Time to revisit this one seeing as how I didn't get much further at the time, a new baby will do that to you!
@lucid, thanks again for the time taken to post such useful information. I'm not a cable connoisseur by any means but what you say is understandable and makes sense. From what I can gather, 75ohm co-ax is fine for the majority of video and audio applications unless over long distances.
Coax is highly suitable for analogue and digital (electrical) interconnects. Distance isn't really a problem either unless the cable is very lossy. Think about the aerial and satellite cable running around your house. Did anyone say "Oh, you can't have your telly/Sky box any more than X metres from the aerial/dish" ? No, probably not. The signal from an LNB is runs at anything up to 2,150MHz which is massively higher frequency than analogue audio at a few tens of kHz. Very high frequencies are where cable losses are seen most acutely. It also has a bit less amplitude @ +/-0.3V for satellite (0.6V pp) versus anything up to 2.0V pp from some audio sources.
So ask yourself, why is it okay to run 20-30-40 metres of cable carrying a 2150MHz signal of lower amplitude through satellite coax at £0.75/m, but the audio industry wants you to believe it's a big problem running more than 3 or 4 mtrs of coax carrying a 20kHz signal at 1Vpp through cable coasting ten times the price and more? Could it be that they don't have a good engineering reason other than some of their cables aren't that good?
To be fair, there was (an possibly still is for some brands) a bit of an issue with impedance matching within the interconnect design.
I use 75 Ohm cable and 75 Ohm rated plugs to avoid signal reflection points within a lead. It's just good engineering practise. What I can't control though is the source and load impedances. The load impedance shouldn't really be an issue. It should be very high; 50,000-100,000 Ohms, something in that range. It's the source impedance that can be hugely variable. A Creek CD50 CD player has an output impedance of 47 Ohms. An Onkyo C-7000R CD player is 330 Ohms, and an Astin Trew AT3000 CD player is just 1.5 Ohms (Yes, one point five. It wasn't a typo.) There's no way to practically and cost effectively engineer cables to cope with such a wide variance. Thankfully any audio reflections from the source/cable interface are dealt with inside the player. The best we can do is to ensure the minimum number of transmission boundaries within the lead.
Phono plugs weren't originally designed as 75 Ohm connectors. Solder joint ones definitely aren't. But thanks to companies such Canare and Belden (F-Conn) responding to the use of the phono socket as a video connection for analogue HD (Component video) it's now possible to buy plugs that are close to or exactly that standard.
With the correct cable I can run analogue audio 30-40 mtrs with imperceptible loss (<0.3dB @ 200Hz) which I do with my micro-diameter sub-woofer installation cable.
I'd be looking to make a mixture of stereo analogue audio, digital co-ax and some short bridging cables to use for my NAD and Schiit stack. I've seen reflection and refraction mentioned when talking about really short cables, is this actually much of an issue and is there a minimum recommended cable length? Max length of the interconnects will probably be about 2m.
This is to do with power transfer. When the signal encounters a high impedance load (as they all should be) such as a 47,000 Ohm impedance for the Line In of an amplifier there'll be a signal reflection. There's a relationship between the way the wave travels (propagates) down the cable, its frequency (and the harmonics) and the cable length that determines the size of the resultant voltage that the destination device sees. This is because the signal is represented by a simple sine wave, and a sine waves varies in amplitude from +100% through 0% to -100%. The power transmission level is affected by what happens when new signal waves encounter the reflected waves.
Since a high impedance load looks like an open circuit in electrical terms, then the reflected wave will have close to 100% of the amplitude of the source wave. At various points where the waves meet there will be constructive or destructive interference. IOW, the signal as seen in some parts of the cable could appear to double or be summed to zero. This is a standing wave and it sounds like a complete disaster for audio and video interconnects. A 20Hz electrical signal has a wavelength of roughly 15,000km. (Speed of light @ 300,000,000m/s in a vacuum divided by the frequency in Hz = wavelength in metres. Divide by 1000 to get km). A half-metre long interconnect really shouldn't cause too many issues. In fact QED are quite happy to make up and sell interconnects in 0.6m lengths. Do you think they'd do that if there was a risk of getting no sound out of the cable?
Looking at amps with pre-power bridging links, the little metal U's have about 2cm of exposed conductor. Once again, did any of these manufacturers suffer massive returns because the amps didn't work?
The current flavour of the month from some cable companies such as Chord is for standard interconnect lengths starting at 1.2 metres. That's double the 0.6m minimum from QED. If you're concerned then I'd say stick with multiples of 0.6m for anything between separate boxes.
I've been comparing the Belden 1694A, 1505F and Canare LV-61S and they all seem much of a muchness spec wise, apart from the 1694A being RG6. Should I be using RG59? I guess all quality 75ohm cable will have similar specifications so it just comes down to price/preference?
RG refers to Radio Guide, and it's a standard physical size guide, so it defines overall diameter, dielectric diameter and centre core size. There's a relationship between these dimensions that defines the cable impedance.
What the Radio Guide doesn't specify is materials. There's a lot of cheap crappy RG6 aerial and satellite cable that's made from a solid steel core anodised with copper, and the outer shield composed of sparse aluminium braid over a metalised plastic (Mylar) foil. It's RG6, but nothing like the quality of Belden 1694A, or even Webro WF100 satellite/TV coax.
RG59 is still a 75 Ohm cable but a smaller outside diameter (aprx 6mm vs 7mm). Again, the Radio Guide doesn't specify materials. 1505F uses a multi-core centre and is a double shielded with braid rather than a foil/braid combo. This makes it more flexible than foil/braid cables but not as suitable for very high frequency signals such as for use as satellite coax. It's good for audio though.
My Go-To RG59 is
Liberty Interflex SD. It's a Serial Digital cable with the electrical properties of 1694A (high bandwidth, low loss) but is almost as flexible as 1505F so it a good all-round solution for audio and for digital coax. I also use
Van Damme Plasma Grade RG6 if I need something very flexible for audio. In the past I tried Belden 1694A but found the solid core and its thickness combined with the foil/braid shielding just made the whole thing far too stiff. The Liberty does everything that 1694A does but in a much more install friendly way.
You mentioned the Canare RCAP and I agree, they are pretty chunky. I've been looking at some compression fittings (namely ICM FS59/FS6). They look easier to work with than the RCAPs.
Yes. I use the F-Conn connectors and also some from Liberty. They're all short-body 75 Ohm plugs and work well. The Canares are just too damned long IMO, especially for free standing gear on a standard living room Hi-Fi rack.
Have you factored in the cost of a decent compression tool? I've tried the cheaper alternatives but found that they don't quite fit or don't close the plugs up completely so I stick with my Liberty compression tool. They're a lot of money, but buggering up the plugs and then having to remake the cables gets expensive and time consuming.
