Complementary feedback pair / Sziklai as DC filtering, solving

[triedgetech]

nvm, didn't get help with thevenin process

[wasedadoc]

No, Rth is closer to 10k.  When you short out the voltage source you have zero Ohms across the 15k.  Then you have two 4k7 resistors in series..

[Marco]

Load at VCC is varying between 5mA to 15mA.
I'd like to solve the circuit to see if either BJTs are in active region.

If that is all you want to know, don't make life hard. 5ma*180 is 900, so the PNP is always active barely.

Increase R33 10x so it's definitely active, then the base current through the NPN is a bee's dick and R36/R37 don't matter much. Though R37 could be a wee bit smaller for lower noise. Currently they will drop around half a volt, which seems a lot.

Your zener is wrong way round BTW. R39 is just a bias resistor for the zener, the low pass filter sees the differential resistance of the zener for AC and a voltage drop from the input for DC approximately. Save Thevenin for when it matters.

[Marco]

You mean D2? That's a schottky, not zener.

Ignoring R36/R37 that's not giving the NPN much headroom to work with, Schottky diode voltage is close to Vcesat. It works right now because R37/R36 are dropping so much voltage, but if you're going to rely on that you might as well make R36 larger and forget about the diode.

LTSpice simulation showed that the bigger the resistance value at the base of NPN, the better sziklai circuit filters out noise...

Did you try changing R37 (the base resistor) separately from R36 (the low pass resistor)?  R37 is just there for stability, but it will amplify base current noise.