The MOST POWERFULL 12V PSU EVER...

[SafeFire]

Incredible, simple incredible. I'm amazed by your pure curage and knowledge to build it. Totally subscribed

By the way. The irony of an earlier post from this topic dawns on me.
The fact that I feel encouraged to also listen better in on my physics lessons from now on, clashes with the fact That I am sitting in a physics class right now, reading XS on my laptop insted of listening in. Oh the irony

[alexio]

Important thing to remember Alexio is that these specs are for with the PSU will be capable of not what they'll be running 24/7. Of course there's not going to be an occasion that the many of these lines will ever draw the current George is designing it for. But, if you can do it, then do it! Overspecing will in fact allow for a cooler design as the power circuits will be under less "stress" to provide power requirements that are closer to their maximum, so will produce less waste heat.

True to some point, but remember that maximum efficiency is achieved at 70-80% of the maximum load on consumer PSU's. I don't know George's design of the PSU ofcourse. You are probable right that less current per components + more components ( more surface area) will result in a lower operating temperature.

1.8kW do it George

[Raja@ASUS]

George will you use a capacitance multiplier?

i have an active amp psu I built from a kit here, basiaclly an rc filter feeding a mosfet stage. Virtually no ripple at 72v output and upto 10amp current draw. Circuit ofcourse is ip of a friend of mine called Dr David White.

You should get far better overclock ability using a conventional linear supply.


There would definatley be a market for a linear psu for the cpu, I think a good supply using a mosfet output stage and real low ripple would give a 100-200mhz advantage in overclocking over conventional switchmode pus's we use.

I am very interested in your results buddy..


regards
Raja

[Pipi]

Looks like i'll have to go to Greece to holiday next year

[celemine1Gig]

...
You should get far better overclock ability using a conventional linear supply.

There would definatley be a market for a linear psu for the cpu, I think a good supply using a mosfet output stage and real low ripple would give a 100-200mhz advantage in overclocking over conventional switchmode pus's we use.
...

I agree with you that a linear design will be nice, but I really doubt that you'll get a 100-200MHz better OC just by switching the PSU. The switching VCore supply on the mainboard still remains.

[esdee]

I agree with you that a linear design will be nice, but I really doubt that you'll get a 100-200MHz better OC just by switching the PSU. The switching VCore supply on the mainboard still remains.

can you guess someone who's lunatic enough to override the Vcore switching supply with even lets say ... a maximizer?

[Revv23]

can you guess someone who's lunatic enough to override the Vcore switching supply with even lets say ... a maximizer?

haha, he almost did that on the dothan!!!

I thought the maxi could only outout 115w though? sureley not enough for CPU.

[nn_step]

HOW many rails?

[Moonman]

As tittle says....

It will NOT be with 4 separate 12V rails BUT with ONE 12V and all the cables will be feeded in paraller from that rail......

One rail to rule them all.

[STEvil]

well crud, and here I was thinking i'd be the smart one by strapping a pair of 60A @ 12v PSU's together.. then I see this lol

[hipro5]

Thanks guys for encouraging me....Though I don't have a lot of spare time, I'll try to do it as complicated as I can....

2 x 8x8cm DELTA fans high output cfms will be used for the heatsinks in puss-pull config....
They will be thermally controlled though....As the temp of the heatsinks goes up, the rpms of the fans will rise automatically and visa-versa.....

Point is, I'm on a dilemma right now about the 12V feed back circuit......Should I get feedback reading from the CPU's slot from the VGA's slot OR should I use them ALL....
"Feedback reading" for those who don't know, is a reading that is been taken from the VERY END of a cable line and by this, the voltage stabilizer circuit "repairs/fixes" it's output voltage so as to be "stabilized" at the voltage we have peaked....

I can chose the very end cable line of the cpu's feeding or vga's feeding OR - more complicated though - ALL of the 12V readings (cpu's, 2 x vga's, etc) and then add the readings and devided them /3 ....Then take this "mixed" reading and with this, control the 12V stabilized output....OR no need for all this shi(f)t....

[STEvil]

why not take the reading at the end of the (PSU) PCB where the cables/wires attach?

If you have a droop situation at one location and not another and the PSU compensates for this then you will have overvoltage at one point.

With thick enough cables and good enough connectors there wouldnt be any issues unless the PSU itself couldnt hold the load.

[hipro5]

why not take the reading at the end of the (PSU) PCB where the cables/wires attach?

If you have a droop situation at one location and not another and the PSU compensates for this then you will have overvoltage at one point.

With thick enough cables and good enough connectors there wouldnt be any issues unless the PSU itself couldnt hold the load.

Because of the long cables, you HAVE to take the feedback reading at the very end of them so as to be more accurate of your output voltage even if they are thick enough - There's always a droop in them....

[STEvil]

Yes, but you dont want to overcompensate for something which will be little or no problem given use of the proper cables.

[Johnny Bravo]

I'd say measure from VGA as this will be the larger of the current draws -> most likely to droop. Mixing the feedback may not be such a good idea as what usually happens is the feedback circuit only partly compensates for a voltage droop and none of the end supplies are fully corrected.

[zakelwe]

Although possibly not likely nvidia and Ati are both now putting out video cards and motherboards which could theoretically have 3 SLi 'd XF'd ..so I would have 6 Pcie cables on the psu. Not sure how that would effect your calulations for amps / power needed either ?

Of course the 3rd slot and double Sli connectors might be a for a gpu based physics card which needs not much at all

Best to be prepared though, just in case.

Regards

Andy

[heikis]

will this beast fit in an ATX-size case? how huge will it be :P ?

[esdee]

haha, he almost did that on the dothan!!!

well he actually did it on the dothans board NB ...

[prosk8]

give the fhoto man

[nn_step]

Be sure to keep the ripple down man, otherwise it'll be harder to get world record overclocks

[marauder16]

give the fhoto man

he broke his camera, so no pics for some time...

can't wait to see if this psu 'll help in OC reults

[blossa]

Looks like i'll have to go to Greece to holiday next year

Do not forget to bring batteries.

[Asazman]

haha, makes my PCP&C look like a toy

[prosk8]

HIPRO GET A camera and fast man








or you can send like hamm 8 your supply to me and i do the fhoto and extreme overclock

[Demo]

would this great invention beat out those PCP&C 1kwatt PSU ?