View Full Version : Question for anyone doing DIY RCA interconnects
I have some Belden twin core shielded wire and just purchsed some Furutech FP-126R RCA plugs.
I just want to check in regards to wiring.
For each single RCA plug, do i use one core for signal, one core for ground and then also connect the shielding to ground at one end of the RCA cable (to connect to the source end)?
Have also seen people run the shielding straight to the chassis earth on an amp to prevent it being near the ground of the RCA lead???
Any advice would be great as all the stuff i google uses single core wire
That is one common way of doing it. You can also use both cores as the signal and the shield as the return (connected at both ends obviously).
Cool, thanks.
I only bought 4 RCA plugs for now to try. But when i get more i will try both ways and see what the difference is. I have plenty of cable to play around with.
Cheers
The Barbarian
18-03-2014, 18:22
Done all mine loike dis
Done all ine loike dis
That's it. I'll try that way for the first set and alternatives for the next few to see how they compare.
Thanks
walpurgis
18-03-2014, 22:50
You can leave the screen unconnected if you choose.
Except why would you want to Geoff?
I'd also not use both cores for signal in a twin core cable with shield. You want the ground return to be lower impedance than the signal ideally for better S/N, so I'd recommend if using twin core, you use one core for signal only, and you can either connect the shield one end or both ends for twin core.
For coax, you have to ground the shield both ends as it also acts as signal return. Leaving it disconnected will result in some rather unpleasant noises coming from your speakers!
If you connect the shield at both ends it will effectively become part of the return. Also, bearing in mind the construction of a braided shield there will be more metal in the shield than in the internal cores and so will offer a lower impedance.
Yeah i'll go with one core for signal, one for return and the shield connected at the source end.
Will try one set with shield connected to the return core at one end and try another set with the shield connected to the chassis earth screw of my amp to see if there is any noticeable difference.
Thanks
Yes, that will be interesting to try - let us know your findings please.
The idea behind the so-called semi-balanced construction is the twisted pair as an interconnect. In this, one of the pair is used as the signal and the second is used as the return. If the pair is tightly twisted the return lead, being at the system's ground potential, will act as a screen/shielding to the signal lead. One advantage of using a twisted pair is that any interference external but common to the pair will induce currents in both leads, so they will tend to cancel, that is the cable will have a high CMIR (common-mode interference rejection). Some interconnects on the market are just that - a twisted pair without an additional screen.
Most however do have an addition screen or shield, acting as an electrostatic shield. This may be a wire braid or it may be an aluminised foil. In either case the screen is connected to the return lead at one end only, as the diagram provided by André (The Barbarian) shows. If it is connected at both ends then the interconnect is little different to a coaxial design, save the capacitance will be increased. If the screen/shield is not connected it will simply 'float' and not provide any additional shielding. All it will do is increase the cable capacitance slightly.
Many people (the manufacturers included) insist that the screen should be connected to the signal return at the source end of the cable, and some manufacturers even go to the trouble of marking the cable with directional arrows; the idea being the cable is installed so that the arrows point away from the source. This is a fallacy. The cable should be installed so that the common junction of the screen and signal return is at the point of lowest electrical potential as regards the external interference. This will be the mains earth. It is not neccessarily true that the ground of the source is at earth potential: if the source is double insulated (i.e. uses a two wire mains cable) it will not be. In these cases it is best to experiment with changing the direction of the cable and not rely on the cable markings, or to wire up the cable so that the screen shield forms no part of the signal circuit and is connected separately to the metalwork of the preamp. Semi-balanced cables, being asymmetric, are the only form of interconnects that can display directionality.
However Barry, common mode noise rejection doesn't occur in single ended circuits so twisting a pair of conductors doesn't act as a shield and will not reject common mode noise. That only happens in balanced audio circuits where summing one to the other against a common reference results in rejection. Nowhere near enough coverage for EMI or RFI rejection by twisting pairs either in single ended configuration. For any sort of effective shield, a proper braided or spiral shield is necessary and absolutely essential for digital cables for example.
Your last paragraph is though correct. Direction arrows tend to be fitted not to denote connecting the arrows pointing towards preamp so that the shield is earthed at source (for example) but so that a reference is there for the grounded end of the shield (opposite end from where the arrows are pointed). Most people though would not know where the lowest ground potential is so the preamp tends to be used as the star-earth point for grounding. Easy enough to swap leads round to see if there's any difference one way or another. Quite often, there isn't.
