WTB - Watercooled Power Supply
 

As the topic states, I am looking to buy a Watercooled Power supply, or a Waterblock that will work on a specific power supply. Either Enermax or one of the new Vantec 520a's. Looking to utilize 1/2" id hose.

Thank you,

Brian Watson

you can just epoxy a regular block onto the hs(s) in your psu...

Interesting idea. Not alot of space to work with though. If anyone has tried this it would be cool to see. I'll put up a post in the liquid cooling board.

I would, however, still like to purchase one if one is available.

Brian W

I thought Koolance did..... www.koolance.com I don't know if they still sell one I guess I could look....:rolleyes:

Didn't find anything on their site. I will keep looking. I have found a pretty good tutorial on making one. Maybe I will try if I can not find one made.

Brian W

Bladerunner of zerofanzone built one.

Im not sure but i think in some power supplies the internal heatsinks actually have LIVE power conducting through them... i dont know if this is true or not... i thought i saw one like that in an old power supply

Yes they do. That is why u use a ceramic plate. It conducts heat good enuff and insulates the power as well.

I have read this somewhere, but do not remember where.

Brian W

bladerunner didn't have any ceramic plate on this wc'ed psu..
i don't remember him doing anything to insulate the wb

no he didnt and i know why, i read his psu mod after i posted that

you know why?????


well spit it out man...
you can just say you know why and not explain it... :(

The heatsink was live, but was not electrically connected to the MOSFets. He had 2 choices:
1-electrically connect the waterblock so that it's live too
or
2-don't connect the waterblock electrically to mains

He opted for #2. Here's why: If the block had been connected, he would have had to isolate the MOSFETs electrically from the waterblock, which would have involved using either a Mica or Ceramic shim. Now of course that would greatly reduce the cooling ability, so that he didn't have to use any shims...

As he put it: "Ideally I wanted the metal back part of each mosfet directly over the block channel path for the best cooling. "

Re-reading, it looks like he opted to use the mica shims anyways. Probably because the waterblock table was so close to the transformer.:shrug:

Or... maybe it was because the block fittings would touch the case side anyways?:shrug:

dont connect?


from wut i saw that wb was screwed directly to the mosfets........................................... .................................................. .............................

Hi I can clear it up. :)

Originally the heatsinks were live as they were connected to live parts of the PCB, (soldered)

All I did in simple terms was not solder the water block to the PCB or any live point, but earthed it instead. The mosfets are not electrically connected to any part of the waterblock becuse I used the original mica shims, (which are a very good isolator, but also transfer heat well).

This was mainly for additional safety as the metal back plate of the mosfets measured earth when tested.

I never did find any reason for this "live sink" thing but whatever the PSU is still working fine a year on. I've done a few developments, one to help cool a coil that was getting a little hot and the other to help natural airflow, which was basically removing all unneeded plug wiring and making large cutouts in the chassis.

http://www.zerofanzone.co.uk/pictures/jpg/psucase5.jpg

Just as a quick addition to the live heatsink topic, my Aopen HX-08 power supply (300W that came with the HX-08 case) has a screw holding the heatsink to the case, so the sinks in this power supply are grounded, and should be easy to replace with a waterblock setup.

Yeah I've got one of the Aopen 300 watts here as I'm still using the HX08. I think only one of the sinks inside is earthed to the chassis from memory, but I've not tested it out recently. Psu's vary greatly in internal layout and design, and a lot have live sinks. Although my water-cooling build worked for me with the Enermax there is nothing to say it would work with every different PSU make or type.

Yes, only one of the sinks is earthed to the back of the case, but we can always try the touch-and-jump method to test the other sink cant we? :)

Rather you than me! with Uk mains at 240v AC I'd use a volt meter myself ;)

Do you have any details on that? (pic not necessary)

Also, I'm having a hard time understanding how a mica shim isolates the MOSFET from the block: a metal screw that touches the top of the MOSFET, or contacts it from within the hole, should connect to the block, no?

Well no because when using a mica plate you (should) use a plastic recessed washer that enters the screw hole and flanges out over the IC, this insulates the screw from the IC's package.

Ok I can only talk about the PSU I modified, (which is an early Enermax 550 watt sample), and as such the mosfets it contains. The mosfets all except the small one have a plastic package with a metal back piece. the plastic moulded part of the front continues around the screw hole all the way to the rear so there is no metal mosfet back plate to screw contact when it's fixed in place on the heatsink. The single small mosfet has a complete plastic package hence no mica shim on that one. Now in the standard PSU it was just the mica shim that was preventing the metal earthed mosfet back plate from making an electrical connection to the live heatsink. This is ok as it's what the mica shims are designed to do, I still don't have any truly confirmed reason why it is made with live sinks like this, I'd love to know the exact reason, some have given educated guesses but no one has said the reason is X. My understanding of electronics is pretty limited so..............

I reused the mica shims in my water-block design as I wasn't 100% sure the backplate of the mosfets remains earthed all the time and also thought it was a sensible extra safety measure in case for some reason the mosfet back plate was to go live via a fault etc. The mica shims are surprisingly good at heat transfer, not sure they would work well at Cpu heat energy levels but they are probably no worse than the chewing gum type "Tims" some heatsink makers use.

The coil that was getting hot, (70C plus under lengthy gaming sessions), may not have been a problem in it's self as the PSU still worked fine. I just didn't like the extra heat it was adding to the PSU insides.

Bit of a bodge really but it appears to work. I just made a copper bracket to the block that's fixed with a few screws and rests on the coil, again using a mica shim to prevent any possible earthing of the coil if the thin protective coating were scratched off. Coil doesn't go above 45C now. I have a pic anyway which may help show what I did, and my site and article will be all updated soon.

http://www.zerofanzone.co.uk/pictures/jpg/psu_coil1.jpg

jeez I like that BladeRunner

you're not inclined to quote a price for one ?
(PM me if so)

nice work

bladerunner.........

so could i just epoxy a wb onto the current sinks?

or would i have to have something to insulate it...??
(i don't want to remove the sinks)

Orgy

As I said every PSU is different so I can't tell what yours is like internally. If it's the same as mine then the sinks would be live so I'd not favour epoxying a water-block to something that's live. If it made electrical contact then the cooling system would become live either shorting to earth or through some other part of the PC.


unregistered

Unfortunately I don't have the time or equipment to make them for others atm, also as said PSU's all differ in design and layout so unless I had access to every PSU I wouldn't be able to ensure a successful build. There can be a lot more to it all than just simply cooling the mosfets effectively, I built mine because I wanted a fanless environment, willing to accept killing it or other parts of my PC. Others have done it have a look on http://www.digital-explosion.co.uk for instance make it sound easy, but I'm not so sure myself and wouldn't suggest you try unless you are fully prepared to accept the risks of failure in some shape or form.