Frankenstein

[ppereira]

I call this thing Frankenstein because it is a collection of many different ideas that may not work in the end. In Spice however it adds up. I know, so did Dr. Frankestein's creation

It goes against many accepted good principles i.e. simplicity is better, no NFB is better etc. (has plenty of both complexity and feedback)

Don't know how it sounds, it's simply an exercise in getting to a circuit that measures well.

Here's some Spice data:

Total Harmonic Distortion: 0.000098% (2V, 1KHZ)
Output Impedance: 416.525

[ppereira]

Frequency Response

[ppereira]

Harmonic Frequency Fourier Normalized Phase Normalized
Number [Hz] Component Component [degree] Phase [deg]
1 1.000e+03 6.645e+00 1.000e+00 179.97° 0.00°
2 2.000e+03 3.602e-06 5.420e-07 -0.02° -179.99°
3 3.000e+03 1.140e-06 1.715e-07 -1.22° -181.20°
4 4.000e+03 2.368e-06 3.563e-07 0.21° -179.77°
5 5.000e+03 1.893e-06 2.849e-07 0.04° -179.93°
6 6.000e+03 1.580e-06 2.377e-07 0.11° -179.86°
7 7.000e+03 1.336e-06 2.011e-07 0.01° -179.96°
8 8.000e+03 1.185e-06 1.783e-07 0.14° -179.83°
9 9.000e+03 1.058e-06 1.592e-07 0.10° -179.87°

[Zen Mod]

well.....if you like it-go for it;

I prefer non-overall feedback for toob preamp

[AndreasS]

How this global feedback works?

Andreas

[rdf]

The big caution with Spice is devices are perfectly matched. Stages 2 and 3 look highly dependent on symmetry (and second order cancellation?) for performance. There a probably a few 6bq5 models floating around, might be worth trying a simulation with mismatched tubes.

[ppereira]

Hi RDF. That was a concern. Then I replaced 1/2 of the EL84's with other tubes (6SN7) and even though distortion increased it was still under .001% total which was totally unexpected... I couldnt say the same for resistor and capacitor values. Any variance there dramatically changes measurements.

[AndreasS]

Hi Ppereira,

your overall feedback dosn't work: R1 must be connected to ground and not to the feedback path.

best regards
Andreas

[ppereira]

It does work. R1 connects to ground via R5. See Jadis JP80 schematic, same arrangement.

[AndreasS]

In the Jadis schematic the grid resistor of the input tube is connected to ground. In your schematic the feedback signal is given on the grid AND on the cathode of the first tube, without influence for the tube.

Andreas