Speaker cables...!

[oscroft]

I'm very much a speaker cable sceptic - beyond a certain minimum thickness and contruction quality, I've never heard any difference.

But at the same time, I try not to adopt the fingers-in-the-ears "It can't, it can't, it can't!" approach when people make claims that I doubt or without an explanation that I can understand. It's come to mind recently after a difference of opinion on speaker cables popped up across some YouTube channels.

I've always thought about speaker cables in terms of LCR (Inductance, Capacitance, Resistance), and those are well understood and easily measurable. And with a decent thickness of copper cable they tend to be vanishingly small. But some expensive speaker cables are made in a twisted or braided form, which raises capacitance, and I've never understood why they do that.

There's also been some debate about the skin effect, where higher frequencies are carried more on the surface of a conductor and don't reach much of a depth into it. That's a genuine problem at radio frequencies in the megahertz range and the like. Some argue that the effect does exist to some extent down at high audio frequencies, but the consensus among audio electronics experts seems to be that it's insignificant.

Then there's the idea of noise picked up from electromagnetic interference (EMI) and radio frequency interference (RFI), and that's what the recent YT disagreements have been all about. It's certainly true (and easily demonstrable) that speaker cables can easily pick up RFI - if I connect a length of speaker cable to my FM tuner, it makes an effective antenna. Whether RFI can have any effect on the audible spectrum is the big question. Is there any possibility of demodulation anywhere in the chain? After all, Lucille Ball reckoned she could sometimes hear radio through a tooth filling!

Surprisingly, while I can find plenty of material covering LCR and skin effect, I haven't found much of any detail on the possible effects of EMI and RFI - other than a sort of hand-wavy "You can try using shielded cable if you think you have a problem" thing.

So I'm going to try a test. A number of people have championed the use of Cat5 (or Cat6) ethernet cable in various configurations. Its twisted pairs are good at noise suppression, which is important at the high frequencies used for ethernet. I've used it myself in the past, but only bacause I had some and with no thought about noise suppression - I just soldered all eight connectors together, and used two cables for my speakers. It's inconveniently stiff and quite ugly, though, so I use QED 79 strand these days.

But I do have some Cat5 cable and some banana plugs, so I can try a twisted-pair noise suppression version at zero cost. And I simply have to try it, don't I? I know I do!

I'm going to make up a pair of cables along the lines of The "Full Frequency Range Cable". (That, and the linked pages, make some interesting claims about speaker cables - though I'm certainly not endorsing them). I'll measure the LCR compared to my QED cable, and the Cat5 will surely have higher capacitance. There's an interesting effect of high capacitance cables on some amps shown at Speaker Wire, so I'll use my scope and see if I can detect any ringing/oscillation.

Anyway, I need breakfast now. Next stage is to make up the cables, and I'll post a photo when they're done.

 

[mushii]

Nice to see somebody doing some real science and not just technobabble and pseudoscience. Nice work

 

[MikeFaulkner]

In my younger days (long time ago) I used to play a lot of 5 a side football. One of the places we played in had a metal roof and we could quite clearly hear a local radio station playing. There was no radio in the building! Moving on several years, I developed an interest in Ham Radio and read about WW2 POW's building their own SW radio's using a point contact diode. Somehow the metal roof was acting as a radio receiver.

So (and you are probably wondering what this has got to do with RFI and EMI) just about anything will pick up radio signals, it's whether the electronic equipment at the end of the wire will do anything with them!

Will be interesting to hear the results of your experiment.

 

[oscroft]

It's hard work twisting wires, and my hands ache now Anyway, the steps...

1) Braid three lengths together, and tape the ends to stop them unravelling...

2) Strip the outer insulation, then separate the solid colour and stripey wires and twist them together...

3) Strip the insulation from the ends of the conductors, twist together, and fasten on some banana plugs...

4) Finally, Just tighten up the twists a bit and add a bit more tape to neaten it up a little...

More in next post...

 

[oscroft]

My usual cables are about two metres long. So I used six three-metre lengths of Cat5, expecting to lose some length due to the braiding. They ended up about 2.5 metres, which is fine. As an aside, when I read about people testing cables, they tend to use five metres or longer - that's great for those with AV systems and long cables to reach all the speakers, but any possible cable effects are obviously going to be less for two-channel systems with short cables.

The individual conductors are 24 AWG, so 12 of those on each side gives me an effective 13 AWG approx (according to a calculator I found somewhere). It's similar to the 14 AWG of my QED 79 strand cables, so I think that's a good amount of copper.

I might tape up (or heat shrink wrap) the end pieces with red and black, but I haven't decided yet - I do quite like the look of the exposed colours.

Anyway, that's it for now. I haven't measured LCR yet, I've just done a quick check that I haven't shorted anything. I'll do that tomorrow, and then have my first listen. Or maybe I should burn them in for a few weeks first

 

[hamzamian]

Interesting experiment and I look forward to seeing the results.

are you using solid core? What kind of cable is it? Shielded? Foiled? Neither I guess if it’s plain old CAT5?

 

[Paul7777x]

Interesting indeed. But it still must be done double blind with a large number of people and a decent number of attempts to have any weight.

 

[oscroft]

Interesting experiment and I look forward to seeing the results.

are you using solid core? What kind of cable is it? Shielded? Foiled? Neither I guess if it’s plain old CAT5?

Plain old Cat5, each core solid 24 AWG, no shield.

 

[oscroft]

Interesting indeed. But it still must be done double blind with a large number of people and a decent number of attempts to have any weight.

It will have to remain weightless, as there's only me and double blind is impossible.

