Changes to PSU Testing

Want to talk about one of the articles in SPCR? Here's the forum for you.
Post Reply
MikeC
Site Admin
Posts: 12285
Joined: Sun Aug 11, 2002 3:26 pm
Location: Vancouver, BC, Canada
Contact:

Changes to PSU Testing

Post by MikeC » Fri Feb 11, 2011 6:18 am

This is to open a discussion with SPCR old timers. People who have been following SPCR closely for a while.

I've been working on the usual periodic updates to recommended/reference articles... and realize that both the case and PSU fundamentals need a closer look. To be specific...

My "hotbox" for PSU testing was unique when it was introduced way back in 2004 -- http://www.silentpcreview.com/article148-page1.html Every PSU review since then has featured the hotbox to simulate realistically hot conditions inside a case, and SPCR reviews are still fantastically useful if you want to know the relationship between intake temp and PSU fan noise.

However, in early 2011, how relevant is this hotbox? Consider that the vast majority of enthusiast oriented cases put the PSU on the bottom, often with the intake from beneath the case -- PSUs just aren't subject to the same uber hot conditions any more.

The PSU tester needs to be rebuilt. It's been abused for 7 years. Time to retire it. I'm thinking to stop hotbox testing. It would shorten the PSU test time to half. Maybe just have a single run with added controlled heat at 300W load.

I'm also making changes in the case basics/recommendations -- top-mount PSU cases (standard ATX tower) recommended only for system up to ~200W TDP; only bottom mount PSU cases for higher.

Thoughts?

Modo
Posts: 486
Joined: Wed Apr 16, 2008 3:32 am
Location: Poland

Re: Changes to PSU Testing

Post by Modo » Fri Feb 11, 2011 6:40 am

I'd keep the hotbox in some form. There are still many cases with the PSU at the top, and some PCs simply run in hot places, so testing this aspect of PSU performance remains important.

MikeC
Site Admin
Posts: 12285
Joined: Sun Aug 11, 2002 3:26 pm
Location: Vancouver, BC, Canada
Contact:

Re: Changes to PSU Testing

Post by MikeC » Fri Feb 11, 2011 6:50 am

Modo wrote:I'd keep the hotbox in some form. 1. There are still many cases with the PSU at the top, and 2. some PCs simply run in hot places, so testing this aspect of PSU performance remains important.
1. OK, but I don't recommend any of those cases unless your system is <200W TDP, at which point with a half decent PSU, its fan will never ramp up.
2. yeah, this is true. I'm still not sold on doing it w/ the hotbox tho. Am thinking about a radiant heater I can turn on for a minute, then take a quick check of air temp at the intake before running tests at selected power levels -- not continuously as I do now. The time required for PSU testing is one of the reasons you see them so rarely these days. It's such tedious work!

frenchie
Friend of SPCR
Posts: 1346
Joined: Tue Jul 08, 2008 4:53 am
Location: CT

Re: Changes to PSU Testing

Post by frenchie » Fri Feb 11, 2011 7:12 am

I like the hotbox because it gives a good idea of general PSU stability and reliability. If it can take the hotbox (voltage values stay good, PSU operates normally), it says a lot to me about design, testing, and quality control at the manufacturer's level. On the other hands, a PSU that fails the hotbox does scream "failure" to me.

I understand your point about more and more "enthusiast" cases having separate intakes for PSUs, but it doesn't mean that your computer should burn and die the day the airconditionning stops working on a hot summer day. Not to mention if you don't have an airconditionner :) .
Maybe you could test the PSU in the hotbox only at maximum rated output (instead of taking measurements at regular power intervals). A PSU that passes this test will have a very easy time in one of those sperate intake cases.

Falkon
Posts: 115
Joined: Tue Feb 08, 2011 6:59 pm
Location: Canada

Re: Changes to PSU Testing

Post by Falkon » Fri Feb 11, 2011 7:17 am

MikeC wrote:I'm also making changes in the case basics/recommendations -- top-mount PSU cases (standard ATX tower) recommended only for system up to ~200W TDP; only bottom mount PSU cases for higher.

Thoughts?
What about cases like the NSK3480 where the PSU is top mounted but draws fresh air from the exterior?

