Not sure about a twisted pair, but it's better than no shielding at all. And yes, one end, the sensitive end, should be grounded, not both, from what I have read.
Better though , to use shielded wire, like a guitar cable or even coax, to isolate it from interference. Some RCA cable hi-fi connectore wires are shielded too, if you find decent ones. My impression too, is that if you have a big enough chassis and there's nothing nearby within an inch or so (heater wires, etc.), that you don't even need any shielding.
Space is good, and also rotate your tube socket mounting so that the filament supply is away from the input. Shorter lengths are the best, as well, for the sensitive signal input.
Better though , to use shielded wire, like a guitar cable or even coax, to isolate it from interference. Some RCA cable hi-fi connectore wires are shielded too, if you find decent ones. My impression too, is that if you have a big enough chassis and there's nothing nearby within an inch or so (heater wires, etc.), that you don't even need any shielding.
Space is good, and also rotate your tube socket mounting so that the filament supply is away from the input. Shorter lengths are the best, as well, for the sensitive signal input.
jarthel said:
You are indeed discussing a twisted pair with a shield around it, just to make sure, and are suggesting that only one end of the shield is grounded? Your responders assumed this, but I just wanted to make sure. As Miles & frank have suggested, yes this is good. If you really want to go all out, you can make that ground connection via a resistor, but this is usually overkill.
maybe I wasn't very clear.
================
I have 2 pieces of wire (1 red and 1 black) that I have twisted together. The red wire is the signal wire. The black wire is connected to ground.
only 1 end of the black wire is connected to ground.
<b>There is no shield (like those found in coax) around the twisted wires.</b>
================
I have 2 pieces of wire (1 red and 1 black) that I have twisted together. The red wire is the signal wire. The black wire is connected to ground.
only 1 end of the black wire is connected to ground.
<b>There is no shield (like those found in coax) around the twisted wires.</b>
Miles Prower said:
When returning from the speaker connector + the feedback cable, if coax is used with shield connected to local earth node of first or split stage (where feedback is fed in a schematic anyway), there is also a - speaker connector point that needs to be ground referenced back there. Will it suffice to be a simple wire returning, twisted around the coax or it must be a second coax maybe twisted with main return coax, with both core & one way shield earthed in the small stage ground again?
salas said:
That ground reference from the speaker output should already exist as a connection to the star point. Adding another return in parallel is likely to create a ground loop that injects PS noise.
Nothing like this should be necessary, and would most likely cause a ground loop. The whole idea behind star grounding is to provide one central ground connection so that the filter charging current pulses have but one way to go, so that there won't be any noise occurring in any other part of the circuit due to ground loops.
"Be careful about the additional stray capacitances that this can cause. The cure may be worse than the disease."
Of course, you take that into consideration. On a project where I used 75 Ohm RF coax to connect to the gNFB summing node, I didn't need an external compensating capacitor since the cable capacitance was enough to serve the purpose.
There are two reasons to use twisted pair that I know of.
(1) When you have a balanced signal driving a differential amplifier. In this case any noise is picked up by both wires equally and with the same phase. The diff amp rejects the noise because it is all common mode; it is the same in both wires. The diff amp is sensitive only to signals that are different between the wires, but the noise appearing in the two wires is the same.
Note that balanced does not mean that there is necessarily the same music signal appearing in each wire (but out of phase of course.) Balanced means that both wires see an equal impedance to ground at both ends. That assures that any noise picked up is equal in both wires and can be 'canceled' by the diff amp. The actual signal to be amplified might be on one wire only. Noise will still be rejected.
(2) The other reason to use twisted pairs is when there is a large AC current through the wires and they are part of a loop (i.e. one wire can be thought of as the return path for the other.) In this case the radiated fields of the two wires cancel. Heater supply wires are a good place to do this because it reduces hum radiated to other parts of the circuit.
The application proposed here serves no really good purpose that I can think of. The twisted-in ground wire does not act as much of a shield. It does contribute additional capacitance to ground, which we generally want to avoid. I probably wouldn't bother, but don't let me stop anyone from experimenting.
-- Dave
(1) When you have a balanced signal driving a differential amplifier. In this case any noise is picked up by both wires equally and with the same phase. The diff amp rejects the noise because it is all common mode; it is the same in both wires. The diff amp is sensitive only to signals that are different between the wires, but the noise appearing in the two wires is the same.
Note that balanced does not mean that there is necessarily the same music signal appearing in each wire (but out of phase of course.) Balanced means that both wires see an equal impedance to ground at both ends. That assures that any noise picked up is equal in both wires and can be 'canceled' by the diff amp. The actual signal to be amplified might be on one wire only. Noise will still be rejected.
(2) The other reason to use twisted pairs is when there is a large AC current through the wires and they are part of a loop (i.e. one wire can be thought of as the return path for the other.) In this case the radiated fields of the two wires cancel. Heater supply wires are a good place to do this because it reduces hum radiated to other parts of the circuit.
The application proposed here serves no really good purpose that I can think of. The twisted-in ground wire does not act as much of a shield. It does contribute additional capacitance to ground, which we generally want to avoid. I probably wouldn't bother, but don't let me stop anyone from experimenting.
-- Dave
Hang on, guys,
While all the advice is sound, Jarthel is talking about signal wires inside a chassis. How long could they be, unless the layout is horrifically misplanned?
I also wondered about the "high impedance end". The output and following input impedances may differ, but once they are coupled by a wire of a few cm. the impedances are exactly the same at audio, viz. that of the lowest impedance side (hopefully the driving side). If the topology is such that earth loops are of consequence inside a chassis, then earth only one side of the screen, but I really think it is immaterial which side. (Such can be problematic with steel chasses and poorly placed power transformers.)
But I would agree that a twisted pair would not really make a difference in audio. One need not fear a few pF of screen wire capacitance unless this is in the wrong place as e.g. a NFB (long) lead as Miles said, and then only concerning feedback stability in the supersonic region. Or have I just been singularly lucky in the past?
Regards
While all the advice is sound, Jarthel is talking about signal wires inside a chassis. How long could they be, unless the layout is horrifically misplanned?
I also wondered about the "high impedance end". The output and following input impedances may differ, but once they are coupled by a wire of a few cm. the impedances are exactly the same at audio, viz. that of the lowest impedance side (hopefully the driving side). If the topology is such that earth loops are of consequence inside a chassis, then earth only one side of the screen, but I really think it is immaterial which side. (Such can be problematic with steel chasses and poorly placed power transformers.)
But I would agree that a twisted pair would not really make a difference in audio. One need not fear a few pF of screen wire capacitance unless this is in the wrong place as e.g. a NFB (long) lead as Miles said, and then only concerning feedback stability in the supersonic region. Or have I just been singularly lucky in the past?
Regards
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.