Power Supply Unit

I added a full-wave rectifier and found the diodes dissipate maybe <30W. Which is a lot.

Try a current source on the zener, it may allow you to ditch the 2mF cap.

A current source was my first attempt, it works but the ripple rejection was a little lower. It's an either-or situation, so whatever suits you best it's as good as the other. P.S. A diode bridge would dissipate a lot less than 30W. At a 3.3A consumption the dissipation per diode would be 2.3W with a total for the 4 diodes of 9W. If you use germanium diodes the total dissipation will be around 3W.

Why this extreme ripple rejection, if a little change of drawn power cause a 1000times higher voltage drop? So the powered circuit has to handle the drop/RR anyway by itself. No need for 122dB RR here in this supply concept. And this RR is for the price of extreme slow startup time >>5s! One more very unrealistic thing, your 37V has zero impedance. Place at least a more realistic internal resistance to it like 1Ohm which still gives 37A short circuit current!! You do have such a voltage source available? I don't have it.

5S start up time is not bad, i don't mind waiting 5s :D. Yes, this doesn't have short circuit protection, but as i said, this is probably most suitable for low power but high ripple sensitive circuits, like preamps and oscillators.

I updated my current source example, it would start fast (I did not even think of that as an issue I just thought how much I pay $$$ for one of those capacitors). http://everycircuit.com/circuit/6713707085168640

20 cents per cap, that's not an issue. It occupies the same space as a constant current source, and the delay is not an issue. I mean come on 5s shouldn't be an issue, moreover some preamps need a soft start before the power stages stabilise.

What is this power supply now for? A preamp or a 100W load as you have in your circuit? And i don't talk about a missing short protection. J just talked about a realisitic "internal resistance" of a power source! Again why 122dB, if the regulator easy drops a few 100mV by change of load?

Oh, the UPW1J222MHD I use cost 2.5$, what is your part #

>>37V (63V?) 2.2mF for 20cent? Cool, i'd like to have 100 of them, were?

http://www.elimex.bg/product/7537/c2200m-35vvn.html there you go. A 35V 2200uF cap for 60ст (30 cents). These caps are 20% voltage tolerant so i don't see why you'd buy a 63V cap, but hey the 63V cap is 50 cents, so whatever suits you best.

And hurz, if you bothered to read a few comments above, you would've noticed, that i suggested this mostly for preamps and oscillators where the draw won't be sufficient to cause and ripple, and the full -122dB can be obtained.

P.s. The website i sent you, is known to be somewhat pricy. I bet i can find the 63V version for 30 cents and the 35V version for 20cents. I know a few very cheap shops near my neighborhood, but i don't think they have a website.

Samyoung SHL series capacitors. Ok thanks, I'll look around and see if I can get them from one of the vendors I have experience with. Though looking at the data sheet, I think the 35V working voltage should be taken as a no joke sort of limit. https://www.samyoung.co.kr/download/miniature/SHL.pdf

I've had those same caps working flawlessly on 40V, plus that cap on the resistor divider won't need to handle enormous amounts of current like a Smoothing capacitor will, it only works with 20mA give or take, so it won't be stressed so much. Trust me, most electrolytic capacitors won't notice such a small voltage tolerance. Some might have a problem, i think old Russian capacitors are quite intolerable, but most of these new ones have a 20% tolerance.

Anyway, you could definitely use a constant current source as a substitute, i just decided to use this method specifically for the delay. As i said, sometimes it's important to delay the start-up of the preamplifiers, until all the modes have had time to set up correctly in the power stages of an amplifier.

I don't trust engineers who use tolerances for their bad designs. You have to give a tolerance and not take it for you from the cap. What you do is not engineering its suicide. Electrolyt caps are very sensitive against overvoltage doesn't matter how much current you draw from them. Static overvoltage does kill them quickly. If not in hours then in days or weeks. This is very bad design! Anyway, from 20 cent we are already at 60 for a not working 35V version and for 63V ...... As thumb of rule its recommended to double the cap voltage!!! So absolute no reason to trust you in the past and after this suicide instruction i will not trust you in the future!

BTW, the tolerance is given to the capacitance, but manufactures never give a tolerance for the voltage of cap. They fix it for good reasons. Capacitance could have a high tolerance and even unsymmetrical like +25%. Double the voltage and you are safe, everything else will shorten the lifetime dramatically. Anyway, ultra cheap electrolyt caps are made for only 2years operating time.... Spend some more money to have long fun with your circuits.

Look, i told you. I had 35V caps working at 40V for years and i haven't had any problem. From an ideal point of view, you should always go for the higher value, but from a practical point of view 3-4V won't kill it.

P.s. I sent you a link for a 30cent 35V cap. How did you get to 60 hahaha. Moreover i also said that the website i sent you is known to be pricy, i bet i can find the 63V version for no more than 40 cents in the local component shop. I know a few quite cheap. I bought a few 47uF/450V caps for exactly 50cents each for my vacuum tube amplifier. I don't know where you buy your stuff, but i recommend you find a different retailer. ;)