But did you make all of your buying decisions using only double blind testing with large listening panels? And did you refuse to consider the possibility of differences between components if you couldn't perform such a test?

In the real world, simple ad-hoc comparison is the only kind any of us can really use - as double blind tests with large listening panels are simply not attainable for ordinary listeners. Either that, or just buy some kit and never upgrade.

If I think I detect any differences, it might help others in their decisions over whether to consider the possibility of cable differences. And that's all I can offer, plus maybe a bit of fun.

 

[Paul7777x]

It will have to remain weightless, as there's only me and double blind is impossible.

But did you make all of your buying decisions using only double blind testing with large listening panels? And did you refuse to consider the possibility of differences between components if you couldn't perform such a test?

In the real world, simple ad-hoc comparison is the only kind any of us can really use - as double blind tests with large listening panels are simply not attainable for ordinary listeners. Either that, or just buy some kit and never upgrade.

If I think I detect any differences, it might help others in their decisions over whether to consider the possibility of cable differences. And that's all I can offer, plus maybe a bit of fun.

Usually, I’ve bought blind. No other choice as most of the stuff I’ve bought over the years has been used.

I did do a boy of reading before buying just to get the gist of what I might be paying for.

All except for cables. I’ve always bought 2.5mm inexpensive copper speaker cable. And until someone somewhere shows me an engineers reasons and reproducible data why there’d be a difference between any two given speaker cables, I’ll continue to buy that when I need speaker cable.

When I’ve needed XLR cables I’ve bought professional ones at around 20 quid a pair.

Again, engineering reasons and data might make me buy £100 stuff.

But anecdote never will.

 

[oscroft]

Usually, I’ve bought blind. No other choice as most of the stuff I’ve bought over the years has been used.

I did do a boy of reading before buying just to get the gist of what I might be paying for.

All except for cables. I’ve always bought 2.5mm inexpensive copper speaker cable. And until someone somewhere shows me an engineers reasons and reproducible data why there’d be a difference between any two given speaker cables, I’ll continue to buy that when I need speaker cable.

When I’ve needed XLR cables I’ve bought professional ones at around 20 quid a pair.

Again, engineering reasons and data might make me buy £100 stuff.

But anecdote never will.

Sure, perfectly reasonable. I would never suggest that anyone should ever buy anything on anecdote - but anecdote has led me to consider things that I can then evaluate by my own standards.

Again, engineering reasons and data might make me buy £100 stuff.

Actually, a small additional comment in response to that. Engineering reasons and data are not alone sufficient to make me buy something, just to consider buying it. I'd still, ideally, want to listen too.

 

[oscroft]

I've done some LCR measurements on the cables (Cat5 and QED), but before that I did a bit of listening. No proper comparisons, I just used the Cat5 cable for a couple of hours of evening background listening. And nothing struck me as "Hey, this is better/worse/different".

Resistance: Both are too low to measure accurately with a simple multimeter. They both hover around 0.2 to 0.3 ohms - but just shorting the meter leads shows 0.1 to 0.2 ohms anyway, and that's the absolute bottom of the meter's range. So, resistance is negligible in both cables.

Inductance: QED inductance came in at 5.3uH (microhenries), Cat5 at 5.8uH. That's at the bottom end of what my meter can measure, so probably not the greatest of accuracy, but at least it shows there's no significant difference

Capacitance: QED cable capacitance (one two-meter run of straight cable) measured at 108pF (picofarads). The Cat5 cable (three three-meter runs of twisted cable, braided) measured 1,560pF. So that's massively more capacitance in the Cat5 cable. But it's still tiny compared to the capacitors used in the speaker crossovers.

 

[oscroft]

Walking to the shops this morning, I passed a skip of discarded building trash, and I spotted on the top of it... a spool with about 20 to 30 metres of Cat6! And... and... I managed to resist the temptation, so I'm not going to waste today making up any new speaker cables

 

[oscroft]

OK folks, I've been doing some careful listening...

The frustrating thing is that I can't do quick A/B tests (and even more frustrating, I can't do them blind). So I keep thinking I can hear subtle differences between the two cables, but not clear enough to be confident. I can't do a quick comparison to make sure, and the suspected differences are too small to survive in my memory long enough for a cable swap.

So two preliminary conclusions... One is that if a suspected difference is so slight that I'd need a rapid A/B test to be sure if it's there, it's not a difference worth considering. Secondly, when I suspect I hear small differences but can't A/B confirm them, my only safe conclusion is that they don't actually exist. (And I think that's probably two ways of expressing the same conclusion).

I think I'll continue swapping the cables, but maybe only every couple of days, to see if I get any longer-term feeling (like listening fatigue, for example), though that's hampered by the fact that my ears and brain vary from day to day anyway.

To come back to a comment from The Naked Truth about Speaker-Cables, that a Cat5-based cable has "a performance more akin to serious High-End cable" - if that's the case, then as far as I can tell, so does QED 79-strand.

And the claim at The "Full Frequency Range Cable" that "the usual Figure 8 Cable with fairly large, multi-stranded Conductors is sonic poison. If you use this stuff, even switching over to simple 1.5 mm solid core mains cable will bring substantial improvements" is obvious cackpipe.

Still, my new cable does have one advantage over the QED 79 strand - it's actually a bit cheaper, judging by current Cat5 prices.

 

[oscroft]

I've just thought today (today!)... wouldn't it have been fun if this had all been leading up to an April 1 revelation of the massive benefits of twisted-pair noise-rejection speaker cable?

I always have my best ideas too late...