DIY Modular power supply

Skrog Productions

Hello , i've finally finished my diy power supply upgrade for the modular & learned a lot in the process about the balance of regulator supply voltage / current & heat , lots of heat , almost sausage cooking heat . Razz

I decided to try a dual bridge design a friend sketched for me and have my 12V regulators on each of the 6 bus bars / distros which has cured my heat problem i've had for a few months .
I ran a patched up test and left the modular running for 6 hours , mildly warm heat sinks , hurrah! .
Below is a diagram of how my supply is laid out and some pic of the unit, spray painted electric blue to skrog standards , heh Very Happy

.

Hope some of the info is useful to owners of rapidly expanding diy synths.

Dave.

Skrog power 2011.jpg

Description:

Filesize:

379.29 KB

Viewed:

162 Time(s)

This image has been reduced to fit the page. Click on it to enlarge.

blue power.JPG

Description:

Filesize:

75.22 KB

Viewed:

66 Time(s)

This image has been reduced to fit the page. Click on it to enlarge.

power guts.JPG

Description:

Filesize:

188.18 KB

Viewed:

71 Time(s)

This image has been reduced to fit the page. Click on it to enlarge.

working power.JPG

Description:

Filesize:

122.79 KB

Viewed:

76 Time(s)

This image has been reduced to fit the page. Click on it to enlarge.

Cynosure · Posted: Wed Jun 08, 2011 12:33 pm Post subject:

Awesome stuff. I will be making something similar in the future. Very versatile.

kjackman · Joined: Sep 05, 2010 Posts: 69 Location: Utah, USA

Very nice.

sompost · Joined: Aug 17, 2010 Posts: 24 Location: Switzerland

The toroidal transformer isn't really a 15V one, is it? From your numbers it rather looks like a 18V one. The 20V output slightly puzzles me.

Cheers

sompost · Joined: Aug 17, 2010 Posts: 24 Location: Switzerland

Never mind my comment above. I can see that it's a 15V transformer (but your numbers still puzzle me).

In your design you split the to paths (red-black/orange-yellow) and rectify/smooth them separately. Can you do that? I though that you're supposed to combine them immediately (red-orange) to get a center tap, and then go on from there. Both Thomas Henry and Ken Stone do that in their designs (and they use 18V transformers to get 15V regulated power).

In my calculations I never got more than +/-19V from a 15V transformer, and some commenters doubted that I actually get as much (they suggested accounting for 10% line variation). That would be too close to the required threshold for the 7815. You get +/-20V which looks a lot better.

I believe your measurments. And I am confused.

Skrog Productions · Posted: Fri Jun 10, 2011 8:50 am Post subject:

Hi sompost.

I believe there is no huge difference between having a one rectifier or dual rectifier setup , i just tried this design as it was new to me as i usually have just one rectifier in my other power supply's , i did read somewhere that any AC frequency discrepancies between a two secondary transformer system are settled using two bridges.
The 20V output after the bridges baffles me as well , all my transformers i have bought so far (toroidals) always output a couple of volts more than stated on the side sticker (under load) , and once through a rectifier seem to jump up another couple of volts , perhaps it may be the two sets of 4700uF caps maintaining the peak input cycle , i don't know .

I usually have an indicator led per supply side to act as a minimal load for when there is no modules connected (very rare) and have measured 19.8V , could be the electricity in Scotland Smile

Dave

sompost · Joined: Aug 17, 2010 Posts: 24 Location: Switzerland

If your Scottish electricity is only half as good as your single malts, then your measurements make total sense! Laughing

Cheers, and very nice build I might add. Do you have any pictures of the modules seen in the background?

slacker · Posted: Fri Jun 10, 2011 11:43 am Post subject:

On paper, 20 volts after the rectifiers is about right 15 x 1.414 = 21.21 minus a couple of diode drops for the rectifier gets you about 20 volts.
Looks like in this case the transformer has less load than it's rated for so it's really giving you 17 volts AC to start with.

Skrog Productions · Posted: Fri Jun 10, 2011 12:59 pm Post subject:

Yes , pics & sounds of the diy modules can be seen by clicking on the blue number 59 under my avatar to the left of here.

Room left for 4 last modules ......
http://electro-music.com/forum/phpbb-files/karplus_strong_track_patch_up_202.jpg

Dave Smile

tazer · Joined: Oct 18, 2010 Posts: 49 Location: UK

bloody hell that looks a tasty setup, do you need ladders to get up to the top?

diablojoy · Posted: Sun Jun 12, 2011 10:03 pm Post subject:

not sure if i am reading this correct
you are running 6 x 7812/7912 in parallel ? with 10 paralleled outputs off each so 60 x 3 pin outputs in total at -/+12volts should be good for five or six amps total each side roughly heatsink dependent
would like to see a schematic of that section as
normally there are issues with power sharing in such arrangements
and very low ohm resistive paths are required to balance them
but i think you have done something else here perhaps
I am interested as i have to do something soon as I need to make some supplies capable of
+/-15volts at 5 amps or better

Skrog Productions · Posted: Mon Jun 13, 2011 10:04 am Post subject:

Hi diablojoy.

