Hi all, I've followed Dave for years now, and I've been reading up on the forum for a few weeks now trying to figure out a nice power supply design for myself.
Given that I have never designed anything electronic in my life, there were a few moments where I looked at my design and said to myself "WTF did I do this for?" and many other things like that. BUT I think it's about ready to go.
My design requirements were:
-Linear
-Two positive 1.25V-16V outputs with independent current/voltage adjustment
-AC Mains voltage input (120VAC @60Hz)
-Modular (each functional group would be build on their own boards, so I can re-design at a later time or re-purpose)
-Beefy external heatsink to save me putting in a fan circuit.
I'm pretty well ready to purchase my parts, but I'm still out on the fence about a transformer. Given I have ~4.5V drop from my LM317s and rectifiers on each rail. Any suggestions? And any pointers or recommendations about my current design?
Thanks all!
(Attachment Link)
First of all, nice and clear circuit diagram.

I have a question, Do you have any design requirement about output current? Or are you going up to the maximum output current of the LM317 (1.5A not considering power dissipation)?
My biggest gripe with your current circuit is the potentiometer in the current circuit. The LM317 works such, that it always tries to get 1.25V between the adjust and output pin (i.e. it drives the output to get 1.25V drop towards the adjust pin). This means that there is always 1.25V voltage drop over that potentiometer. Power is calculated as P = U * I, so with the constant 1.25V, at 1.5A, it would give a nice and toasty 1.875W of power in the resistor! Potentiometers are usually limited to a couple of hundred milliwatts of power, and it is usually not a good idea to run any substantial current through them. I have seen this particular circuit floating around on the internet before, but I do not think it is a very good one. This circuit would be okay though, I think up to 80 or so mA.
The problem is, having a current limit requires a bit more complex circuitry. Here is an example:
https://www.circuitsonline.net/downloads/schakelingen/lm317_labvoeding_v2_b.pdfNote that the P1 current adjustment potentiometer is not in the direct power path.
Anyways, your output voltage adjustment section seems quite good. The only remark I have is that it might be hard to find 3K potentiometers. Generally standard values are 1, 2 and 5 in multiples of 10 (i.e. 1k, 2k, 5k, 10k etc).
Don't be discouraged though! It is a lot better than my first circuit (and I am sure for a lot of others too!). If you have a breadboard, you can just try and build it and see what happens. Even if it fails in some way, try to figure out why it failed and improve on it. That is the best way to learn, and you will learn fast!