Are you really fine with IC voltage regulators ?

[ginetto61]

Originally Posted by Mooly

No, not after Can you see why?

Hi no
if the voltage drop across the cap multiplier is constant i will have just to add this V drop when setting the voltage with the IC reg stage.
If the cap multiplier has no self noise, while the IC regulators have it, i would put it at the end as a last filter
If not the self noise of the IC stage will reach the circuits to be supplied unfiltered.
Anyway I really have no clue about how big is this self noise of the IC regulators The datasheet report only the psrr of the IC regulators
I will check better
Found something and i am reading
What Linear Regulator Chips Have the Lowest Noise Figures?

[Mooly]

The volt drop is one thing. The whole point of a regulator is to provide a stable voltage that doesn't change in response to load variations.

A cap multiplier would be very poor in that regard. Also you have the volt drop to contend with.

Most voltage regulators like the 78/79 series work best with just a small cap at the output for stability (see the data sheets for details). Large caps destroy the transient response of the regulator.

[ginetto61]

Originally Posted by Mooly

The volt drop is one thing. The whole point of a regulator is to provide a stable voltage that doesn't change in response to load variations

Thanks but i am mainly interested in low consumption class A circuits so the current draw should stay constant. And while at the output of a IC regulator is not recommended to put big uF after a cap multiplier i could put thousands of uF without any issue i guess ... a sea of capacitance that should take care of any peak of current draw

A cap multiplier would be very poor in that regard. Also you have the volt drop to contend with.
Most voltage regulators like the 78/79 series work best with just a small cap at the output for stability (see the data sheets for details). Large caps destroy the transient response of the regulator

yes as i said above ... the response to transients will be taken care by the last big cap after the cap multiplier. Those caps could be placed also locally on the preamp board. Like 1000-2200 uF per rail ... it is a lot.
But after understanding all these issues about IC regulators i wonder if a discrete regulator with zener would be a more hassle free option. Together with a cap multiplier ... before or after ?

[jean-paul]

A resistor and a Zener diode. That will do

[ginetto61]

Originally Posted by jean-paul

A resistor and a Zener diode. That will do

Hi thanks ! i was thinking to that. But then someone told me how noisy are zeners ... it never ends

[jean-paul]

It was just joking as you think a lot (good!) but lack basic knowledge. Of course IC regulators are better than a resistor/Zener diode combo. Of course are ultra low noise regulators (and not their 40 year old cousins) the way to go certainly when you lack basic electronic skills. Of course you should not limit everything to just 1 application. Etc.

As answers are opinions in many cases you could use your time to design a PCB for the PSU you want. Include all necessary stuff on 1 board, add extra pads for options and experiment in practice. You could add pads for both LT4320/MOSFETs and also Schottky diodes, you could add CLC filtering as many don't because coils are expensive, you could choose footprint of a few modern regs so your designs a bit more universal, you could add various footprints of various sizes filter caps (tip: snap in caps are nearly always better). You could make it RF proof while the air is full of RF. You could show the design here and the progress you make and get useful directions. In the end you could end up with a real PSU.........

Cap multipliers are no substitute for regulators.

[ginetto61]

Originally Posted by jean-paul

It was just joking as you think a lot (good!) but lack basic knowledge. Of course IC regulators are better than a resistor/Zener diode combo. Of course are ultra low noise regulators (and not their 40 year old cousins) the way to go certainly when you lack basic electronic skills. Of course you should not limit everything to just 1 application. Etc.
As answers are opinions in many cases you could use your time to design a PCB for the PSU you want.
Include all necessary stuff on 1 board, add extra pads for options and experiment in practice. You could add pads for both LT4320/MOSFETs and also Schottky diodes, you could add CLC filtering as many don't because coils are expensive, you could choose footprint of a few modern regs so your designs a bit more universal, you could add various footprints of various sizes filter caps. Etc.
Cap multipliers are no substitute for regulators

Hi thank you very much again. I have decided to start with LTSpice. I understand that reality can be quite different from ideality ... but i believe in simulation a lot. I have been hugely impressed by the sim result of a little more than basic cap multiplier. I see that a very cheap tip142t with its hfe=1000 can be a great cap multiplier. And i could put another stage (before the cap multiplier) to regulate the voltage with a zener.
Years ago i was playing with a Bride of Zen preamp by Nelson Pass
Its power supply uses zeners followed by RC filtering (schema attached)
I think i could use the cap multiplier (CM) instead of the RC filter and get a greater psrr ? that CM will take care also of the noise from the zener diode ... hopefully
And the preamp was quite silent after all.
I would prefer the tip142t to the tip29 for its much higher hfe.
I would use just one zener ... even if i read that higher voltage zeners are noisier I have great expectations on the CM ... really great after seeing the results from sim Unbelievable psrr

[jean-paul]

Believing is done in church. I never simulated anything but built stuff and it got me there. Of course simulation can save time (which is the goal after all) but IMHO the knowledge should be in the head not in a program and experience is gained by doing stuff and measuring what you have created.

[PRR]

Originally Posted by jean-paul

Believing is done in church. ..... IMHO the knowledge should be in the head not in a program and experience is gained by doing stuff and measuring what you have created.

Amen.

[ginetto61]

Originally Posted by jean-paul

Believing is done in church.

Hi point taken ... i should have said trust. I trust math and i trust also measurements.

I never simulated anything but built stuff and it got me there. Of course simulation can save time (which is the goal after all)

this is not a little thing. In a matter of second you can change some parts value and see if the circuit is stable or not ... its distortion ... calculate its gain Zout ... for me stability is the very key ... i do not understand at all the concept of stability.
And this is a very big problem for me. I had a little integrated from Rotel and a decent headphone amp with too much Vgain ... i wanted to use the HP amp as line preamp so i tried to reduce the Rotel gain changing the FB resistor The amp was buzzing ... i put an hand on the heatsink and the sking remains attached ... what a pain ... the heatsink was unbelievably hot
I knew after that i had made the amp unstable ... and burned some parts with some smoke as well

but IMHO the knowledge should be in the head not in a program and experience is gained by doing stuff and measuring what you have created

i would leave a sw to do the rough job ... then of course the knowledge of the designer is fundamental to get a nice result