Complementary Input Gain = 10000 !!!

[lineup]

This is an idea I got today.
And the sim tests shows very good results.

When we use 4 input transistors in Complementary LTP configuration
we want to utilize the input gain as much as possible.
There are many ways to do this.
But not many ways take full advantage of full gain of both pairs.

This idea uses a mirror part to add the gain from the inverted input.
See red part in my circuit.
The output of resulting gain is taken from collectors of the blue output (VAS).

In this circuit I use the very good 2SC2362/2SA1016. Excellent transistors that can take high voltage.

All transistors in my schematic works at 5 mA current.

Lineup

[lineup]

Hello everybody
I attached one International Rectifier MOSFET output stage.
The harmonics spectra is shown below.

How can this sound Lumba, please tell ....
Some 'nastys':
--------
3rd -148 dB
7th -154 dB

[lineup]

A some of you didnt get my idea from first diagram.
Here is the simplified.

The description is still in my introductory post.

[janneman]

Hi Lineup,

The gain of your circuit is determined ultimately by your load resistor R6, which can be made arbitrarily large in a simulator. But once you start to drive a real impedance, you will most probably see that the gain collapses and that - without feedback - the distortion rises if you drive a (non-linear) Vas input R.

Your circuit is interesting and neat, but very often reality smashes dreams....

Jan Didden

[AndrewT]

If U5 & U6 swing nearly rail to rail, what happens to the swing capability of U7 & U8?

[AKSA]

Hi Lineup,

mount a complementary feedback pair on U5 and U6 and for most real loads it will be essentially load invariant.

You might also use the Lender VAS drive, base to one side of the LTP and emitter to the other, with suitably amended load resistors to keep it balanced, and you can eliminate the four transistors in the current mirror. This will have much less phase shift.

Hugh

[Lumba Ogir]

lineup,
Karma came around and bit you in the ***...
I would go for a less messed up VAS. Also, high open-loop gain does not lead to Nirvana, which should be considered established beyond the shadow of a doubt, after decades of heroically insistent experiments.