Hi everyone:
Total newbie here.
Building a chipamp.com dual mono LM3886-based amp.
After having tested the PSU boards okay, it is now time for me to test wire the amp boards to the PSU boards.
1. I noticed in Peter Daniels' thread about building Gainclone-type amps, that he mentioned using something really thing (22-26 gauge) wires for signal to RCA. Does it matter whether you use thinner or slightly thicker wire for RCA to signal on amp boards? I have Cat3 wiring and Cat5 ethernet cabling I can use for this. Is thinner better? Doesn't thinner create more resistance?
2. Also, is it a bad idea to use thicker wire than the PCBs are designed for, as long as you cut some of the strands? Does this create any problens, other than possibly soldering mess? I've
3. Finally, is yellow a bad colour to use for RCA to signal pads? I'm already way overbudget and don't feel like going out to buy white or any other colour of new wire for signal connection. Also, they don't sell it anywhere nearby.
Thanks
Total newbie here.
Building a chipamp.com dual mono LM3886-based amp.
After having tested the PSU boards okay, it is now time for me to test wire the amp boards to the PSU boards.
1. I noticed in Peter Daniels' thread about building Gainclone-type amps, that he mentioned using something really thing (22-26 gauge) wires for signal to RCA. Does it matter whether you use thinner or slightly thicker wire for RCA to signal on amp boards? I have Cat3 wiring and Cat5 ethernet cabling I can use for this. Is thinner better? Doesn't thinner create more resistance?
2. Also, is it a bad idea to use thicker wire than the PCBs are designed for, as long as you cut some of the strands? Does this create any problens, other than possibly soldering mess? I've
3. Finally, is yellow a bad colour to use for RCA to signal pads? I'm already way overbudget and don't feel like going out to buy white or any other colour of new wire for signal connection. Also, they don't sell it anywhere nearby.
Thanks
For input signal wiring the thickness does not matter. Unbalanced connections are best done in coax; balanced using twisted pair.
If you can only connect some of the strands at each end, then you may in effect be only using the strands which happen to be connected at both ends. Better to use thinner wire and use all of it.
Use any colour you like; the audio signal doesn't care. Don't do all the wires in the same colour as you will find debugging very confusing! Invent a colour code and stick to it.
If you can only connect some of the strands at each end, then you may in effect be only using the strands which happen to be connected at both ends. Better to use thinner wire and use all of it.
Use any colour you like; the audio signal doesn't care. Don't do all the wires in the same colour as you will find debugging very confusing! Invent a colour code and stick to it.
DF96:
Thank you, that was very helpful. You mean a n RCA input might be done using RG59 or RG6 coaxial? I'm curious, what is the advantage of coaxial?
You mean with shielding removed, right? Just the conductor, right?
OH, and what gauge are the Chipamp.com LM3886 boards designed to accept? Does anyone know?
Thank you, that was very helpful. You mean a n RCA input might be done using RG59 or RG6 coaxial? I'm curious, what is the advantage of coaxial?
You mean with shielding removed, right? Just the conductor, right?
OH, and what gauge are the Chipamp.com LM3886 boards designed to accept? Does anyone know?
Last edited:
You don't need RF coax. Any shielded cable will do. For some reason I have never understood, many people seem to want to wire their inputs with twisted pair (and sometimes quite loosely twisted). Twisted pair is for balanced circuits. Use TP unbalanced and you can pick up hum and interference.
Your external input cables will, hopefully, use a coaxial construction so this should be continued up to the circuit input.
Your external input cables will, hopefully, use a coaxial construction so this should be continued up to the circuit input.
because a twisted pair help attenuate all interference impinging on the two cores of the signal circuit.
There is a second trivial reason.
Those that use coaxial cable read about "grounding faults" and think that if they break the screen connection at one end they cure all their problems, not realising they have just CUT the RETURN part of the circuit.
I am unclear whether you are merely reporting the confusion of others?AndrewT said:
Twisted pair attenuates interference in a balanced circuit in two quite different ways:
1. Electric field interference affects both wires equally, and a balanced circuit ensures that this means only a common-mode signal appears at the receiver - which it can reject.
