Quote:
Remember that we are not talking about dead-end streets here......for currents to flow there must be a return path, correct? No matter how much voltage you apply to one lead of a light bulb, you need to connect the other one back to the other side of the battery or AC source before it lights up.
Keep that in mind.........
Let's assume that we have a source driving a power amplifier, and that source does indeed have a DC offset at its output. That would not be a good thing, because the amp would multiply that offset by its gain, and pass that as a much higher DC current through the speakers/cans, heating up the voice coils, all sorts of bad effects.
How does one prevent that? With a blocking cap on the input of the power amp, in series with the signal. A cap will not pass DC, only AC.
If DC current can't flow past the input cap, it cannot flow out of the source, can it? Where would it "go", because its path to ground is blocked by the input cap? It cannot build up in the cable, right?
If the DC current cannot flow forward through the signal lead to the amp, no DC current can flow back through the ground side to the source. Meaning.......there is no net DC current flow through the loop with an input cap present.
The input cap cannot eliminate the voltage offset, but it can block the DC current flow.......sort of like turning off a faucet stops the water flow, but it doesn't make the static pressure in the supply pipe go to zero. The potential for flow is not eliminated, it's just blocked.
DC offset of a source output thus means only that there is a potential for a DC current to flow, if the balance of the circuit allows that.....and that cap prevents the flow.
There are DC servo designs that prevent that offset, too, so that an input coupling cap is not always needed.
Also, let's presume that a device does have a 30 mV DC offset (about ten times what many actually have). Let's first assume that means the signal lead at idle is 0.03V positive with respect to ground. The output stage of the device is works on a differential basis. A typical 2.0 Vrms full scale output means that the max voltages, if "centered" on true ground, would run from +2.85V to -2.85V. If there is an offset, that would mean the output voltage would vary from +2.88V to -2.82V......and that would allow a small "forward" current to exist. But what if the offset was still 30 mV but negative with respect to ground? The output signal would vary from +2.82V to -2.88V. The same offset would be said to exist, but the potential DC current flow would be in the opposite direction.
I'm not an expert on circuit design, but posts on various audio forums show that offsets are not necessarily just positive or negative with respect to ground, so the direction of any potential DC current flow would be a crap shoot. If that would be responsible for directionality of cables, then wouldn't you have to know that.....and attach the cable in the direction that was right for that particular case? Never heard of a manufacturer recommending that!
Originally Posted by circularlogic /img/forum/go_quote.gif 
|
Remember that we are not talking about dead-end streets here......for currents to flow there must be a return path, correct? No matter how much voltage you apply to one lead of a light bulb, you need to connect the other one back to the other side of the battery or AC source before it lights up.
Keep that in mind.........
Let's assume that we have a source driving a power amplifier, and that source does indeed have a DC offset at its output. That would not be a good thing, because the amp would multiply that offset by its gain, and pass that as a much higher DC current through the speakers/cans, heating up the voice coils, all sorts of bad effects.
How does one prevent that? With a blocking cap on the input of the power amp, in series with the signal. A cap will not pass DC, only AC.
If DC current can't flow past the input cap, it cannot flow out of the source, can it? Where would it "go", because its path to ground is blocked by the input cap? It cannot build up in the cable, right?
If the DC current cannot flow forward through the signal lead to the amp, no DC current can flow back through the ground side to the source. Meaning.......there is no net DC current flow through the loop with an input cap present.
The input cap cannot eliminate the voltage offset, but it can block the DC current flow.......sort of like turning off a faucet stops the water flow, but it doesn't make the static pressure in the supply pipe go to zero. The potential for flow is not eliminated, it's just blocked.
DC offset of a source output thus means only that there is a potential for a DC current to flow, if the balance of the circuit allows that.....and that cap prevents the flow.
There are DC servo designs that prevent that offset, too, so that an input coupling cap is not always needed.
Also, let's presume that a device does have a 30 mV DC offset (about ten times what many actually have). Let's first assume that means the signal lead at idle is 0.03V positive with respect to ground. The output stage of the device is works on a differential basis. A typical 2.0 Vrms full scale output means that the max voltages, if "centered" on true ground, would run from +2.85V to -2.85V. If there is an offset, that would mean the output voltage would vary from +2.88V to -2.82V......and that would allow a small "forward" current to exist. But what if the offset was still 30 mV but negative with respect to ground? The output signal would vary from +2.82V to -2.88V. The same offset would be said to exist, but the potential DC current flow would be in the opposite direction.
I'm not an expert on circuit design, but posts on various audio forums show that offsets are not necessarily just positive or negative with respect to ground, so the direction of any potential DC current flow would be a crap shoot. If that would be responsible for directionality of cables, then wouldn't you have to know that.....and attach the cable in the direction that was right for that particular case? Never heard of a manufacturer recommending that!
