34 PSUs reviewd. 34!

PSUs: The source of DC power for all components in the PC & often a big noise source.

Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee, Devonavar

Post Reply
meglamaniac
Posts: 380
Joined: Thu Jul 15, 2004 12:44 pm
Location: UK

34 PSUs reviewd. 34!

Post by meglamaniac » Mon Sep 26, 2005 7:02 am

Over at Hexus they may not be silence orientated, but they certainly do a thorough job!

There is some interesting info in there, and also some intriguing results.
Worth a quick look.

[edit]
It might be a bit slow right now as it appears to have been posted on slashdot as well...

moritz
Posts: 167
Joined: Sat Mar 26, 2005 7:06 am
Location: Germany
Contact:

Post by moritz » Mon Sep 26, 2005 7:12 am

No Seasonics and no Antecs, though. And somebody on Slashdot commented: The review tested all these products in FSG Groups facility, an employee of FSG Group is said to be "sexy" in the review, and a product from FSG Group won? Yeah, right... Hohum.
I get my kicks on channel six.

MikeC
Site Admin
Posts: 12283
Joined: Sun Aug 11, 2002 3:26 pm
Location: Vancouver, BC, Canada
Contact:

Post by MikeC » Mon Sep 26, 2005 3:43 pm

The review was actually posted a few days ago and I went carefully through the whole thing, first thinking, "WOW!, great that someone else is taking this seriously"... then increasingly becoming disenchanted as I looked at more of the details... and finally realizing that it's almost as misleading as the usual hooey that passes for PSU reviews.

OK, having said all that, I guess I have to be specific. Where shall I begin? How about with what moritz mentioned already...

1) Too many brands missing: Seasonic, Antec and many others
2) Misleading Voltage Regulation graphs: PSUs with the highest voltage positioned on top, the ones with the lowest on bottom. The implication is that the highest is the best and lowest is worst -- when in fact the one closest to the target voltage is the best. The 12V VR graphs has a 12.8V PSU on top (6.67% too high, which is beyond the +/-5% VR tolerance spec) and 11.58V PSU on bottom (just 3.5% low, which is in spec).
3) Misleading Efficiency graphs: No indication of at what power this was measured or whether all the PSU efficiency was calculated for the same power level. We know how much it changes with load -- it can go from <75% at 90W to >80% at 400W! This graph could be completely doctored, we have no idea.
4) Misleading PSU Temp graph: Again, a single number, with no indication of power level -- just that it's "at load". Sorry, that's just not good enough.
5) No acoustic testing, despite access to anechoic chamber and equipment, only 2 sets of data on two FSP PSUs.

The best thing about this review is that it helps debunk the "need" for high power.

But then they write, "What holds the majority of capable supplies back from powering a system properly isn't the fact they can't make the power." What is? According to them, it's that "the vast majority of issues are down to vendors not tuning the internals of the supply to cope with the demands a modern system will make on a certain voltage rail, usually +12V." The implication is that if they did, all the PSUs tested would be equally capable of provided good stable power to any PC?! Then why bother with testing? Why not just buy the cheapest PSU that says it's ATX12V v2?
Mike Chin, SPCR Editor/Publisher
Support SPCR by buying your gear through this link: Amazon

slau
Posts: 11
Joined: Sun Aug 11, 2002 3:26 pm

Post by slau » Mon Sep 26, 2005 10:48 pm

Yeah, I was thinking, oooh Hexus did something good... But from what I managed to see - during the great Slashdotting, it seems the review was off the usual Hexus quality. Lot's of sensationlism, the old standby of British tabloid journalism.

By the way, how much would that PSU tester you have cost?

dukla2000
*Lifetime Patron*
Posts: 1465
Joined: Sun Mar 09, 2003 12:27 pm
Location: Reading.England.EU

Post by dukla2000 » Mon Sep 26, 2005 11:23 pm

This was linked here a couple of days ago. Agree although on the face of it the review had potential, the real shame is too little specific data is reported. I was pleased to see the LC (aka Allied aka Mercury aka ...) that I have stood up for here in the past actually survived but that was about all I could find.
MikeC wrote:2) Misleading Voltage Regulation graphs:
Coming back to the 'too little data' point: my beef is they give a single number average for each psu. So the psu that alternately provided 11.0 and 13.0 on the 12V line comes out looking good whereas at no time during the test was it in spec! Given all the kit they had access to I would have though max, min & standard deviation would have been available which really would have been interesting based on their test method which involved several variations of load on each psu.
[size=75]2009/Oct: Jetway JNC81-LF * 4850e naked under fanless Xigmatek Apache * Antec mini Skeleton w/Nexus 120mm PWM fan * Delta 90W brick w/Skeleton DC-DC board * WD2500BEVT 250Gb blue[/size]

Devonavar
SPCR Reviewer
Posts: 1850
Joined: Sun Sep 21, 2003 11:23 am
Location: Vancouver, BC, Canada

Post by Devonavar » Tue Sep 27, 2005 12:04 am

My reaction was the same as Mike's:

Wow, isn't it great that somebody's actually paying attention to efficiency! Then, looking at the details, and thinking ... so much potential, and so little done!

Like 99% of other site out of there, they insist on testing at full output power. Why???? Ditto the thermal testing... nice to have, and seems to be consistently done, but why only at full load when the fans are running at their fastest?

Someone else mentioned the "averaged" voltage regulation, and this also struck me as an extrememly poor way of judging a PSU, especially when the range measured is only for the upper 50% of the output.

They also suffer from trying to directly apply Intel's formal PSU testing guidelines. Testing at 50% and 100% load doesn't produce comparable results unless all the PSUs are the same output level ... which they're not. And, why oh why didn't they list the data for 20% load, as Intel lists. They also must have seriously fudged some of the load tables, since IIRC, Intel only lists tables up to 450W output, and the majority of the PSUs tested were above this.

Oh well ... I do appreciate the effort. Their attitude seems to have been excellent. I like emphasis on efficiency and thermals, and the admission that high powered PSUs are unnecessary was a huge step forward ... but their results don't seem to back this up. Maybe the testing was done before they reached those conclusions... <shrug>

TheLostSwede
Posts: 6
Joined: Tue Sep 27, 2005 1:53 am

Post by TheLostSwede » Tue Sep 27, 2005 2:45 am

Not wanting to slag Hexus off, but I want to let you in on some information here. I work for another UK IT publication, which yes I know it sounds dodgy, but let me explain a couple of things here.
First of all, I have had access to a labs in Germany that belongs to Tagan, a European PSU manufacturer. Hexus was offered to go the same place for their test, but declined that offer as they didn't want to be seen as partial. So instead they spent some time in Taiwan around Computex this year to do their testing at FSP's facility.
Now what I don't get is the matter of fact that FSP seems to be using older and less technically advanced equipment that that I have access to in Germany. Hence some of the convoluted calculations etc. that Hexus has used to explain parts of their testing. I had some serious concerns with their test results and I would go as far as to say that they are meaningless.
There are some problems when it comes to testing PSUs, the lack of information is the most frustrating one and they guys here at SPCR is doing a great job and I've tried to take onboard as much as I can from what I have read here. This is the only site I link to when it comes to PSU test as I don't see any other source as reliable, due to the way people test PSUs. You can't use something like motherboard monitor and think that that'll do and that you'll have proper numbers, as they're usually way of if you use a multimeter.
Anyhow, I was expecting something quite spectacular considering the time it took Hexus to write up this group test, but I was highly disapointed with the outcome.

Post Reply