2. Magnetic field interference is reduced by two means: closeness of the wires means the loop is small, twisting of the wires means that the loop orientation keeps reversing and so cancellation occurs. Whatever interference signal is left is differential-mode so the receiver accepts it.
When used in an unbalanced circuit the first method fails - all you get is a little screening (but not very much). The second method still works, provided that the far end is electrically floating.
You can use screened/shielded twisted pair in an unbalanced circuit, but then you are using three conductors where two will do the job.
1. fails !
For an unbalanced pair, 1. does convert some of the interference to common mode. That common mode signal is far below the interfering differential mode signal. i.e. the interference is attenuated. That is my understanding of reading H.Ott for unbalanced twisted pair.
My various experiments and builds seem to show that an unscreened twisted pair have a very good tolerance to interfering RF fields.
I have used in times along way back, a parallel pair in a screened format. This did not attenuate interference sufficiently for my needs. I abandoned using it.
I find that unscreened twisted pair are far superior.
For an unbalanced pair, 1. does convert some of the interference to common mode. That common mode signal is far below the interfering differential mode signal. i.e. the interference is attenuated. That is my understanding of reading H.Ott for unbalanced twisted pair.
My various experiments and builds seem to show that an unscreened twisted pair have a very good tolerance to interfering RF fields.
I have used in times along way back, a parallel pair in a screened format. This did not attenuate interference sufficiently for my needs. I abandoned using it.
I find that unscreened twisted pair are far superior.
A parallel pair (presumably untwisted) will pick up magnetic fields. Screening won't stop this for audio, but it will for RF.
Unscreened twisted pair will pick up electric fields when used unbalanced.
Three ways to stop electric fields causing interference:
1. screening - electric fields can't penetrate a good conductor
2. balancing - ensure both signal polarities receive the same interference (which means identical conductors, may mean twisting, balanced impedance load at both ends, good common-mode rejection at the receiver)
3. very low impedance at one or both ends for both conductors
Two ways to stop magnetic fields causing interference:
1. coaxial construction - in ideal case this gives zero loop size
2. reduce loop size - conductors close together, and twisting to reverse loop sense
- in both cases the cable itself must be the only signal return connection
Coax, and other single screened cables, use method 1. Twisted pair uses method 2.
I remain baffled why people want to use twisted pair for unbalanced connection, when one of the requirements for electric cancellation (identical impedances) is explicitly ruled out. The only thing which rescues them is if the source impedance is very low - but then it is the context which provides the protection, not the cable construction.
Unscreened twisted pair will pick up electric fields when used unbalanced.
Three ways to stop electric fields causing interference:
1. screening - electric fields can't penetrate a good conductor
2. balancing - ensure both signal polarities receive the same interference (which means identical conductors, may mean twisting, balanced impedance load at both ends, good common-mode rejection at the receiver)
3. very low impedance at one or both ends for both conductors
Two ways to stop magnetic fields causing interference:
1. coaxial construction - in ideal case this gives zero loop size
2. reduce loop size - conductors close together, and twisting to reverse loop sense
- in both cases the cable itself must be the only signal return connection
Coax, and other single screened cables, use method 1. Twisted pair uses method 2.
I remain baffled why people want to use twisted pair for unbalanced connection, when one of the requirements for electric cancellation (identical impedances) is explicitly ruled out. The only thing which rescues them is if the source impedance is very low - but then it is the context which provides the protection, not the cable construction.
I would use something like this: Audio Single Miniature Screened Single Core Grey (priced per metre) | Maplin
Cheaper than RF coax, and OK for short audio runs inside a metal case. I imagine something similar is available near you?
If you want to use RF coax then 75ohms impedance will have less capacitance than 50ohms impedance.
Cheaper than RF coax, and OK for short audio runs inside a metal case. I imagine something similar is available near you?
If you want to use RF coax then 75ohms impedance will have less capacitance than 50ohms impedance.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.