suggestion for proceeding with new paradigm for power supplies.

Tube DIY Asylum

Do It Yourself (DIY) paradise for tube and SET project builders.

OK, this one isn't as crazy as the tube regulator one.
Here is the idea: perhaps there is a way to keep R, L and C's low so that slew rate is kept really fast but ripple is eliminated, by playing with phase angles each step of the way through a three or four part LC network.
So, when (and if, frankly) I can get the time, I would construct a series of networks, each one FOLLOWING A SPECIFIC, DIFFERENT RULE, and measure ripple and slew rate for each, with the same load on each.
Notice I am not talking about computer simulation. The problem with computer simulation is that the software may or may not allow a new paradigm to be truly tested. Testing real life is the only way to find out. But God knows I could be wrong and anyone with a good simulation program by all means please play with this idea.
Now, my intuitive idea is this (here is where I get very afraid that BBECK or Mark Kelly will shred my idea with some fact I have either forgotten or never knew) - if the LC network keeps shifting the resonance point, the ripple keeps getting split into higher frequency fractions as it is also reduced in amplitude, and hopefully, before the presumed advantage of low DCR H C is lost, it becomes part of the low-level noise and, while it may still be a few mv's or tens of mv's higher than we'd like, at least it's not an obtrusive 120 Hz anymore.
So, the ratios between R, L and C keep changing as the LC network progresses, according to a preconceived "rule".
Here are examples of rules that might be tried (this is different from cutting and trying; the idea is to test a number of different ways of specifying ratios, and see if there is either a trend or perhaps one even hits the jackpot).
THESE ARE JUST IDEAS, I am not claiming any knowledge here; just pop-corning.
for the following, "s.r." means square root, I don't know how to get the square root symbol from the keyboard (for the heck of it I decided to go with four LC stages; if the chokes are hand-wound and the R kept really low, and the H and C's are also kept low, I think we can get away with four stages - maybe even five stages!):
1. The rule of progressive square root irrational ratios: the first LC is L/C = s.r.2; the second is L/C = s.r.3; the third jumps to L/C = s.r.5; the fourth is L/C = s.r.6.
2. The rule of alternating numerator/denominator progressive square root irrational ratios: the first LC is L/C = s.r.2; the second is L/C = 1/s.r.3; the third jumps to L/C = s.r.5; the fourth is L/C = 1/s.r.6.
3. The rule of same square root irrational ratio with simple multiplication of L and C each stage by 2, 3 and 4: the first LC is L/C = s.r.2; the second is L/C = s.r.3 but values for L and C are doubled; the third jumps to L/C = s.r.5 but the values for L and C are tripled; the fourth is L/C = s.r.6. but the values for L and C are quadrupled.
And so on. Many, many possibilities. Only way to implement a nutso idea like this is to wind up chokes to spec, and combine caps as needed to get an approximation of the right value. Or maybe computer simulation would work, who knows? God knows I don't.
Now, I'm just playing around, but isn't this at least a more systematically applicable idea than some vague notion of what Jeff and Dennis might be driving at and trying to put that into practice by cutting and trying? At least there are rules here to test and see; we might get a "scent" from one of them.
I used irrational square roots for the ratios from the classical discovery that the relationship between the diagonal and the side of a square are irrational, and thought perhaps that might apply nicely to introducing a "noise-pattern" to the breaking up of the ripple frequency. :)

2,3,5,6,7,8
1. Rule of irrational number phase shifting, each one alternating, progressing from 1/3 through
"It ain't a comeback until it's left the shop" Jimmy Dunne, the first man to drive a VW Beetle faster than 200 mph, and he has the forehead scar to prove it; I will always honor him for taking a chance on me when I wanted to be an engine mechanic.

This post is made possible by the generous support of people like you and our sponsors:

VH Audio

Topic - suggestion for proceeding with new paradigm for power supplies. - radiance 22:55:23 07/23/06 (3)

nobody's going to at least tell me I'm nuts? (nt) - radiance 18:42:01 07/24/06 (2)

Re: nobody's going to at least tell me I'm nuts? (nt) - mach1 23:47:11 07/24/06 (1)

Ah, thank you, I just realized from your post that I've been.... - radiance 00:48:27 07/25/06 (0)

You can not post to an archived thread.