Okay. I take your points. The ATX isn't helping me here and all an what I
really want is an AC-DC rectifier leading into a switching PSU circuit (giving slightly above my max DC volts, say 50 or 30) leading into a linear LDO regulated linear drop circuit controlled by the front pot. If this is to be believed anyway:
The key to a clean output with reasonable efficiency is a switch mode power supply on the front end and a linear regulator on the back end.
cleaner output and better regulation of a linear
BTW I assume you meant to say the 'back end' (i.e. directly connected to the AC rectifier then the wall) is the switch mode and the 'front' (with the DC volts out control pot and banana plug) is the linear... otherwise I completely misunderstand... but I think I'm right? Doesn't make sense (to me) the other way...
I meant with this not using a off-the-shelf voltage regulator but build one using some opamps and such. The basic circuit is very simple, but what I like about it is how much you can learn in going from a very basic zener diode with a transistor to a decent supply can be a path with a lot of "gotchas". Of course, if digital is what you want to go into, I can understand it not being a fun thing.
I do like the idea of building up my own completely custom supply in stages! This would allow me to understand the concepts of each stage I go through too, giving me a much deeper insight into the whole thing. It would totally be fun. I'm not only interested in digital.
So what I have now is two switch-modes connected inefficiently. The ATX is basically providing the rectifier and that's it. Eventually what I want to do is
build all of these components myself (well maybe not the
rectifier [/edit: I mean transformer] coils...) and put them in a custom enclosure.
Wall -> Transformer -> Rectifier -> Smoothing -> Home made switch mode (to get around 50V+) -> More smoothing -> Home made linear LDO with a ten turn pot (to get approx 1-50V adjustable). Opinions?
The way to "make" the ATX work for me is adjust the output voltage of the PSU 'internally' somehow and upgrade any components on the other side of that to handle the new DC voltage. Then I could just add the LDO circuit. Unfortunately I've got no idea how to do that, although I know its possible. Better to build my own. I'll learn more you're right.
The one someone like the OP is going to use is most likely going to be one of the "recycled" ones, not a brand new high quality one. So likely old design, dried out capacitors, etc.
Someone like the OP? I am a fully qualified nerd thank you very much
(software side anyway) and my PSUs are all good quality - wouldn't trust my work on a machine which has a higher than necessary chance of failure during use. This particular one is only out of my current number 3 PC (so approx 3 years old - I build or buy a PC or buy a laptop just about every year or so and move them down 'the stack of usability' when they are replaced) and is 500W+. It's heavy with nice caps (no bulge and I have measured ESR) and proper heat sinking and it does a great job. I bet you're surprised at the ripple... or maybe I will be... I'll let you know.
Anyway if the ATX is too noisy for use with whatever microcontroller I'll build something to smooth it. I'll learn something doing that. Or maybe the ATX should only be for high power jobs where my little adjustable wall warts won't cut it.
So what is the point of such exercise? To prove that it can be done?
Honestly - have you never heard of doing something just for the sake of it? For fun? To learn something? I do not fully understand power supplies and I intend to hack them and read about them until I do!
- Same approach I take with a new software system.
Also, actually I will have both current limit and adjustable outputs on my proposed design, if you read the thread.
Thanks all!