Using a regulated ps for a power amp input stage.

[MASantos]

In the first watt amplifiers we have the input stage being feed from the unregulated power rails.

I like the idea of using a regulated supply for powering the input stage of the amp, and then having a separate unregulated supply for the output stage which is more tolerant as far as I understand.

From reading around I believe that cascoding can be seen, to some extent, as some sort of regulation, since it protects the input transistors from higher voltages in the power rails, does it also improve the quality of the supply?

What would be a simple regulator one could piggyback into an f5 board for example?

 

[AudioSan]

cascoding is no way the same as regulated PSU. the cascoding just shave of some amount of the voltage. the cascoded voltage is determined by the rail voltage and will swing paralell to it.
a simple LM317/337 regulator is the easiest way to regulate the inputstage. or LM78xx/79xx.

 

[Nelson Pass]

On some amplifiers like the F5 and J2 you cannot separate the main rails from
the front end as the output devices are operate "Common Source" and the
front end develops voltage to drive them referenced to the main rail.

Otherwise, it's perfectly OK to regulate the front end supply.

 

[Zen Mod]

If construction of amp in question allows it , then reg is welcome

edit: for second time today , Pa was faster

so , considering that Big Man is in thhe office , I'm going to Lalaland now , to squeeze that flu out

 

[AndrewT]

cascoding is no way the same as regulated PSU. the cascoding just shave of some amount of the voltage. the cascoded voltage is determined by the rail voltage and will swing paralell to it.
a simple LM317/337 regulator is the easiest way to regulate the inputstage. or LM78xx/79xx.

The cascode could be referenced via a resistive voltage divider between rail and ground. That type would swing with every change in supply voltage including interference.

The cascode can also be referenced to a fixed voltage set some way above the signal ground, or to the LTP emitter voltage and be virtually impervious to variations in supply voltage and interference.

 

[2 picoDumbs]

Build yourself a shunt regulated BA3

 

[MASantos]

So I've been going throught the ba3 article and schematic, and due to my limited electronics knowledge can't seen to understand the need for the capacitor on the output of the ba3 input stage. the circuit is quite similar to the f5 and it doesn't have that specific cap. it is because of not having a feedback loop?

 

[flg]

Do you mean at the "D" output? Because that node would be connected to the P channel bias of the BA-2 output stage. Or N Channel bias on BA-1 output stage

 

[MASantos]

Yes, I mean the D output, but it is in the signal path right?

 

[flg]

Yes.

 

[Zen Mod]

Buddha is everywhere .

even in signal path .

 

[generg]

So I've been going throught the ba3 article and schematic, and due to my limited electronics knowledge can't seen to understand the need for the capacitor on the output of the ba3 input stage. the circuit is quite similar to the f5 and it doesn't have that specific cap. it is because of not having a feedback loop?

Nelson in the BA-1 article to a similar problem

"and I chose to use a 10 uF polypropylene type here, as the only capacitor in the signal path. Strictly speaking you do not have to use this capacitor, but it does make adjustment and long term DC stability very easy to achieve. Without it, you will be using P201 to adjust the output offset, and there will be considerably more drift."

 

[2 picoDumbs]

Buddha is everywhere .

even in signal path .

And ZM is always in more than 1 place at the same time.

 

[flg]

Nelson in the BA-1 article to a similar problem

"and I chose to use a 10 uF polypropylene type here, as the only capacitor in the signal path. Strictly speaking you do not have to use this capacitor, but it does make adjustment and long term DC stability very easy to achieve. Without it, you will be using P201 to adjust the output offset, and there will be considerably more drift."

Using the BA-3 FE for a BA-1 output stage without that cap will mean you'll have to adjust your BA-3 for a +4 or +5V offset to bias the output stage. There is no "P201" in BA-3. He may have also been refering to removal of the bias circuit on the output side of C203 and letting the diff output set the bias.

 

[generg]

You are right!

I would not do it.....!

The only possible valuable of my post may be that using no cap from frontend to outputstage is an arduos affair....and maybe not worth the effort.

 

[flg]

Yes, I would leave the BA-3 output cap in place and use the suggested biasing on a BA-1 or 2 output stage. Using BA-3 only as a preamp it may be possible to go without it?

 

[Zen Mod]

same problem

 

[MASantos]

Then why doesn't the f5 need an output cap?
What changes to the circuit would be necessary to remove the cap? Feedback?

 

[Zen Mod]

yup

due to lack of feedback , BA3 FE is less stable then F5 , DC offset wise

but , only generally - it depends of actual physical execution

so - make it , then observe - if you have enough DC stability prior to cap , then - don't use it

if not , some sort of servo is mandatory