Super Symmetric Cork Sniffers' PI

[jazbo8]

Jaz, I was suggesting eliminating the current sharing between the cathodes, change to independant biasing / bypassing.
Pete

Got it.

Thanks,
Jaz

[VacuumVoodoo]

Alex, apologies but I can't see the positive feedback that you mention as being key to its operation?
Pete

Without the positive feedback into right hand triode grid combined with low value shared cathode resistor
the gain imbalance can be as large as 15db. As can be seen in the graph it is less than 0.1 db when this feedback is applied.
Or, if it makes it easier to grasp call it bootstrapping the right hand grid.
In a proper paraphase, with separate cathodes the feedback loops are broken.
Perhaps you'd like to simulate and investigate the circuit yourself?
After all, mental gymnastics is supposed to be good for us.

[jazbo8]

https://dl.dropboxusercontent.com/u/132 ... erters.pdf

Finally had sometime to run some sims... Without the positive feedback, the differential PI is indeed very un-balanced (as shown on the upper right chart).

The above sims show the results for perfectly matched sides, but when the two sides are not matched - with the differential pair + CCS tail configuration (and the cathodyne since only one side is used), the outputs remain pretty much in balance. However, with the proposed circuit, the outputs will be un-balanced unless further adjustments are made to the output load and/or the feedback ratio. Is that right?

Jaz

[roberto]

Thank you jazbo8!
I will search some plots I've done for a DC coupled concertina for a small amp I'd built in the second half of the year.

[VacuumVoodoo]

https://dl.dropboxusercontent.com/u/132 ... erters.pdf

Finally had sometime to run some sims... Without the positive feedback, the differential PI is indeed very un-balanced (as shown on the upper right chart).

The above sims show the results for perfectly matched sides, but when the two sides are not matched - with the differential pair + CCS tail configuration (and the cathodyne since only one side is used), the outputs remain pretty much in balance. However, with the proposed circuit, the outputs will be un-balanced unless further adjustments are made to the output load and/or the feedback ratio. Is that right?

Jaz

Nice Which SW are you using? MicroCap is pretty good but I'm not very happy with the graphics it delivers.
Correct observation re sensitivity to loading on the outputs.
The gain balance, or gain tracking if you will - balance trimpot affects gain in both sides, is affected by unequal loads.
However, loads differing by 5% either way from nominal skew the gain balance less than 0.25db.
Fun begins once you overdrive not the circuit itself but the power tubes it is feeding.
I'm pretty confident it'll work fine until it gets overloaded by a power tubes grids drawing current.

[jazbo8]

I am using TINA-TI, but only small circuits (<50 parts) can be sim'd... So I had to cut and paste the schematic and charts to put them together on a single page for printing.

Yes, the next order of business is to observe what happens when 1) the power tubes are over-driven and 2) when the PI itself is over-driven. Although, given the non-linearity involved, the simulation results may not be that close to reality...

Jaz

[jazbo8]

The best investigation into PI and PA clipping I have found is by katopan over at Wattkins, definitely worth studying. Will try to run some simulations with various PI configurations, to see if they even come close to the actual results.

PI Input & Output: http://www.wattkins.com/files/wattkins/PI_Clamping.pdf

PA Grid & Output: http://www.wattkins.com/files/wattkins/Pwr_Grids.pdf

Jaz

[VacuumVoodoo]

I'll be back...