Re: Isolation transformer questions
Posted: May 9th, 2020, 12:46 pm
tomp wrote:
Your statement that on a balanced feed the noise would be equal on both lines and out of phase is only true if the transformer is passing a noise from the primary that is also equal and out of phase. If the noise on the primary is being received from an external source that is radiating into the lines depending on the frequency and the layout of the feed lines it would most likely be in phase on both lines. That is where balanced circuits especially with transformers reject common mode signals, that is signals that are in phase so that in phase noise would not make it through the transformer, ignoring capacitive coupling for this discussion. If they were out of phase, that is differential, the common mode rejection (CMRR) would not reject them because they are not common mode and they would pass to the secondary.
The out of phase voltages that would appear across the center tapped secondary would not truly cancel at the center tap. The voltage at the center tap is the middle of what ever total voltage appears across the secondary. Even if you ground that point the voltage relationship between the leads does not change and no current will flow to ground since the secondary is isolated from ground except at the point where you have made the connection. There is no circular path for any current to return to either lead of the secondary, again ignoring stray capacitance. All you are doing is referencing the center tap to a ground but that does not change the secondary lead voltages with respect to each other. If you connect a load across the whole secondary it would see the whole noise voltage on the secondary. If you connect from one lead to the center tap it would see half of the total noise. On the transformer, you cannot connect the two leads to each other or you would have a direct short. In that case the out of phase voltages would cancel along with the production of a lot of smoke.
If you look at the circuit diagram of the Corcom filter, there are only two lines plus a ground. They do not differentiate between a hot and neutral line. In fact, when the Corcom is used at 240 volts on a traditional split phase, ie balanced center neutral house feed both inputs are hot. The key to the figure is that both lines have inductors and capacitors in combinations that shunt a lot of the noise from either lead and either polarity to ground regardless where it comes from.
"Balanced" has nothing to do with equal voltages. Balanced is about equal impedances to ground. If the Corcom filter is fed from a balanced source, common mode signals of equal value would theoretically cancel to 0v at ground in the filter. If a Corcom is fed from a grounded-neutral 120v supply, the neutral is grounded at the source, where the hot side is not. (However, with the neutral grounded at the source, the common mode noise may not be equal on both legs in the first place.) If a "balanced" 240v w/neutral feeds the Corcom with equal impedances to ground, common mode noise can be cancelled at ground.