Capacitor cooling - Heatsinks for caps

[TELVM]

Thermal Management of Electrolytic Capacitors

Cap Coolers




This is an intriguing idea and I'm considering ghetto-modding my own homebrew version.

What do you think of the concept?

[Sparkey55]

Re: Capacitor cooling - Heatsinks for caps

Thermal Management of Electrolytic Capacitors

Cap Coolers




This is an intriguing idea and I'm considering ghetto-modding my own homebrew version.

What do you think of the concept?

IMO, you would be wasting money and time. For mobo caps just put a cool breeze from a fan on them, all you need.

[mariushm]

Re: Capacitor cooling - Heatsinks for caps

I think you should actually MEASURE the temperature of capacitors before even starting to think about fabricating or buying such heatsinks.

The capacitors in a power supply don't heat that much to require heatsinks.

[Pentium4]

Re: Capacitor cooling - Heatsinks for caps

I think it's a cool idea, but not really worth it. It looks like they take up a lot of space, and the only electrolytic caps I can think of that might need them are on the output of VRM's, but polymers should be used there anyways. If you're already ordering from digikey, might as well throw in a couple and do some basic testing.

[TELVM]

Re: Capacitor cooling - Heatsinks for caps

... the only electrolytic caps I can think of that might need them are on the output of VRM's ...

Which is exactly what I have in mind. Some VRM lytics in old mobos are surrounded by coils and mosfets in close proximity and have to endure a heavy thermal beating by radiation no matter how many cfm you flow thru the case. Those caps would profit from some form of sinking.

To reiterate, I'm not gonna buy any of the gadgets that were posted above only to illustrate the concept. If anything I'd implement my own homebrew solution.

[mariushm]

Re: Capacitor cooling - Heatsinks for caps

Those capacitors are there because they often HAVE TO BE there, as close as possible to inductors and other hot parts.

Moving them further away may make the power supply oscillate or increase ripple and so on.

If they're already so cramped up in there, how do you hope to add heatsinks around the capacitors?
At best, you can move further away the minimum load resistors and maybe, just maybe you could desolder the large chokes and cover them in some potting compound to increase the heat dissipation area before soldering back to the pcb.

And after all that... good capacitors are 2000h+ @ 105c , with estimations doubling the lifetime with each 10c....8000h @ 85c , 32000h @ 65c, 64000h@55c
In worst power supplies, the capacitors on secondary side are probably going to run at about 60-70c... maybe not even that much.
What are you going to do with some heatsinks, increase the lifetime of the capacitors from 10 years to 15 years? You probably won't use this power supply in 10 years anyway.

[TELVM]

Re: Capacitor cooling - Heatsinks for caps

Those capacitors are there because they often HAVE TO BE there, as close as possible to inductors and other hot parts.

Moving them further away may make the power supply oscillate or increase ripple and so on.

Who has talked about moving anything?

If they're already so cramped up in there, how do you hope to add heatsinks around the capacitors?

I was thinking on top of them . This could also made an interesting ballistic second stage if the cap ever explodes .

What are you going to do with some heatsinks, increase the lifetime of the capacitors from 10 years to 15 years? You probably won't use this power supply in 10 years anyway.

I'm talking VRM caps in vintage motherboards. I love tinkering and hate hot components, that's enough motivation for me.

[Sparkey55]

Re: Capacitor cooling - Heatsinks for caps

Which is exactly what I have in mind. Some VRM lytics in old mobos are surrounded by coils and mosfets in close proximity and have to endure a heavy thermal beating by radiation no matter how many cfm you flow thru the case. Those caps would profit from some form of sinking.

To reiterate, I'm not gonna buy any of the gadgets that were posted above only to illustrate the concept. If anything I'd implement my own homebrew solution.

In order for those types of heatsinks to do any good the plastic sleeves on the caps need to be removed where they come into contact with heatsink. Very little heat will be conducted with such small surface area on the caps. Futile but knock yourself out.

[mariushm]

Re: Capacitor cooling - Heatsinks for caps

I assumed you want to do some moving because I couldn't imagine how you would add those round heatsinks on the capacitors.

The tops of those capacitors by the VRM circuitry is probably the coolest point of the capacitors.

Now I didn't test so this is just a guess, but the ESR is so low on those capacitors that they don't really heat up all that much by themselves.
I would guess they heat up more through the leads because the ground plane acts as a heatsink for the mosfets in the VRM circuitry. I think adding some heatpipe system/heatsinks on mosfets would probably have a way stronger effect.

In this case, adding heatsink on top of capacitors would probably be like painting stripes on a car - stripes don't make the car faster, contrary to some people's opinion.

// yes, I'm saying a lot of probably. Again, everything I say is just guesswork.

