The "Menno Cell" - Trans Technology

[SpreadSpectrum]

By the way, I never did like to have current sources in both the bottom and the top of a circuit, not single ended, not differential or invertor . . . very tricky.
The only good way to do that is to set one current source to slightly less current than the other current source, and then put a resistor across the lesser current source to 'suck up' the excess current from the greater current source. And now, the lesser current source becomes a high resistance imperfect current source.
Making two independent current sources have exactly the same current over changing voltages and changing temperature is like a dog chasing its tail.

It doesn't just have current sources for loads, it's a current mirror. Again, this is simply a copy of the input stage of practically every solid-state amp out there with tubes substituted for some of the transistors.

The multiple current sources really aren't tricky for this purpose; the delta between the current set point for the tail and the current set point for the mirror loads is simply the quiescent sink current to drive the output stage feedback resistor. You need some quiescent current flow there, so it's just a simple subtraction.

 

[Tubelab_com]

I never did like to have current sources in both the bottom and the top of a circuit, not single ended, not differential or invertor . . . very tricky.

I found this out 20 some years ago when I reinvented the flip-flop.

It doesn't just have current sources for loads, it's a current mirror. Again, this is simply a copy of......

Current mirrors are required learning in the IC design world. The research group where I spent the last 12 years of my career at Motorola slowly morphed into an IC design center. I preferred building real prototypes to sitting in front of a computer simulatin' and calculatin' but I had no problem lifting their circuits and using them in a tube amp......wanna make a 50 watt type 45?

Take a real 45 load it with a resistor, CCS, a small OPT with a resistive load, or even a reactive speaker simulator. Insert a current mirror in its cathode circuit with a 5X or so current boost (multiple identical transistors) in the cathode circuit of a BIG fat TV sweep tube, which is driving a big OPT. It takes some tweaking to get it right but it does work.

This can be reduced to nothing more than the 45 and the current mirror. The fat sweep tube can be replaced with a big chunk of silicon, and you can make some pretty curves.....load it into an OPT and crank it up, and you will learn all about silicon's SOA all over again.

I wound up following a different path with the UNSET, but the SOA / secondary breakdown thing is a real problem even with mosfets. The SOA ratings presented in the data sheets of many parts, especially anything from ON Semi / Fairchild are WRONG. That 150 watt mosfet becomes more like a 20 to 50 watt mosfet when operated anywhere above 1/3 of its maximum voltage rating. I have a big pile of blown / exploded mosfets and a few shattered vacuum tubes to prove it.