Rethinking Fanless psu's impact on system heat

sgtpokey · Post by **sgtpokey** » Tue Aug 31, 2004 7:43 am

Here's a link from HTPCnews regarding a 0dba PSU. It is basically fanless except when the power load>250watts, at which load 80mm fan kicks in.

What caught my eye was I'm a semi-regular HTPCNews reader. Their reviews are:
a) pretty good from a silent perspective
b) relevant to us because the components used are generally challenging to silence due to case space/power requirements/etc.

So it was interesting to see that in a media case (one with a P4 and a Radeon 9800) they were able to get decent system temps out of a fanless PSU. The heat's gotta go somewhere, but it doesn't seem to be impacting their critical components too much..

http://www.htpcnews.com/main.php?id=epower_1

One note, they list the case as a Custom HTPC case, but I believe the case only has one 80mm case fan at low rpm (1500 rpm)... I'll go have to ask them about it...

StealthGirl · Post by **StealthGirl** » Tue Aug 31, 2004 8:58 am

This is a really cool idea ... but, why not just mod any PSU this way? I bet we have a couple of EE types in the group that can whip us up a schematic of a current switch. Guys? Any volunteers?

sgtpokey · Post by **sgtpokey** » Tue Aug 31, 2004 10:41 am

Hmm, now that you mention it, I wonder exactly what is the difference with this particular model vs a standard psu.

From the images, it doesn't look to have amazingly different thermal engineering, except for that the fan cage/heatfin thingy that hopefully dumps some heat outside of the case.

Hmm, maybe someday a better review may come along, from one of our favorite websites...

Or I'd buy one and give it a look-see, 'cept I'm not that curious enough to pay $150, esp. since I'm waiting for that Zalman external fanless psu.

greeef · Post by **greeef** » Tue Aug 31, 2004 11:23 am

surely current is irrelevant, it's only temperature that matters.

I think that wattage trigger thing is a little bit of hype - people are obviously getting tired of thermally controlled fans.

Post by **MikeC** » Tue Aug 31, 2004 11:43 am

Very dubious that the fan trigger is current or wattage. Just as easily done with temp, usually is (cheaper to do):

1) Assume 40C as typical in-case ambient
2) check internal temp of PSU at 250W load in 40C temp
3) ensure core components can withstand this internal temp
4) calibrate thermistor to kick fan on at this internal temp

Too bad there is so little info about the rest of the review system, esp the ventilation & airflow -- and what kind of power load the PSU was at -- and what its efficiency is. Looks like max DC power draw was not more than ~100W.

What we do know from the reviewer's MBM5 screen capture is that there were two fans in the case, one at 3000-3500 rpm, the other at >5000 rpm. I am guessing the first fan would have been the case fan, the second fan the CPU cooler fan. I can't imagine that these fans would be quiet, and I bet they create a LOT of airflow. No wonder the PSU fan never turned on...

jamesavery22 · Post by **jamesavery22** » Tue Aug 31, 2004 12:39 pm

greeef wrote:surely current is irrelevant, it's only temperature that matters.

I think that wattage trigger thing is a little bit of hype - people are obviously getting tired of thermally controlled fans.

Thermally controlled fans are one thing but turning it on and turning it completely off a bunch of times is something else...

Even if the user has no solid load (no FaH or anything in the background) what will more than likely happen is the heatsink will heatup enough, force the fan to kick in, cool down, turn the fan off, heat up, kick in the fan, over and over and over.
Doing this "stepping" effect in temps is really bad on any component.

If you want a silent PSU dish out the extra bucks to buy the actual silent ones that dont have a fan at all...

saiyajin · Post by **saiyajin** » Tue Aug 31, 2004 2:36 pm

I have seen many reviews lately about fanless PSUs but everybody here are saying that we should not use fanless PSU... I think that this is due to bad experiences in the past... but come on... I would guess that someone somewhere has a Fanless PSUs that is just working as the manufacturer claims and it doesnt cause any problem/overheating in the the computer!!!! If you have a Fanless PSUs, share your experience with us.

sgtpokey · Post by **sgtpokey** » Tue Aug 31, 2004 3:38 pm

oops, I completely missed the MBM chart with the fan rpms... kinda funny though, those guys usually run their fans at acceptable silent pc RPMs.

Yeah, that probably does explain his case temps and fanless PSU... he just has a loud case fan.

oh well.

Mats · Post by **Mats** » Tue Aug 31, 2004 5:06 pm

By the way, some pics of the long awaited Antec Phantom have been used in their site here and here. Small pics, nothing interesting.
What does it mean? Absolutely nothing, but it's strange that they're showing products that haven't been introduced. No info at all.

...or, have I missed something again?

Mats · Post by **Mats** » Tue Aug 31, 2004 5:41 pm

Here's another review (not so good) of the Epower PSU with some pics of the inside.

HammerSandwich · Post by **HammerSandwich** » Tue Aug 31, 2004 11:13 pm

I emailed Epower to send Mike a review sample. Join in to make it happen!

silvervarg · Post by **silvervarg** » Wed Sep 01, 2004 1:04 am

The small blue PCB mounted upside down on the heatsink looks very much like a fan controller. The most common reason to mount it on a heatsink is to put a sensor on the heatsink, so this is a good indication that it uses temperature to controll fan speed.
The size of the heatsinks looks promising, so I do hope they send a review sample to MikeC.

Bluefront · Post by **Bluefront** » Mon Sep 20, 2004 3:24 am

FWIW...This E-Power PSU is now available at NewEgg.

I sure like the idea of a back-up fan that turns on at high temps. This fan control seems to be determined by a remote sensor. Hard to tell....

Rethinking Fanless psu's impact on system heat

Rethinking Fanless psu's impact on system heat

Fanless unless needed