MikeC
Site Admin
Posts: 12285
Joined: Sun Aug 11, 2002 3:26 pm
Location: Vancouver, BC, Canada
Contact:

Re: Changes to PSU Testing

Post by MikeC » Fri Feb 11, 2011 7:27 am

frenchie --

You know, the hotbox was never about testing for reliability under heat, and very few PSUs have actually failed in SPCR testing over the years. Lots have been noisy, but few have actually failed. I doubt it is that good a test for reliability under heat, because it is too short. Typical PSU failure due to heat occurs after long continued exposure, not just a single burst -- tho I am sure that happens sometimes, and a heatwave is probably a trigger for a PSU on the verge of failure. I've had a small handful of PSUs go on me in actual use over the years, and the most common failure was with PSUs I had modded with near-silent, barely spinning fans, mounted in standard ATX mid-tower cases. Yup, the heat got to them over the long term, even though those system drew very little power (under 100W, typically).

I am the first to say -- don't rely on SPCR hotbox testing as an indicator of reliability.
Falkon wrote:What about cases like the NSK3480 where the PSU is top mounted but draws fresh air from the exterior?
Obviously that's an exception.

MikeC
Site Admin
Posts: 12285
Joined: Sun Aug 11, 2002 3:26 pm
Location: Vancouver, BC, Canada
Contact:

Re: Changes to PSU Testing

Post by MikeC » Fri Feb 11, 2011 7:33 am

frenchie wrote:I understand your point about more and more "enthusiast" cases having separate intakes for PSUs.
It is not just "more and more" -- it's ALL of them. Show me one high performance case that has the PSU in the conventional spot. Yeah, that is a challenge. You will not find any such cases released in, say.... the last 2 years.

But aside from that, this is SPCR, and we're focused mostly on quiet stuff, and the simple point is that anyone who really wants quiet should get a case with an outside intake for the PSU. Period. Once that position is taken, there is ZERO reason for the hotbox. It has become... an historical artifact. (Survive long enough and we all get there! :lol: )

frenchie
Friend of SPCR
Posts: 1346
Joined: Tue Jul 08, 2008 4:53 am
Location: CT

Re: Changes to PSU Testing

Post by frenchie » Fri Feb 11, 2011 8:06 am

Thanks for all the replies Mike !! :)
I should have realized by now that what it's my head doesn't automatically get written on my screen if I don't type it on the keyboard... I was thinking about fanless PSUs...
So how about fanless PSUs ? If I remeber correctly, during the fanless PSU torture tests you ran last august, the PSUs that failed (read : were not able to deliver full load output without forced airflow) failed beause of the heat (except for the SilenX POS that just sucked). Do you think hotbox tests would still make sense for fanless PSUs ?

MikeC
Site Admin
Posts: 12285
Joined: Sun Aug 11, 2002 3:26 pm
Location: Vancouver, BC, Canada
Contact:

Re: Changes to PSU Testing

Post by MikeC » Fri Feb 11, 2011 8:14 am

frenchie wrote:Thanks for all the replies Mike !! :)
I should have realized by now that what it's my head doesn't automatically get written on my screen if I don't type it on the keyboard...
:lol:
I was thinking about fanless PSUs... So how about fanless PSUs ? If I remeber correctly, during the fanless PSU torture tests you ran last august, the PSUs that failed (read : were not able to deliver full load output without forced airflow) failed beause of the heat (except for the SilenX POS that just sucked). Do you think hotbox tests would still make sense for fanless PSUs ?
I doubt anyone else will venture into that product sector. Seasonic has sewn it up. Why would anyone else even try? The market just isn't big enough.

Big Pimp Daddy
Posts: 175
Joined: Mon May 30, 2005 11:26 am
Location: Sunny Swansea

Re: Changes to PSU Testing

Post by Big Pimp Daddy » Fri Feb 11, 2011 9:04 am

Obiously this is just my personal opinion, but I don't see much point in the hotbox, never have. I can appreciate that 5 or 6 years ago some people would not have fresh air going to the PSU, but SPCR types all did even then surely? That was the first mod I made to my case - a duct for the PSU. I honestly haven't looked at hotbox data for any PSU tested in the last few years, I skip straight to the "out of case" data. I'm sure there are some people who find hotbox data useful for various reasons, but surely this niche is outweighed by the time (and hence cost) of reviewing PSUs in this way. Plus the implication that cutting this part of the test would result in more PSU reviews seals the deal for me.

