How do passive PSU's work?

[Carlos1010]

Hi all,

So recently i've heard alot of talk about these types of power supplies and i'm just curious how they work. I heard that they are close to silent but how can it do that without a fan. I mean like I thought that a PSU without a fan can explode or something. Also, why don't alot of people get these types of PSU's, I haven't seen alot of people with these kinds of PSU's in their signature. Are they like a bad tier or something? Thanks in advanced! 

[Canada EH]

There is no fan.

There are some server PSU's that sound like a 747 taking off. Like my 12V 75A 750W Dell PSU I use for not its intended purpose, but rather its modified for the RC world.

Regular ATX PSU's with fans are good. Not too sure why people would choose passive PSU.

 

[Carlos1010]

1 minute ago, Canada EH said:

There is no fan.

Yeah I know

[W-L]

1 minute ago, Carlos1010 said:

Hi all,

So recently i've heard alot of talk about these types of power supplies and i'm just curious how they work. I heard that they are close to silent but how can it do that without a fan. I mean like I thought that a PSU without a fan can explode or something. Also, why don't alot of people get these types of PSU's, I haven't seen alot of people with these kinds of PSU's in their signature. Are they like a bad tier or something? Thanks in advanced! 

Usually they have a block of heatsinking to help dissipate the heat outside of the unit or look a bit like swiss cheese with lots of extra vent holes for heat to passively rise out of the unit. 

[Canada EH]

Just now, Carlos1010 said:

Yeah I know

So then why the title??????????????????

Quote

How do passive PSU's work?

 

They just use Heatsinks. And conventional air flow, like an old VW Beatle.

[mariushm]

Man, you sure ask a lot of so basic questions that any curious person would find answers for with some Google searches.

 

A power supply converts a higher voltage to lower voltages used by computers. This conversion, like anything in this world, is not done perfectly. There are some losses which manifest as heat.

 

Integrated chips, components, can handle some heat but only up to some point. Some chips can be up to 85 degrees Celsius, some more basic but important components for power supplies (mosfets, diodes) may function properly at up to 125.150 degrees Celsius .. but of course, it's never a good idea to let these components reach such high temperatures. For example, one 125c rated component may be seen by an engineer as one that can only go as high as 105c.

 

In order to keep those hot components cool, heatsinks are used. The heatsinks take the heat from the tiny chip surfaces, spread it all over the surface of the metal in the heatsink and then this heat slowly transfers in the air that's around the metal.  A heatsink can only "suck" and push in the air so much heat until it reaches its limits. By moving a lot of air over the surface of the heatsink (which happens when you install a fan in the power supply), the heatsink can move a lot of heat into the air much faster, so it manages to keep the small chips touching it much cooler.

 

As engineers design more efficient power supply designs and more efficient chips, there's less waste produce as heat, so in general the chips and components don't heat as much.

 

In the case of semi-passive power supplies, engineers know that up to a particular power output (let's say 200w of 650w), the performance of the heatsink chosen for a part is good enough to keep that chip or component lower than the maximum temperature allowed (for example 100c for 125c parts).  When the power output goes over that value, they know they'll have to help the heatsink with some extra air from the fan, to maintain the temperatures below that threshold, so that's when the power supply turns the fan on and from that point, the rotation speed is often adjusted based on how hot some of the most hot components are in the power supply (there's usually one or two temperature sensors in strategic places).

 

Completely passive power supplies are very very efficient (gold or higher) so they produce from the start very little heat, and they also employ smart designs like placing some on the hotter chips on the bottom of the circuit board, touching the metal case of the power supply. This way, the whole case of the power supply acts as a heatsink for those components and the case of your computer itself spreads some of the heat (because you screw the power supply to the case and there's metal to metal contact). The other heatsinks inside the power supply are much bigger and more capable (but usually also much more expensive), to make sure the components will stay within allowed temperatures over the whole output power of the power supply.

 

The huge surface of the power supply case acting as heatsink and the normal air movement inside the power supply is enough to keep all components in a fully passive power supply within reasonable temperatures.  (natural convection, warm air rises so cold air comes from the bottom of your case through the power supply and naturally raises to the top of your case)

 

 

I mean like I thought that a PSU without a fan can explode or something.

 

When a chip or component gets too hot, parts of its insides can melt together so there can be short circuits inside. That makes the component blow up or just crack in separate parts and make smoke. In the worst case scenario, the fuse in your electricity panel can be triggered to protect your mains wiring and the power supply is dead.

if you keep these components within allowed temperatures, passive or not, the power supply will work fine.

[Streetguru]

20 minutes ago, Carlos1010 said:

 

A 500W Passive PSU for example is more like a 1000W PSU that just runs very efficiently and produces very little heat, it's also why it costs as much as a 1000W PSU

and also why some higher end PSUs have a zero RPM mode that doesn't kick on the fan until it's needed, making it 100% silent at idle

[Carlos1010]

6 minutes ago, mariushm said:

 

Man, you sure ask a lot of so basic questions that any curious person would find answers for with some Google searches.