[TELVM]

Re: Capacitor cooling - Heatsinks for caps

... I would guess they heat up more through the leads because the ground plane acts as a heatsink for the mosfets in the VRM circuitry ...

You got a point there. Backside heatsinks would probably be a much better solution.

[TELVM]

Re: Capacitor cooling - Heatsinks for caps

That's the spirit! :

[SeanB]

Re: Capacitor cooling - Heatsinks for caps

If the capacitor is getting hot then it will be either something else warming it or it being run way beyond ratings or it is in backwards. Blowing air over it from a small fan will help with the first 2 more than any passive heatsink ever will.

[PeteS in CA]

Re: Capacitor cooling - Heatsinks for caps

Thermal Management of Electrolytic Capacitors

Cap Coolers




This is an intriguing idea and I'm considering ghetto-modding my own homebrew version.

What do you think of the concept?

Well, a couple of things. First, the heatsink manufacturer, Aavid is a significant and reputable company. More to the point, these caps are large "beer can" (aka "Large Can") capacitors, not the "Miniature" capacitors used for motherboards and in P/S O/Ps. So it's more curiosity than practical. I suppose bits of sheet metal could be glued with some sort of thermal epoxy atop motherboard capacitors, but a higher speed fan - i.e. more cooling air - would probably do as well or better and not risk messing up the airflow.

IIRC, Rifa had/has a series of large can lytics that have a "hole" in the center to allow conducting heat out of the part.

[eccerr0r]

Re: Capacitor cooling - Heatsinks for caps

Again, if the capacitor has low ESR it should not be generating much heat to begin with. The dielectric/insulators to prevent the can from touching the outside of the capacitor will virtually negate any heatsinking done. Best you can do is heatsink the leads/terminals but there's a huge thermal impedance through the thin leads.

In my opinion it's pointless to heatsink capacitors - just make sure other things that dissipate heat dissipate it far away or somehow make sure it does not heat the capacitor.

[TELVM]

Re: Capacitor cooling - Heatsinks for caps

These are the two guys I'm worried about:

>Expand >Expand

They are placed in a murderous cross-fire between inductors and mosfets in close proximity, that already destroyed their predecessors:


I move tons of air thru the case, and have ghetto-modded a heatsink for the mosfets:

>Expand

And yet those two caps remain too warm to my taste, for as Mariushm says they are heating up through the leads because the ground plane acts as a heatsink for the inductors and mosfets. Chokes can get even warmer than unsunk mosfets, brrrrr:


I'm now brainfarting a backside flat heatsink to see how that works. The last resort will be a fan blowing on the mobo backside right under this particular zone.

[Kiriakos GR]

Re: Capacitor cooling - Heatsinks for caps

What do you think of the concept?

I think that empty pockets born unworthy ideas so to hook peoples mind and exploit their wallet.

Every single capacitor is designed to handle specific conditions of operation.
Those engineers who design our capacitors for more than 50 years they are not idiots.

[Kiriakos GR]

Re: Capacitor cooling - Heatsinks for caps

They are placed in a murderous cross-fire between inductors and mosfets in close proximity, that already destroyed their predecessors

Your thermal camera image proves that you capacitors are cool as ice cubes.

[TELVM]

Re: Capacitor cooling - Heatsinks for caps

I wish that thermal camera was mine. That's another guy's pic of another comp, posted here just to illustrate how hot inductors can get.

[Kiriakos GR]

Re: Capacitor cooling - Heatsinks for caps

Any motherboard comes with five layers of copper.
Any produced heat it is dissipated that rapidly in all the board which makes it work as a giant heat-sink.
Thermal camera just for motherboard PCB inspections is almost useless.

[mariushm]

Re: Capacitor cooling - Heatsinks for caps

Don't generalize. The number of layers on a motherboard varies.. there are motherboards with 3-4 layers,there are motherboards with 8-10 layers. Nowadays, pretty much 4 layers is the minimum due to large pinout of processors and chipsets and all the decoupling capacitors required to be soldered around the socket.

Some video cards are made with even 12 layers.

The layers of copper do act like heatsinks but they're not that thick, most inner layers are just 15-30 micron thick of copper (1 oz copper = 1.37 mils = 0.0347 mm = ~35 microns)

Gigabyte and maybe other manufacturers use now 3oz copper, which just means the top and/or bottom layer use more copper.

re image.. the capacitors there are indeed relatively cool, looks like around 90f/32c for the capacitors. Everything else around is much hotter.

BUT...
the outer can of the capacitor is somewhat insulated from the insides of the capacitor so the inside temperature may be higher
AND
the tops of the capacitors are shiny white aluminum which may mess up the thermal camera, unless the camera is smart enough to adjust the coefficient. For proper measurement, the tops should be painted black or covered with a thin black material. So the temperature could be a bit higher but reported incorectly by the camera.

Either way, adding heatsinks would bring minimal benefits like I said.