[Porksoda]

So, there are three TLEs in the ppa. Seeing as the rails on the three op-amps will not be identical, as they will all have different voltage drops across the transistors, won't the vitrual ground at each op-amp's power supply be different? I would think this could cause problems, since all of the virtual grounds are tied together...

 

[Glassman]

good point, I talked about that with tangent, his conclusion is - it works, so don't bother

(something alike) and I second that after all even if it's suspicious..

 

[tangent]

We haven't actually measured it, but what we believe is happening is that the voltage differences among the three TLEs' outputs causes a constant current flow through that big ground plane to keep everything balanced. Since it's a direct current and it's small, it isn't worth worrying about. It would matter if the current were large enough to warm the TLEs, or if the current were AC, since that would couple into the input lines running under the ground plane.

 

[Porksoda]

My thinking is that as the power the op amps are drawing will vary, the drop across the filtering transistors will vary, and so the op amp rail voltages will vary at a frequency similar to the output of that channel, causing each of the grounds to oscillate at different frequencies, possibly creating crosstalk, since they are all tied to the same ground?

Actually the main reason I brought this up is I am building my own PPA-esque amp on protoboard and I THOUGHT the one op-amp power suppy was for all 3 op amps, not one PER op amp, so I am having to go back and buy more parts now.

Perhaps I should just put one TLE on the ground channel and leave it at that, since it will be the most stable, then leave the TLE out of the other two power supplies?

 

[morsel]

We would not have made the PPA as it is if the TLEs fought each other in a way that caused oscillation or instability, however, it will work with only 1 supply for all 3 opamps, with the provisos that all 3 opamps should have bypass caps near their power pins, and the FET current sources are cranked up to support 3 opamps. PPA v1.0 provided jumpers to do this.

You are making inaccurate speculations, then contemplating decisions based upon those speculations as if they were fact. Perhaps you might search for and read the voluminous threads on the PPA before proceeding, they will provide all the background you need on PPA design and development.

 

[tangent]

Quote:

My thinking is that as the power the op amps are drawing will vary, the drop across the filtering transistors will vary,

Yes, the difference being on the order of microamps causing a drop on the order of microvolts.

That voltage drop, though, will be absorbed by the C4/C5 caps. Or if you want to look at it the other way, the dynamic current will come from them in addition to the Q3s. The point being, you won't even see the microvolt ripple on the rails that would be there without the caps.

EDIT: Furthermore, the dynamic current across the V+ Q3s will cause the same dynamic current to go through the V- Q3s, causing a similar drop. Since the voltage at the TLE OUT pin is simply 1/2 the IN to COM voltage, when both rails drop towards zero by the same amount, the TLE output voltage shouldn't change.

Quote:

the op amp rail voltages will vary at a frequency similar to the output of that channel, causing each of the grounds to oscillate at different frequencies

If your PPA clone is oscillating, this isn't the reason.

 

[ITZBITZ]

I knew I spent hours matching my transistors for some reason...