Any external interference will induce identical currents that will flow in the same direction in both conductors of a twisted pair and so will have no effect on the wanted signal, which of course has currents flowing in the opposite direction in each conductor. Twisted pairs are used in digital computer circuitry. Furthermore, since one of the conductors of the twisted pair is at ground potential, tightly twisting the pair will provide a measure of overall screening, albeit with a low 'optical coverage' of ~ 50%. There are some proprietary interconnects that consist solely of a twisted pair without any additional shielding.
In the '80s it was briefly fashionable to use unshielded 300 Ohm ribbon feeder for interconnects. I believe that ribbon style interconnects are still available from some manufacturers (Max Townshend's "Isolda" cables) and these have received good reviews in the audio press.
Common mode noise rejection:
Sorry Barry, but it's not factually correct to state that single ended systems can (or do) operate common mode noise rejection. Common impedance coupling takes effect in single ended systems where the shield (whether that be a braided shield or a core twisted around the signal) carries a ground voltage (common mode voltage) between two interconnected devices with NO common mode rejection at the receiving end. There is NO common mode rejection possible. Indeed, the only property of s single ended cable contributing towards reduction of noise is the screen impedance. The lower the better.
For fully balanced devices, there does exist common mode noise rejection. I have given a full description of how this operates in an article published HERE (http://referencefidelitycomponents.co.uk/design-of-interconnects/) for anyone interested.
Don't take my word for it though as there are plenty of papers available on the subject. Just a point that often confuses hence the offering of a fuller explanation.
I Have read that digital coax interconnects length can affect noise in the cable and to buy in multiples of 1.5m.
Just wondering if any of this applies to analogue RCA interconnects? I'm about to build my interconnects with Furutech ends so was wanting to find out if length is going to matter? can they be too short? I'd rather have short cables between my components on the hifi rack but if there is a problem with this i can make them longer.
walpurgis
05-04-2014, 18:55
I just make my cables of various lengths to suit my needs. Never had a problem.
I don't think you can have a too short cable, unless of course you have a special situation and are depending on it to provide some sort of capacitive, resistive or inductive effect and require a particular length to achieve this. But that would be a funny way to go about things.
The only problem with very short interconnects is the practical issue of it limiting component positioning - you won't hear any difference between. 0.5 and 1M interconnect.
I won't be moving my gear around so i will stick with shorter interconnects to keep it neat.
Just didn't want to make 0.5m or 1m interconnects and find out that they affect sound or anything.
I Have read that digital coax interconnects length can affect noise in the cable and to buy in multiples of 1.5m.
Just wondering if any of this applies to analogue RCA interconnects? I'm about to build my interconnects with Furutech ends so was wanting to find out if length is going to matter? can they be too short? I'd rather have short cables between my components on the hifi rack but if there is a problem with this i can make them longer.
I had heard this about digital cables as well and once read a pseudo-scientific explanation that made sense - something to do with reflections. I was using a 0.7m digital cable at the time and it worked fine. For analogue cables I think the rule is the shorter the better - less capacitance, but not a big deal for a meter or so. More significant is that you don't stress your wires by turning them in too narrow a radius.
I'd also not use both cores for signal in a twin core cable with shield. You want the ground return to be lower impedance than the signal ideally for better S/N, so I'd recommend if using twin core, you use one core for signal only, and you can either connect the shield one end or both ends for twin core.
Paul,
I have made an unshielded interconnect from 8 strands of 24AWG. I used 4 for signal and 4 for return. In order to have a lower impedance would I have been better to have used 3 for signal and 5 for return?
('pologies Alex for thread-drift.)
Paul,
I have made an unshielded interconnect from 8 strands of 24AWG. I used 4 for signal and 4 for return. In order to have a lower impedance would I have been better to have used 3 for signal and 5 for return?
('pologies Alex for thread-drift.)
No Problems, i'm interested to know this also
Paul,
I have made an unshielded interconnect from 8 strands of 24AWG. I used 4 for signal and 4 for return. In order to have a lower impedance would I have been better to have used 3 for signal and 5 for return?
('pologies Alex for thread-drift.)
You're better off using most for the strands for ground return tbh. Better still to use shielded cable as you wont get much shielding effect from twisted strands on single ended...it really needs to be a braid or spiral screen to be effective. If it was all I had to hand though, I'd use 1 strand of 24 AWG for the signal and 7 strands for the ground return.
The only problem with very short interconnects is the practical issue of it limiting component positioning - you won't hear any difference between. 0.5 and 1M interconnect.
Ahem... I beg to differ! ;)
The difference is subtle, but noticeable, which is why I use 0.5m interconnects whenever possible, especially in analogue applications where protecting 'delicate' music signals from vinyl is involved.
