Class AB Power Supply Design
Ok well after asking lots of questions and looking around, I came up with a design with minimal noise and reasonable cost and size.
The design does use a CLC design, so if you feel there is good reason to avoid this, let me know. I've tried to read up on it, and from what I understand, lower voltage higher current supplies don't need as large a value inductor. I cam up with minimal of 5mh for a 45 volt at 10 amp supply, and I will be using a 10mh inductor. I know tube amps use 5h+ inductors, but thats at 300 volts or more. When I looked at the Class A SS amp supplies that used them, like from Nelson Pass, they used smaller 20mh inductors I believe, and I noticed that the Musical Fidelity with "Choke regulated" supplies used somethin like 5mh inductors, so I figure this is in the range. I think I read somewhere that inductors need at least a constant minimal current running through them to work effectivly, which is one reason they aren't commonly used in Class AB amps. However, with a higher bias Class AB amp like I'm using, I'm thinking it would be ok.
Anyway, the design will use IXYS 1200 volt 30 amp schottky diodes with .01mf MKT caps across them. In Nelson Pass's words, to slow them down. I tried this in others and liked the improvement, though it does make me question the use of Freds in the first place. Anyway, then it will go into 10,000uf's of capacitance made up of a single panasonic cap. Then a 10mh inductor with a .1 ohm resistance, and followed by 40,000mf of capacitance made up of 4-10,000uf caps. I think if I get the caps I want they will be rated for 80 volts or 100 volts and 105c, not sure though. I also might try adding some higher voltage Elna or Nichicon audio grade caps of smaller values to lower the esr. Say like a few 200-600uf caps on each side. Should have a similar effect to bypassing with small film caps, but over a broader range, no?
The transformer isn't picked yet, I was asking about that, but probably will be a Plitron 1000va tranny with 45 volt secondaries. However I'm really thinking of saving the money and getting a cheaper one from Parts Express or Ebay, I've had good luck that way in the past.
This power supply was showing between 100-200 microvolts of ripple with a 5k load and like 2.8 with a 20milliohm load, which also gave almost 8 amps across the load resistor. I figured that would equal close to a full out load. Could someone also explain why ripple goes up so much as load goes up? Thanks.