power supply porn

[Voodoochile]

Now that I'm almost finished this PS, I though it fitting to put up some pics. This is my smoothest/quietest PS to date... I keep meaning to post some pics of the smaller single and double units.

This one will end up in a 1-unit rack chassis (13"x8", without the ears) alongside an amp (yet to be named, but some bastardization of the Apheared #47). Possibly evolving into a preamp. On to the ****:

The whole nut.

Hexfred bridge detail.

Rails under load, at uncomfortable volume.

More details- The tranny is a Talema 2A dual 12v sec, which nets an input v of 17v per rail (12v x 1.41 = 17v) after the bridge. The bridge is full-wave, built from eight HFA08TB60 hexfreds, bypassed with .33uF polyester box caps. Regulators are 317/337s, with a 1k feedback resistor, and 25-turn/20k bourns trimpots on the adjust pins. Input pins are bypassed with .1uF tantalum. Both regs are protected with diodes in the event of sudden collapse/reverse discharge of the caps.

The other caps are polyester philips and polypro wimas. I had a 1,000uF panasonic between adjust and ground (again with a diode to discharge to +/- rail) to slow down the turn-on, but it was too slow, so it's out now, I'm going to put in maybe a 220uF cap instead. All electrolytics are panasonic FC, a pair of 10,000uF before and after the regs.

For testing purposes, output is trimmed to 14v, but my intention is to run it at 15.5v. It could deliver 1 amp, but will most likely end up delivering something more like 200ma, as the amplifier is to be buffered with either eight buf634s, or four ha5002s. It is silent, and shows approximately 9mv ripple.

I'll put up some other pics someday when I get this cased up with the amp. Thanks for checking this out... kind of geekish to get worked up over a PS, but I do

.

 

[Wodgy]

This looks really cool. Do you have a schematic? Also, who manufactures the 317/337 regulators?

 

[Voodoochile]

Thanks, Wodgy-

Right now, my schematic is on paper, but this may help. The bridge is basically following Benny Jørgensen's on Headwize. From the regs on out, it is similar to Sheldon Stokes' PS. R1 and R3 in Sheldon's have been replaced with the bourns pots, and R2 and R4 are changed to 1k.

The 337 and 317 are available from both TI and National (and many others, I'm sure) I happen to be using National pieces.

 

[da_burl]

Very nice work! I take it you designed and produced your own PCB, or is it just hard wired underneath?

8 Hexfred's in a bridge? Dual mono??

You could also go even further out and use some inductors here and there, on the xfrmr secondary at least.

In my humble opinion, that supply is worthy of more than a 47, I think it would be great for a 42 or even a Gilmore!

I've got some MUR810's, and am going to bypass them with .01 SOD's, ala an old "text" power supply Apeared was talking about for my attempt at something similar to this. Like the idea of the 25 turn pot instead of trying to pick resistor pairs. Sometimes I feel a little safer using a lower voltage when first powering a new build up!

 

[Voodoochile]

Yes, a seperate fw bridge for each secondary... then neg of upper half to g, and pos of lower half to g.

The PS isn't done yet, just testing and tweaking it out. There is a choke at the mains inlet already, and the plot at the rails looks decent enough that I probably won't do too much more out there.

Yes, the pots are nice! I used a pair of single-turn tiny pots on another mini-dual supply, and they were very hard to adjust. These would be fine if they were 10-turn, but they happened to be 25-turn. It takes nearly an 1/8 turn to effect a .01 volt change.

It's definitely overkill, but I've been wanting to do this for a while now, and I figure it will be good for many different amps, should I want to change things around. With a tranny swap, it can push +/- 25v easily. At +/- 15v, with the 2v overhead, the regs barely warm up.

The board is Vector, with 3-pad per hole, and a bus every row.

 

[Voodoochile]

Quote:

I had a 1,000uF panasonic between adjust and ground (again with a diode to discharge to +/- rail) to slow down the turn-on, but it was too slow, so it's out now, I'm going to put in maybe a 220uF cap instead.

I just soldered in a pair of 100uF nichions where the 1k caps were. Now the output rises from 0-14v in about six seconds. With the 1k caps, it was more like 30 seconds. With no cap at all, slam-its-on-now.

 

[da_burl]

I suppose if I had looked close enough, I could have seen the numerous holes!! My last PS had a small section of 3pad/hole protoboard, and the rest was big holed waferboard or whatever its called (no copper, just holes). I found that easier for the larger caps, inductors, diodes, etc. I am considering just hot glueing the big caps upside down in the bottom of the case this time, easier to swap them in and out!

I am starting on my new supply right after the prototype "not_a_meta42" is finished (hopefully this weekend, or if I'm real lucky, maybe tonight!!). Then I will compare sound with the old and new.
D#^n, with all this talk of millivolts of ripple, you are gonna' make me bust out my Tek O'Scope!! Haven't used one in years, but I think I still remember how

As I told Cap'n Bubba, IF it works, and IF it looks half-assed, I will post some pics too...

db out

 

[guzzler]

what are all those TO220 packages on the left hand side of the board in the first picture?? nice specs as well by the way!!

g

 

[Wodgy]

Those are the eight HFA08TB60 hexfreds if I'm not mistaken...

 

[Voodoochile]

Yes, they are hexfreds. One bridge for each secondary. Very fast, but also a soft recovery, not noisy at all.

By the way, this fits like a dream in the left-side of the 1 unit chassis I posted about

.

More pics to come later.