My total load, allowing for 4 more modules, is about 2 Amps per side DC current before the regulators , i chose to make 6 busbars so each set of +/- 12V regs were drawing about 300mA per side which solves my "more heat with less regulators" problem i had before & i can use smaller heat sinks.
Although there are 10 sockets per busbar , i don't use them all, i think about 42 of them.
All the busbars are wired back to the main psu box individually using 16Amp rated water heater electrical cable flex which sorts any volt drop issues Very Happy

.
This is not a true Star ground system i have but , each of the busbars 0V comes direct back to the psu caps & 0V tie point as on the drawing at the top of the thread .

Here's a diagram of how my distro / busbar boards are fed.

Obviously if you are building a 5 Amp per side system supply you would need a transformer far larger than the one i use , may be 400VA to allow a bit of headroom for the transformer.

Thanks for looking Smile

Dave.

power distro.jpg

Description:

Filesize:

87.88 KB

Viewed:

67 Time(s)

This image has been reduced to fit the page. Click on it to enlarge.

diablojoy · Posted: Mon Jun 13, 2011 2:08 pm Post subject:

ahh thats so cool and far easier than my first idea of using pass transistors
thanks
now to find some nice large toroidals Smile

diablojoy · Posted: Thu Jul 21, 2011 7:18 pm Post subject:

didn't manage to find anything larger than 18v -0 - 18v 300VA
toroidal
but that should give a max of 8.3 amps per secondary winding so there is some leeway as I reworked my plan down to 40 modules per transformer and 5 modules per regulator board so it should come in OK allowing an average of approx 50mA per side per module - 4 Amps total per side
as i plan to build my cabinets at 40 modules each this works out very well
have drawn up a regulator board design for pnp blue
for the bridge rectifiers and the main supply caps
[ thinking 4 x 15000uf chassis mount caps ] and the transformer
will go for point to point wiring in an ally enclosure which i have already started to make for this.
one thing I am not certain of yet is if i need to allow for some heat sinking for the bridge rectifiers ? maybe mounting directly to the case will be OK
Dave do your rectifiers get hot at all ?

Skrog Productions · Posted: Fri Jul 22, 2011 8:57 am Post subject:

Hi , on a previous design , my rectifiers were very hot , the new 15A ones run quite cool , they are bolted to some alum "L" section then they are bolted to the case.

Watch out for inrush current ,on the pictured supply above , in the first second of switch on , the secondarys go up to 4Amps per side then settle back to normal 2Amps , my 3.15A anti surge fuses last a month , i will have to get some 4Amp slow-blowers , i think i have around 150 (maybe more) electrolytic caps in total which charge up on switch on Smile

.

Is your transformer 8.3Amps max (300 / 36 = 8.3) .

Think big when using jumper wires on the rectifier pcb , i used 2.5mm sq ring circuit solid copper wire as wire links & make sure any protection diodes are rated well above your normal current draw on the same pcb & big juicy etched tracks for the supply tracks Very Happy

good luck with your power design , is it all 15v stuff you have ????
Dave.

frijitz · Posted: Fri Jul 22, 2011 9:19 am Post subject:

Skrog Productions · Posted: Fri Jul 22, 2011 9:43 am Post subject:

Something Blue Very Happy

chaos & 5pulser & multiples.JPG

Description:

Filesize:

68.52 KB

Viewed:

66 Time(s)

This image has been reduced to fit the page. Click on it to enlarge.

diablojoy · Posted: Fri Jul 22, 2011 9:48 pm Post subject:

frijitz · Posted: Sat Jul 23, 2011 1:15 pm Post subject:

diablojoy · Posted: Sat Jul 23, 2011 6:54 pm Post subject:

Skrog Productions · Posted: Sun Jul 24, 2011 1:39 am Post subject:

I just had a big pile of 78 / 79 's left from when Rapid uk had the non-RHoS sale 5 years ago , use up what you have , (typical tight Scotsman he he Smile

)

Inventor · Posted: Sun Jul 24, 2011 2:32 am Post subject:

All this talk of hot rectifiers reminds me of my work with FET Rectifiers. Seeing this as a major problem worth solving, I created some designs for rectifiers that used Low Rdson FETs such as HEXFETs instead of diodes. I was able to show in a prototype and in concept that they were more efficient than diodes. I even made a full wave bridge out of HEXFETs.

The bridge worked fine except it chirped backwards during zero crossings, which I am certain is a solvable issue. Now, I don't want to risk anyone's precious modulars on such an experimental technique, but I do want to make folks aware that this technology exists.

Les

frijitz · Posted: Sun Jul 24, 2011 11:36 am Post subject:

prgdeltablues · Joined: Sep 25, 2006 Posts: 141 Location: UK Audio files: 8

Outside my competence zone, but this article by Rod Elliott might be useful for addressing the inrush current problem:

http://sound.westhost.com/project39.htm

Peter

diablojoy · Posted: Mon Jul 25, 2011 10:08 pm Post subject:

thanks that is a very good link and very useful too
had only been considering inrush current in terms of the modules connected to the supply not of the supply itself
4 x 15000uf chassis mount caps + 32 x 4700uf mounted across 8 regulator boards + 300VA transformer windings, it does all add up.
will need to study that page some more when i get time.
had been studying the CGS supply delay cct for idea's
the way ken has it looks simple enough, though for me i would change some things so my 8 reg boards came on line sequentially as well as the reg outputs which may make it somewhat more complex. i have to determine min loads first ie will a couple of indicator leds be enough load for each reg board for the short duration between start up and the load switching in ? is it even an issue ?
still more learning to do Very Happy