The only cables I've found that sometimes benefit from being longer are digital interconnects :)
[The worst cables to use long lengths of are speaker cables].
Marco.
Ah - imagine if you were using 0.4M interconnects ;).
Good idea - I'll hack a bit off mine later! :eyebrows:
Marco.
Seriously though, is your gear that close together - obviously it is. I'm surprised though cos, as you know, I only have one source and even then I couldn't use cables that short.
Yup, my DAC is on the shelf on my rack above my preamp, so a 0.5m cable is fine there, and my Paul Hynes MC head amp is also directly above my preamp/phonostage, so ditto in terms of cable length required for that job.
Only between my preamp and power amp do I need a 1m pair of cables, and also from my T/T to my MC head amp.
If I use a 1m (or longer) pair of interconnects in either application where I'm currently using a 0.5m pair, it softens the sound slightly by reducing dynamic impact and also makes it a little less open. The effect is easily heard in my system :)
Marco.
I must admit, I've never personally tried short cables as it would compromise my component positioning too much (or make it very inconvenient). A friend did hugely shorten his arm cable once and in his system it made no difference other than making his phonostage/pre-amp positioning a right pain.
Lol! I don't obsess over it, but I'll always aim to keep signal cables as short as possible, which invariably produces sonically beneficial results (although I'd never overly compromise convenience, in order to achieve that) :)
Marco.
Common mode noise rejection:
Sorry Barry, but it's not factually correct to state that single ended systems can (or do) operate common mode noise rejection. Common impedance coupling takes effect in single ended systems where the shield (whether that be a braided shield or a core twisted around the signal) carries a ground voltage (common mode voltage) between two interconnected devices with NO common mode rejection at the receiving end. There is NO common mode rejection possible. Indeed, the only property of s single ended cable contributing towards reduction of noise is the screen impedance. The lower the better.
For fully balanced devices, there does exist common mode noise rejection. I have given a full description of how this operates in an article published HERE (http://referencefidelitycomponents.co.uk/design-of-interconnects/) for anyone interested.
Don't take my word for it though as there are plenty of papers available on the subject. Just a point that often confuses hence the offering of a fuller explanation.
Hello Paul,
Only just got around to reading your well written article cited above. I see now my misunderstanding: the impedance to earth will be much smaller than the load impedance, so there will only be a very small cancellation (virtually none) of the induced noise currents. Thanks for the explanation.
As I said, your article is very well written, though I do have some questions to ask and would point out one (possible) error. But as this lays outside the OP, I'll discuss this with you via a PM.
Regards
I had heard this about digital cables as well and once read a pseudo-scientific explanation that made sense - something to do with reflections. I was using a 0.7m digital cable at the time and it worked fine. For analogue cables I think the rule is the shorter the better - less capacitance, but not a big deal for a meter or so. More significant is that you don't stress your wires by turning them in too narrow a radius.
In theory it does depend on how well matched the DAC is to the 75Ohm system impedance. If it is poorly matched, then there will be reflections and you don't want these to alias with those signals created during the sampling window.
There are a few items written and published "out there", claiming that digital interconnects need to be at least 1.5m long. If this were true, is the connection between the CD reader in the transport and the internal DAC of an integrated CD player of similar length? I think not.
I have played around with digital cables of various lengths between 0.3m and 2m, and have heard no difference at all.
For the sake of neatness, I favour using interconnects that are only as long as they need to be.
How did the cables work out, what were the sonic differences between the different connection ideas?
I have only made one set so far and replaced the cheap ones from my Pre-amp to power amp.
They are a great deal better. Apologies as i am not up with all the lingo that the Audiophiles use but i can hear a big improvement.
Initially i thought they lost some bass compared to my cheap cables but i think the cheaper cables were just more muddy sounding and not as tight in bass so this made them seem Bass'ier. I turn up the bass knob slightly on my pre amp and got it back to the level i like. The cables are far less harsh at high volumes for higher pitched tones (voices/cymbals etc) so overall very good.
I ended up going with Furutech FP160(G) connectors and wrapped the shield strands around the return wire copper core. I used Cardas solder for the shield to the body of the plug. The signal core is screwed in as i didn't want to solder yet in case i had to remove the wiring.
I only bought 4 connectors to try out so need to get probably 12 more to complete my setup. When i have more i will make a set with the shield going to a spade terminal and connect this to my pre amp ground. Then i'll try and report on any differences.
Thanks
Powered by vBulletin® Version 4.2.3 Copyright © 2024 vBulletin Solutions, Inc. All rights reserved.