JimL's shunt regulator PSU would be great too. 

One other thing to consider is the board dimension and its implication to chassis options and layout. 

Edited December 24, 201510 yr by mwl168

I just did the gerber files for my PS.  Also have a schematic but this is all in KiCAD, so not sure how to proceed from here.  The PS supply is the same as the schematic in the SRX revised thread except for the addition of an 0.1 uf/1000v cap on the input side between B+ and B-.  The PS does not include the bias supply but I used KG's design anyway and I believe that is available as a separate board.

Edited December 24, 201510 yr by JimL

If it´s possible, I would be interested in an amp (psu+amp) too.

On 2015-12-24 at 4:44 AM, JimL said:

so not sure how to proceed from here.

Well, I can install KiCAD and export something suitable for ordering PCBs and then pass on to mwl168 if that is the question?, or is there something else I can do to help? Would really like to try your PSU :), so just let me know how I can support...

I can put up jiml's gerber file on my server, or convert over to my software with a decent picture.

 

Well, adding the files to Kevin's server definitely makes most sense, so that I would suggest as plan A, with the other proposal as plan B.

So, I thought I had finished the PCB design, but there were a couple missing holes, and then the more I tried to correct things the worse it got, so I don't have working Gerber files at present.  GRRRR!  I'm traveling today, so will try again tonight.  If worst comes to worst, may have to start from the beginning.  Part of the problem is that I tried to incorporate a ground plane into the PCB, then have the mounting holes connect to the ground plane.

Edited December 27, 201510 yr by JimL

2 hours ago, JimL said:

So, I thought I had finished the PCB design, but there were a couple missing holes, and then the more I tried to correct things the worse it got, so I don't have working Gerber files at present.  GRRRR!  I'm traveling today, so will try again tonight.  If worst comes to worst, may have to start from the beginning.  Part of the problem is that I tried to incorporate a ground plane into the PCB, then have the mounting holes connect to the ground plane.

Jim, is this for the Amp or the PSU?  

Dr Gilmore designed a board for the amp, I'm trying to design a board for the PSU.

So just a quick update.  I am currently stuck in trying to design a PCB for the PSU.  I've got it mostly done but there is one connection that I haven't been able to make - I mean literally KiCAD won't let me connect two parts.  I've tried deleting the connection in the schematic and reconnecting it, and reloading it in the PCB design module, I've tried manually drawing it in the PCB design module - it just won't let me do it.  I'm going to try starting from scratch again this weekend.

Can you post the schematic file in Kicad? I have Kicad but haven't used it. I assume it has a sch and a pcb file similar to the way Eagle works? Maybe not.

I have nothing to do  for the next couple of days, so I could do the board. the one pass fet is not available at mouser, so something different is needed, need sizes of the caps and how much power (i.e. on board heatsinks, or aluminum bracket) etc

 

Pars, thanks for the offer, but if you haven't used KiCAD you're probably going to have at least as much trouble as I had.

 

Dr. Gilmore, Thanks you for the very generous offer.  Here are jpgs of the schematic, bottom copper with and without the component layout, and top copper.  I believe there is only one missing connection.  My layout uses an aluminum bracket because the MOSFETs should be able to withstand 35 watts or so if the amp boards go south and the entire current has to run through the shunt.  Most of the connections are on the bottom copper, with two vias to the top copper, and a couple ground connections on the top copper which could be replaced with a ground plane on top.  This is my first attempt at laying out a PCB so if you find a better arrangement feel free to change it.

 

The 10M90S and DN2540 are available on Mouser, and I just checked that the 19NM50N MOSFETs are also available at Mouser, as is the SPX431.  I chose those MOSFETs for their low on resistance, which is better than most, but other 500 volt rated MOSFETs can be used.

 

So the input electrolytic caps are 35 mm diameter.

The 0.1 uf/1000v caps (C3, C4 C9) are 27x11mm rectangular, 22.5 lead spacing

The 22 uf/450v caps (C7,C8) are D-Link 36.5x31.5 rectangular, 27.5mm lead spacing

 

C6 is 2.2 uf 50v

 

R12 and R18-22 are at least 1 watt, 500 volt, other resistors are 1/2 watt.

 

On ‎1‎/‎1‎/‎2016 at 8:03 PM, JimL said:

 

The 22 uf/450v caps (C7,C8) are D-Link 36.5x31.5 rectangular, 27.5mm lead spacing

31.5 ??  that's a really huge cap and the picture does not show that.

 

I moved a whole bunch of parts

 

 

 

 

Edited January 2, 201610 yr by kevin gilmore

Oops, one boo-boo.  The labels on the board for R16 and R17 should be exchanged, R16 should be to the left of R17..

take a look at the picture which I just updated

 

Yes, the board layout doesn't show it because KiCAD doesn't have a footprint that large.  But that's the specs of the TDK D-Link on the Mouser website.  As Donald Trump would say, they're UUGE!  On my point to point build I used Solen polypro caps which measure 2.3" long by 1.5" diameter, and they're UUGER!. That's the reason I left space on the board layout instead of bringing the surrounding resistors close to the cap footprint.  I see you added the bias board.

 

BTW, it occurred to me that with an amp using outboard heatsinks, you could use this as a mono PS, one for each channel, and between the amp and the PS the current draw should be absolutely stable, although it's pretty close in any case.

 

Also, the reason I didn't include the rectifiers on the board was in case someone wanted to use a choke input, in which case the rectifiers would have to come before the choke, so the transformer, rectifiers and chokes would be wired point to point and the raw B+/B- and transformer CT to ground would then go to the board.

 

Um, don't see the SPX431, protective zener and caps that are supposed to control the voltage.  Also the output block should be 6 pins rather than 7.

Edited January 1, 201610 yr by JimL

updated picture, everything is there.

need something for C- supply

board is 8.9 x 5.05

need to find hole sizes for the big caps

 

Ah ha!  The C- supply is just a dropping resistor.  I don't think it needs anything else since it goes to the input tail current source which should isolate it from noise and hum.

 

Guess that board is too big for two mono supplies. My original board without the input rectifier and bias supply was about 6.6x 3.6 IIRC, but I cut it pretty close.  As long as we're at it, do you want to have provision for normal bias headphones as well?

 

BTW, I like your layout of the SPX431 much better than mine, cleaner pathway.   Told you I was new at this.

 

Also, if anyone wants to use a off board rectifier and choke input they can just jumper two of the input rectifiers and leave the other two unpopulated.

Edited January 1, 201610 yr by JimL

Pardon my ignorance, but if the 22uF/450V caps are for bypassing purpose, wouldn't they ideally be located on the amp board where they are closest to the load?   

Edited January 1, 201610 yr by mwl168

Yes, bypass caps should be put as close to the amp power entry as possible, but the main reason I put the caps on the output is based on a comment by John Broskie, "My experience with shunt regulators has been that without some capacitance on the output, expect some oscillation."

Also, to go back to my previous comment about the SPX431 layout, I realized after some more thought that there was another missing connection between the output of Q4 and R12, which KG noticed and I didn't.

Edited January 2, 201610 yr by JimL

Thans JimL for the explanation. 

So it looks like Dr. Gilmore has posted the shunt reg PS board on his web site.  As I said it looks correct to me but someone else less involved should probably check it out to make sure.

Kevin and JimL: 

Any particular reason why the shunt regulator PS board is set up for center tapped secondary?

Thanks!

because that is how jiml gets away with half the amount of parts.

for the srx circuit it works because a low impedance ground is not necessary.

i myself would not do this, i would do the usual 2 supplies tied together. It only took me about 1 hour to do that layout, so no biggie.