You ever get that $500?
-
WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!
Speaker wire, CCA vs OFC
- Thread starter EDMLover
- Start date
You ever get that $500?
@Speedskater Thanks for pointing out Jim Brown's work. His website has a lot of very good information on the subject of RFI. http://audiosystemsgroup.com/publish.htm
I quickly looked through a few of his papers/articles, and a relevant one to our discussion here is http://audiosystemsgroup.com/AES-RFI-SF08.pdf.
On page 65, Jim did say always use twisted pair cables for loudspeakers. He didn't give a detailed explanation on exactly how RFI picked up by speaker cables would lead to problems. My interpretation is that somehow the RFI at the amp output can travel back up the signal chain and causes problems upstream, or the RFI can mess up the feedback mechanism of the amp and causes it to misbehave. I don't believe the tiny amount of RF power picked up by speaker cables and the high frequency can possibly drive any audible output at the speaker.
Earlier in the article (page 43), he showed a test using a microphone, a mixer, and an AC generator operating within a stone's throw distance from an AM radio station antenna. These signal level and source/load impedances are vastly different from those of a loudspeaker/power amp circuit, and I don't think one can credibly extrapolate the results from mics to loudspeakers.
I am not an electrical engineer (I am a mechanical engineer) and certainly not an amplifier designer. However, knowing a little bit about how RF noisy inside a class D amp already is, I will be very surprised that this is a widespread problem. If this is indeed the case, given how universal it should have been as Jim Brown described it, I am extremely surprised that this problem is so little known, at least in the consumer side.
My cynical self thinks that as Jim Brown's specialty is in solving RFI problems, he would tend to see RFI everywhere. As the saying goes, "if all you have is a hammer, everything you see looks like a nail." I am not sure how real the problem is. Anyway, he was not asking people to buy super expensive speaker cables, and certainly twisted pairs won't hurt, there is no harm in it.
This is probably as much as I can contribute on this subject. May be someone else can devise some tests to verify.
Last edited:
IME, which does not include many class D amplifiers, RFI on the speaker leads caused problems when it snuck into the feedback (mainly) or bias path to be mixed (modulated) back at the input and amplified at the speaker leads. Worst-case it would cause the amp to go unstable. For decades now any decent amp includes filtering to prevent that from happening.
Since a class-D amplifier is pretty much internally already a HF noise source, it must be designed to handle it and it's hard to see them being terribly susceptible to RFI.
Of course a bad amp is a bad amp no matter it's class. (Insert no-class, low-class jokes here.)
FWIWFM - Don
Since a class-D amplifier is pretty much internally already a HF noise source, it must be designed to handle it and it's hard to see them being terribly susceptible to RFI.
Of course a bad amp is a bad amp no matter it's class. (Insert no-class, low-class jokes here.)
FWIWFM - Don
Twisted pair speaker cable also reduces the self inductance of the cable. That is probably a better reason to use them. I use some certified installer in wall cable that is 12 gauge twisted pair. I covered it up in a pretty woven sheath and put banana plugs on the end. They look like audiophile approved cables and cost me about $40. This way I get the full placebo effect without paying much for it. 
Speedskater
Major Contributor
The first seven pages of this Jim Brown paper, explain a lot of things about interference. The remainder of the paper is about ferrite chokes.
"RFI, Ferrites, and Common Mode Chokes For Hams"
http://www.audiosystemsgroup.com/RFI-Ham.pdf
is this a poor choice for internal wiring between a DAC and amp? its what I have on hand https://www.amazon.com/Shaxon-SO18-...per+wire&qid=1566564194&s=gateway&sr=8-2&th=1
It will work. If the outputs and inputs are unbalanced I would run coax, if balanced twisted pair just to cut down on potential noise.
Inside an enclosure any normal electrical wire will do just fine.
Speedskater
Major Contributor
An auto parts department will have several different usable wire kits.
So a headphone amp, external power supply (wall wart)... 18 AWG is pretty large for that but you'd probably be OK twisting it and using it for the interconnect.
That depends on what else is inside that enclosure. Once upon a time, I had a computer that was so electrically noisy that parallel ATA drives wouldn't work at the higher speeds, not even when using a cable with extra shielding. SATA was just becoming a thing, so I got a controller and adapters for the drives. The differential signalling had no problem working at full speed.
Sure, if you are working with high speed digital circuits it is a different discussion.
Or a switching power supply or power amp, which is what I was wondering about when the question was first asked.
My point was that high speed digital circuits need proper differential connections. Audio doesn't at reasonable distances (but they do help avoid ground loops).
Sorry, I guess I'm missing the context of your point. Most of the on-chip CPU and support logic is single-ended and running from GHz clocks. Some interfaces, like PCIe and SAS/SATA, as well as Ethernet, use differential serial lines, but memory data is single-ended and a huge noise source. Or are you saying that audio signals in the presence of digital circuits should use differential circuits? Coax shields work well enough though I agree with differential where you can.
The OP asked about using standard 18 AWG wires for interconnects from DAC to amp inside a single box. I thought he meant a speaker amplifier and not headphone so had higher power in mind and was not sure if it was a class D amplifier. There is digital logic in the DAC and such, and maybe in the headphone amp, so even with the power supply external (not sure if SMPS or not) it made sense to me to twist the wires to provide some noise rejection.
I think we're all on the same page, just looking at it from different angles. Sorry to be obtuse.
Last edited:
I guess so. My point is that differential becomes important for high speed digital connections over distances longer than a few inches (so, as you state, PCIe, SATA, Ethernet and USB), as well as low-level analog signals for long distances in noisy environments, but doesn't really matter at short distances. Of course twisting the wires is never a bad idea, but only really works if the connection is differential.
Hmmm... IME twisting wires does provide some shielding effect, frequency range and such depends on the twist ratio. When we're routinely looking at -100 dB kind of numbers even a short distance can lead to coupling that degrades SINAD. Whether it matters is very much dependent upon the application and environment.
Twisting wired, as I wrote, really only works if the connection is balanced/differential. Twisting doesn't hurt, but as long as you don't run a low-level signal (or especially sensitive signal, such as a feedback path) right next to a noise source, it isn't much of an issue at the distances involved in connections inside an enclosure.