 

Thanks dude for you response, yeah I know I question alot questions and sometimes like this question I cannot find alot of answers too. This forum is the best because people like you do the difference. Thanks bro

[Canada EH]

Yeah he put a lot of effort into this question that would have been so easily found using the interwebs search feature.

[STRMfrmXMN]

Completely fanless PSUs need fans elsewhere to keep cool enough, though they often have absurd heat tolerances anyways.

[mariushm]

1 hour ago, STRMfrmXMN said:

Completely fanless PSUs need fans elsewhere to keep cool enough, though they often have absurd heat tolerances anyways.

Not true.

Some power supplies are efficient enough that they can stay cool just relying on the heatsink mass and natural air convection inside the power supply to function right.

 

You will see sometimes stickers or labels (that are removed before installing power supplies) saying that the power supply is only responsible for keeping itself at reasonable temperatures and to not assume you can build a 100% passively cooled computer.

 

Nowadays, there are mATX and ITX boards readily available with energy efficient processors and big passive heatsinks, and some can even work with 65w or higher laptop adapter bricks which are also passive, so making a fully passive computer is fairly easy to achieve.

 

Yes, it's often a good idea to have a large diameter fan at low voltage spinning slowly just to move air a bit inside the case and prevent pockets of warm air from staying in one place and overheating some areas of your system. You could even add a fan controller that would stop the fan completely when temperatures inside are below some threshold so you'd still have an absolutely passive system.

 

As an example here's a Seasonic SS-460FL fanless power supply : https://www.techpowerup.com/reviews/Seasonic/SS-460FL_V2/6.html

You can see at the bottom of the page that at 500 watts they measured a peak temperature inside the power supply of 20c over the ambient (air inside the case, outside the power supply), which was up to 45c .. so the inside the power supply the temperatures got up to around 65-70c which means the actual components were maybe about 2-3 degrees celsius hotter at maybe up to 75c. Considering most components in such power supplies are rated for 125c..150c you can see that there's still a bit of headroom.  

 

Here's some thermal images of the 520w fanless model in the same series as the one linked above: https://www.computerbase.de/2013-06/sea-sonic-platinum-series-fanless-520-watt-test/9/

You can see the hottest points of the power supply were at around 70c..73c, and the spots were the bridge rectifiers and a heatsink holding 125c or higher rated components. The bridge rectifiers they used (GBJ2506) are rated to work at up to 150c as you can see in datasheet: http://www.mouser.com/ds/2/115/ds21221-82782.pdf  and for the currents they'd meet in worst case scenario (about 3A each), they could still work continuously at up to 130-140c (as you can see in figure 1 on page 2 ...)

So you can still they're quite well designed.and they still have lots of margins.

 

[STRMfrmXMN]

1 hour ago, mariushm said:

Not true.

Some power supplies are efficient enough that they can stay cool just relying on the heatsink mass and natural air convection inside the power supply to function right.

 

You will see sometimes stickers or labels (that are removed before installing power supplies) saying that the power supply is only responsible for keeping itself at reasonable temperatures and to not assume you can build a 100% passively cooled computer.

 

Nowadays, there are mATX and ITX boards readily available with energy efficient processors and big passive heatsinks, and some can even work with 65w or higher laptop adapter bricks which are also passive, so making a fully passive computer is fairly easy to achieve.

 

Yes, it's often a good idea to have a large diameter fan at low voltage spinning slowly just to move air a bit inside the case and prevent pockets of warm air from staying in one place and overheating some areas of your system. You could even add a fan controller that would stop the fan completely when temperatures inside are below some threshold so you'd still have an absolutely passive system.

 

As an example here's a Seasonic SS-460FL fanless power supply : https://www.techpowerup.com/reviews/Seasonic/SS-460FL_V2/6.html

You can see at the bottom of the page that at 500 watts they measured a peak temperature inside the power supply of 20c over the ambient (air inside the case, outside the power supply), which was up to 45c .. so the inside the power supply the temperatures got up to around 65-70c which means the actual components were maybe about 2-3 degrees celsius hotter at maybe up to 75c. Considering most components in such power supplies are rated for 125c..150c you can see that there's still a bit of headroom.  

 

Here's some thermal images of the 520w fanless model in the same series as the one linked above: https://www.computerbase.de/2013-06/sea-sonic-platinum-series-fanless-520-watt-test/9/

You can see the hottest points of the power supply were at around 70c..73c, and the spots were the bridge rectifiers and a heatsink holding 125c or higher rated components. The bridge rectifiers they used (GBJ2506) are rated to work at up to 150c as you can see in datasheet: http://www.mouser.com/ds/2/115/ds21221-82782.pdf  and for the currents they'd meet in worst case scenario (about 3A each), they could still work continuously at up to 130-140c (as you can see in figure 1 on page 2 ...)

So you can still they're quite well designed.and they still have lots of margins.

 

Hmm, I've never felt totally safe having a completely passive system with a fanless PSU but I guess what little air moving about in the PSU there is can keep things cool enough?