My new design or updated old one

[astankov]

Very well Alex you added the missing caps. The 470K resistor in the input marked green is connected to main ground will it be better to be connected to the lifted GND? Also 56K resistor marked red needs to be connected right after the 1uF cap and then to the lifted GND now is wrong.

On the picture is the TO220 version can you please post the TO126 version?

If you fix that I think we will be ready for VER2.0

[astankov]

Alex if will be a big problem to fix the 470K resistor leave it - it is not absolutely needed.
It is recommended but not critical circuit will work well without it.

[astankov]

Originally Posted by keantoken

R36 will degrade cascode quality, why is such a large value necessary? Needing such a large resistor here for anything other than voltage drop would seem to indicate a problem.

According to this simulated with lower value resistor and when the value go under 2K it seems THD get a bit lower. How will this correspond with reality?
What will be right value for R36 1K maybe or not?

Regards!

[keantoken]

R36 magnifies the affect of parasitic capacitance and Early effect in the cascode transistor. This will be observed as a modulation of the cascode voltage while playing. What you have done is to contaminate the cascode reference with the signal you were trying to eliminate in the first place. You can bypass the resistor entirely. If this causes instability, add a small 1k trimmer as R36 and adjust it until you are satisfied with the stability. Then measure the resistance and substitute a fixed resistor. It would not be strange if a small resistor say 50-330R were needed here; but not a large one.

In reality the affect on distortion could be better or worse than simulations show, depending on simulation quality and prototype layout. At low Vce, the transistor will enter quasi-saturation, and this will cause extra distortion not predicted by simulation (models do not include quasi-saturation). You may never discover it unless you test your amp at high volumes. The 10k resistor will make this effect worse.

[astankov]

Thanks very much for the explanation! Example with the trimmer is very good will play a bit with 1k trimmer and see what will happens but let first PCB get ready and make the new build.

Regards!

[Mooly]

This is where it is useful to scope the actual design for real as well as simulate it. Scoping wouldn't show changes in absolute distortion though.

I wouldn't like to say really... perhaps experiment with the ref volts (zener) as well as the resistor value. There are always trade offs and compromises.

[astankov]

Today I played a bit with the ready board. Found the problem with the hum it was a leaky cap in my test power supply After replacing hum disappeared Also fixed the reversed diode and tested with lower value resistor for R36. Decreased it first to 390R and then to 68R. With 390R sound was more detailed and crystal clear and can't hear any difference between 390R and 68R. Well decreasing the resistance makes HF response better. Probably will leave the 68R for this resistor.

My hands are itchy now for building VER2.0 and can't wait

[Mooly]

I'm pleased it's all working out.

Roll on V2.0

[alex mm]

..... I have change position for 56K and 470K and I forget to post in thread
Alex.

[keantoken]

Astankov, can you list the transistors you can use that you are reasonably sure aren't fakes?