Modo
Posts: 486
Joined: Wed Apr 16, 2008 3:32 am
Location: Poland

Re: Changes to PSU Testing

Post by Modo » Fri Feb 11, 2011 9:53 am

MikeC wrote: 1. OK, but I don't recommend any of those cases unless your system is <200W TDP, at which point with a half decent PSU, its fan will never ramp up.
I see your point. I guess it makes sense to assume that people coming to SPCR will take time to read, and not just browse the list of PSUs.
MikeC wrote: 2. yeah, this is true. I'm still not sold on doing it w/ the hotbox tho. Am thinking about a radiant heater I can turn on for a minute, then take a quick check of air temp at the intake before running tests at selected power levels -- not continuously as I do now. The time required for PSU testing is one of the reasons you see them so rarely these days. It's such tedious work!
True, it doesn't have to remain that complicated. I rather liked the straightforward hairdryer method used on the X-400. ;) So yeah, a space heater to simulate a hot summer day seems enough.

goatsandmonkeys
Posts: 91
Joined: Wed Oct 06, 2010 6:32 am
Location: new york

Re: Changes to PSU Testing

Post by goatsandmonkeys » Fri Feb 11, 2011 10:21 am

I really appreciate your exhaustive reviews. I have to say that the hotbox isn't THAT important to me anymore. I think before a PSU gets added to the recommended list or gets an editors choice it should be hotboxed. I consider the recommended parts to be a gold standard and I bet a LOT of people use them that don't read lots and lots of reviews. It's important for us to know how any of the top silent PSUs do in extreme heat.

Just my 2 cents. thanks again.

cordis
Posts: 1082
Joined: Thu Jan 15, 2009 10:56 pm
Location: San Jose

Re: Changes to PSU Testing

Post by cordis » Fri Feb 11, 2011 10:29 am

Yeah, I think a short term heated air intake test would be fine, just to see how fan speeds ramp up. I also the in-depth reviews, but if the depth of the reviews means that we only get a couple psu reviews done a year, I'd rather have more without the hotbox than less with it.

CA_Steve
Moderator
Posts: 7650
Joined: Thu Oct 06, 2005 4:36 am
Location: St. Louis, MO

Re: Changes to PSU Testing

Post by CA_Steve » Fri Feb 11, 2011 11:44 am

A couple of random thoughts.
- It really comes down to being able to show how the PSU fan ramps up at a given power output and local case temp conditions.
- When specifying a PSU, I look at the expected system load and then compare SPCR reviewed PSUs to see what meets the bar for dB.
- My ambient temp/case/airflow/etc is always going to be different than the conditions for the reviewed PSU.

So, having some indicator of fan performance under conditions worse than I will see is helpful.

Do you need to run a long hotbox test under multiple conditions? Probably not. But a test point on either side of where the fan ramps up under ideal conditions would be useful.

MikeC
Site Admin
Posts: 12285
Joined: Sun Aug 11, 2002 3:26 pm
Location: Vancouver, BC, Canada
Contact:

Re: Changes to PSU Testing

Post by MikeC » Fri Feb 11, 2011 11:56 am

cordis wrote:Yeah, I think a short term heated air intake test would be fine, just to see how fan speeds ramp up. I also the in-depth reviews, but if the depth of the reviews means that we only get a couple psu reviews done a year, I'd rather have more without the hotbox than less with it.
The # of psus reviewed has been decling, but I don't think it's that low yet. Something like 9 or 10 were tested last year.

So it seems, in general, that old hands are OK with the planned change: Load/noise testing at room temp, combined with spot checks at specific loads at perhaps 30/35C ambient. Need to figure out an expedient way of getting this heat.

Thanks for the feedback so far... keep it coming, esp if you have other ideas.

Falkon
Posts: 115
Joined: Tue Feb 08, 2011 6:59 pm
Location: Canada

Re: Changes to PSU Testing

Post by Falkon » Sat Feb 12, 2011 8:27 am

MikeC wrote:I'm also making changes in the case basics/recommendations -- top-mount PSU cases (standard ATX tower) recommended only for system up to ~200W TDP; only bottom mount PSU cases for higher.
MikeC wrote:
Falkon wrote:What about cases like the NSK3480 where the PSU is top mounted but draws fresh air from the exterior?
Obviously that's an exception.
Perhaps a more accurate conclusion would be that cases with PSUs drawing air from the interior would be for systems up to ~200W TDP and cases with PSUs drawing air from the exterior for higher TDP values.

Post Reply