchased for another project, sitting next to me right now... Chances are pretty good I also have 6Ohm resistors too (just in case I threw some into my Digikey cart). 

By the way, this article helped me better understand how to think about speaker cables as part of a speaker system:

https://pdfs.semanticscholar.org/e0ad/c4ec8145619f207fa1e23fab41725877de64.pdf

Basically, just model the speaker, its internal wiring, crossover, and speaker cable as a circuit directly connected to the output of your amplifier... Because it is... In most normal (read DIY) circuits i get to just ignore wire because it's of negligible length. But in the case of speaker wire (in all of its fancy permutations), it is clearly possible for the wire to affect the functioning of the overall circuit. 

In 1980, speaker wire came in 2 basic configurations. Twin Lead (Like Lamp Cord), and Coaxial (Like Cable TV Wire). 37 years later we now have a seemingly infinite number of configurations. Some of the types are:

Ribbon

Twin Lead

Quad Lead

Coaxial

Solid Core

Multi Strand

Litz (Multi strand with each strand in its own dielectric shielding)

Shielded

Unshielded 

Braided or Woven

Parallel

Jacketed

Unjacketed  

Of course, every manufacturer claims their cable is the best and uses superlatives like there is no tomorrow.  And they cant all be right... Can they?

For this test, I thought we would try to include a representative sample of the primary cable types. It would also be nice include cables of a variety of gauges, and a variety of price points... But I don't think that will be possible in any reasonable amount of time... So we will simply test 10 different cables of whatever type people will loan us. 

A few that i would like to include in the test if possible are:

1. Mogami 3104

2. Neotech NES-3004 MKII

3. Belden 5000 (AKA BlueJeans Ten White)

4. Canare 4s11

5. Audio Art

6. DNM

7. Morrow Audio

8. Oyaide

9. Pear Cable

10. Transparent Audio Cable

11. Monster Cable

12. Nordost Cable

13. Kimber Cable 8TC

14. Cardas

15. AlphaCore Goerts

16. Audience Ohno

17. Anticable

18. Audioquest Rocket 33

19. Mapleshade Double Helix

20. Cat6a Network Cable

For the listening portion of the test it would be nice to have a few different systems, ideally something like the following 4:

1. Full range box speaker with 2nd or 3rd order crossover, Class A/B Solid State Amps

2. Rull range planer speaker with 2nd or 3rd order crossover, Class A/B Solid StateAmps

3. Full range box speaker with 1st or 2nd order crossover and Class A Solid-State Amps

4. Full range box speaker with 1st or 2nd order crossover and Tube Amps

Anyone have some Zu or Devore high efficency speakers we can use? 

I think it's important to use full range speakers for all the tests, so we have a chance to hear if the cable is coloring the extremes of the audible spectrum. Its probably also important to have some people in the room that are young enough to still be able to hear the extremes of the audible spectrum. 

I have reached out to a few people/groups/companies for help and advice:

1. Nelson Pass (http://www.passdiy.com)

2. The San Francisco Audiophile Society (http://sanfranciscoaudiophilesociety.com/)

3. The Cable Company (https://www.thecableco.com/)

So far everyone has been helpful and expressed at least marginal interest in the project. Its greatly appreciated.

Continued...

Yesterday the DE-5000 LCR meter showed up, and this morning Zach and I started measuring. First the length of each cable. And then Resistance (R), Capacitance (C), and Inductance (L).

The meter's manual came in Chinese. Here is the English version:
http://akizukidenshi.com/download/ds/deree/DE-5000_manu_en2p.pdf

The first thing we realized is that it's important to test and log each value at different frequencies. We are using 100Hz, 120Hz, 1kHz, and 10kHz.

The next thing we ran into is that inductance (L) changes based on coiling of the cable. Laying out each cable in exactly the same way is important.

Also, when testing inductance its worth measuring both the L of a single run and the loopback... As illustrated by the DNM cable, things can look very different in a circuit (loopback).

Here is the google sheet with the data:
https://docs.google.com/spreadsheets/d/1oNbeYMT6nqo-ugWzvi0ZUUUi8GV7ZXNmfMbZMYC5fuQ/edit?usp=sharing

This session was a good start, and now we know what we are doing :). Next up is to test a bunch more cables at different price points.

I ran across some monster z2r speaker cable and some smaller gauge monster cable when cleaning out the garage, too late for the initial round of testing. will add them to the spreadsheet shortly.

We only spent a little bit of time on the subjective listening tests, though we did spend an entire evening on Thanksgiving listening to the DNM and Bluejeans. At this point, my take on the listening tests is that with the exception of the DNM (which has some issues due to insufficient gauge) is that I cant reliably hear any difference between the cables under test.

https://shawnkernes.blogspot.com/2017/11/speaker-cable-experimentation.html

Oh, and it's worth noting that the issues with the DNM cable completely vanish if you biwire (use all 4 conductors for one speaker).

Some more good information on data cable design here:
https://www.quabbin.com/tech-briefs/why-cable-capacitance-important-electronic-applications

Test Systems:

System 1:
Speakers: Usher BE-20
Amplifier(s): Parasound Halo JC-1
Preamplifier: Parasound Halo JC-2BP
Sources: Nakamichi Dragon, VPI Scout w/Dynavector 20xh2 and modified PS Audio GCPH, PS Audio Directream DAC with Tidal Masters MQA
Interconnects: Mogami Gold, Kimber Hero, Monster Cable Interlink 400 MKII, Transparent 1900XL

System 2:
Same as "System 1" but replaced Amplifiers and Preamplifier with Parasound Halo Integrated. It's worth noting that while the soundstage moves farther away - this "little" integrated hangs in there with its bigger siblings remarkably well.

System 3:
Speakers: Magnepan 3.6R
Amplifier(s): Bryston 4B-SST
Preamplifier: DIY First Watt B-1
Sources: Technics SL-1210M5G w/Dynavector 20xh soundsmith modified and PS Audio GCPH, Logitech Transporter with Musical Fidelity VLink192 and Tidal Masters MQA
Interconnects: Mogami Gold, Kimber Hero, Monster Cable Interlink 400 MKII, Transparent 1900XL