Isn't this capacitor backwards? In figure 15 the capacitor, Csvr, is meant to AC couple the AP generator to the DUT and DC supply. Since this is a polarized capacitor this seems backwards to me. I showed this diagram to several other people, and they agree.

Infact, a lot of these capacitors look backwards. Unless this serves to be a general case...I do not know what the internal connections of each audio amp would look like.

http://focus.ti.com/lit/an/sloa068/sloa068.pdf

 

Many schematics do not use a separate symbol for unpolarized capacitors. The only way to determine if a capacitor is unpolarized is by referencing the parts list that accompanies the schematic or by examining the actual capacitor.

Remember that the minus line in this circuit is part of a differential pair. Therefore, minus actually means 180 degrees out of phase with the plus line. Not necessarilly a negative voltage with respect to ground. Both [minus and plus lines] could have signals that vary from 0 volts to some positive voltage. On the minus line, the waveform is inverted.

 

Many schematics do not use a separate symbol for unpolarized capacitors. The only way to determine if a capacitor is unpolarized is by referencing the parts list that accompanies the schematic or by examining the actual capacitor

Unfortunately, this is an application note on various Audio Precision measurements. There is no BOM available.

Remember that the minus line in this circuit is part of a differential pair. Therefore, minus actually means 180 degrees out of phase with the plus line. Not necessarilly a negative voltage with respect to ground. Both [minus and plus lines] could have signals that vary from 0 volts to some positive voltage. On the minus line, the waveform is inverted.

If that is the case, then that would mean the (-) output of the AP is shorted to the ground of the power supply. But here is how the output of the AP is configured:

The output of the Audio Precision is transformer coupled. The app note states that it is set to 'Unbalanced Floating.' In this configuration the (+) output is off the top of the secondary through a low impedance which is 20 ohms in this case. The (-) output is connected internally to the center tap. The (-) output is then connected to ground more than likely earth ground because of the DC power supply.

The node between Csvr and Rsvr would measure the DC supply voltage with whatever AC squiggle from the AP generator. The other side of Csvr (output of the AP) would, to me, measure 0V DC. This is why it looks backwards to me. The app note speaks of using electrolytics because it will be a large value cap to create the LPF, and most electrolytics are polarized in my experience.

 

Thank you for the explanation.

[In fact, a lot of these capacitors look backwards. Unless this serves to be a general case...I do not know what the internal connections of each audio amp would look like.]

Maybe I misinterpreted the above paragraph from message #1. When I wrote about differential outputs, I was referring to capacitors labeled Cin and Cana in figure two. I should have been more specific.

In figure 15, the AP Gen output also has a note stating CHA ON.
Maybe I'm mistaken, but aren't the phrases "... FLOAT" and "CHA ON" contradictory?
Unbalance refers to the fact that the minus pin is grounded. The text above the figure states that the AP Gen and Csvr are not needed.

 

One dodge is to use bipolar caps and duck the polarity issue.

 

In figure 15, the AP Gen output also has a note stating CHA ON.
Maybe I'm mistaken, but aren't the phrases "... FLOAT" and "CHA ON" contradictory?

Well, you have to turn on channel A in the GUI software for the Audio Precision.

Unbalance refers to the fact that the minus pin is grounded.

Which makes me wonder why selecting UNBALANCED FLOAT would have been mentioned when you could select UNBALANCED GROUND as well. You are connecting the (-) to ground anyway.

The text above the figure states that the AP Gen and Csvr are not needed.

No, but it is the easiest method to test ksvr. PSRR is simple to test for, and I do not have a transformer to couple in an AC signal.