before I build this LM317/337 power supply, a few questions?

[I_may_be_drunk_right_now]

I don't think you've found too many power supplies that use a full-wave rectifier with a center-tapped transformer. Study it out, only two diodes are ever working here, and your output ripple is 60Hz, or 50Hz ... The other two diodes are just reverse biased all the time.

I mean if you're using that part because it's all you have, great, but you should understand what is going on here.

Perhaps you are referring to the classic voltage doubler/rectifier circuit that feeds two stacked capacitors on the input of old switching supplies?

Or maybe two separate windings with each its own full-wave rectifier?

I do think you're drunk   It's two separate full wave rectifier circuits, with a common ground.  Perfectly OK for the OP's needs.

Please show me what happens here. It's a simple half wave circuit with extra useless diodes. Draw out the currents and voltages on each half cycle of the CT output.

Show me where I'm wrong, please.

[I_may_be_drunk_right_now]

I don't think you've found too many power supplies that use a full-wave rectifier with a center-tapped transformer. Study it out, only two diodes are ever working here, and your output ripple is 60Hz, or 50Hz ... The other two diodes are just reverse biased all the time.

I mean if you're using that part because it's all you have, great, but you should understand what is going on here.

Perhaps you are referring to the classic voltage doubler/rectifier circuit that feeds two stacked capacitors on the input of old switching supplies?

Or maybe two separate windings with each its own full-wave rectifier?

What I've basically built is a slight variation of various other designs.
Like this
http://cdn.head-fi.org/e/e4/900x900px-LL-e431d7c1_317_337Schematic.png

Or this... (but with a purpose built bridge rectifier instead of four individual diodes)
http://cdn.head-fi.org/0/08/900x900px-LL-08345efb_SP-2-LMSCH.jpeg

or this
http://sound.westhost.com/p44-fig1.gif

It seems to work fine.  I've checked the outputs of the + and - pins of the rectifier with my scope (before adding the smoothing caps) and it looks exactly like the examples I've found.
The + pin gives me exactly what this attachment looks like and the - pin is just the mirror image.

Of course it would, but at what frequency? Or is your schematic wrong?

You're just copying stuff. There is only two diodes ever working here.

[I_may_be_drunk_right_now]

I don't think you've found too many power supplies that use a full-wave rectifier with a center-tapped transformer. Study it out, only two diodes are ever working here, and your output ripple is 60Hz, or 50Hz ... The other two diodes are just reverse biased all the time.

I mean if you're using that part because it's all you have, great, but you should understand what is going on here.

Perhaps you are referring to the classic voltage doubler/rectifier circuit that feeds two stacked capacitors on the input of old switching supplies?

Or maybe two separate windings with each its own full-wave rectifier?

What I've basically built is a slight variation of various other designs.
Like this
http://cdn.head-fi.org/e/e4/900x900px-LL-e431d7c1_317_337Schematic.png

Or this... (but with a purpose built bridge rectifier instead of four individual diodes)
http://cdn.head-fi.org/0/08/900x900px-LL-08345efb_SP-2-LMSCH.jpeg

or this
http://sound.westhost.com/p44-fig1.gif

It seems to work fine.  I've checked the outputs of the + and - pins of the rectifier with my scope (before adding the smoothing caps) and it looks exactly like the examples I've found.
The + pin gives me exactly what this attachment looks like and the - pin is just the mirror image.

Of course it would, but at what frequency? Or is your schematic wrong?

You're just copying stuff. There is only two diodes ever working here.

http://physics.stackexchange.com/questions/55698/what-is-the-role-of-center-tapping-in-a-full-wave-rectifier

It's very simple. A CT with two diodes pointing the same way to one load is a full-wave bridge.

Turn one of the diodes around and add another cap, you have a "voltage doubler" (compared to the same full wave CT configuration) with a tap halfway. You can add all the diodes you want, they'll be reversed biased.

Unless you actually have a transformer with two *separate* windings, each with a full-wave bridge, with the common *AFTER* the bridges.

Two wildly different configurations.

[edavid]

Drunk, you are just confusing yourself.  It is not a half wave circuit or a doubler.  It is 2 full wave circuits.  All the diodes conduct, and the ripple is 120Hz.  Fire up LTspice and see for yourself.

[dentaku]

edavid is right.
I've done it in real life and in simulations and it works.
Just like this.