SPCR

ouch

The behaviour is not much better than my previous NeoPower480, that I bought 2 years ago and recently sold with my old system.
For a new entry in the market of 2007, it's a flop IMO.
PS. I'd like to see a SPCR review of the new Nexus PSUs, they have interesting noise claims.

I have tested three very similar power supplies sold under Chieftec brand.
In Hungary, these are sold in 500, 560, 620W non-modular versions, and a 560W modular.
The 500W and 560W (both modular, and not-modular versions) had pretty good noise characteristics, and effeciency over 80% (~220-230V mains voltage here).

The 620W version had a few percent lower effeciency, and was much louder.
I think it's the same you tested. IMO Lian-Li made a bad decision with choosing the higher power version...
effeciency graph The upper green/red lines are the power factor, the two lower blue lines are the effeciency of the 500 and 560W versions, the lowest red line is the eff. of the 620W version.

Also, effeciency depends on the load distribution. High load on 12V, and almost no load on 3.3/5V lines gives higher effeciency. (lower voltage -> higher current -> more loss). Most PSUs I tested had measurably higher effeciency with high load on 12V line.

I do PSU tests using an electric load for Hungarian www.radikalmod.hu and www.hardwareoc.hu websites. And obviously english is not my main language

Thanks for your comments, lukr. Sounds like we should continue with our testing of the 500W model. It may end up being significantly better if your observations apply to our sample.

For the record, our load configuration is the same for all our PSU tests. We use the proportionate loading formula used by 80 Plus and by Intel. Basically the distribution of load at any aggregate power level is much greater on the 12V lines than any other, as it's based on the maximum rated power for each line.

We find that although high 12V loading will give higher efficiency results, reducing the 5 and 3.3V lines to almost nil does not actually help. There usually needs to be some kind of load on those lines for the 12V line to work well.

PS. I'd like to see a SPCR review of the new Nexus PSUs, they have interesting noise claims.

Seconded. On paper they are off the charts good, so this would be a high priority review IMVHO.

Is spcr going to venture once again into fan swapping/controlling?

I remember how spcr people used to show how all one needed was an efficient % psu and then a swap of the fan would rectify much for low expense.

~El~Jefe~ wrote:Is spcr going to venture once again into fan swapping/controlling?

I remember how spcr people used to show how all one needed was an efficient % psu and then a swap of the fan would rectify much for low expense.

Not in reviews. There are way too many good/quiet off-the-shelf PSUs today to merit that kind of modding discussion in a review. It's cheaper, better and safer to buy one that's already quiet & good... unless you're looking for the kind of silence only ultra fanatics seek. (Not healthy, really, because your heightened sensitivity can become impossible to satisfy and unmanageable; I speak from experience. )

~El~Jefe~ wrote:Is spcr going to venture once again into fan swapping/controlling?

I remember how spcr people used to show how all one needed was an efficient % psu and then a swap of the fan would rectify much for low expense.

I think the Lian-Li's fan is good, it's just got a poor fan controller and poor efficiency. Any PSU can be modded with a quieter fan, but with low(er) efficiency PSUs heat can be a problem; and then the question becomes why buy a PSU that you have to mod when Seasonic/Corsair/etc are quiet in stock form?

jaganath wrote:I think the Lian-Li's fan is good, it's just got a poor fan controller and poor efficiency. Any PSU can be modded with a quieter fan, but with low(er) efficiency PSUs heat can be a problem; and then the question becomes why buy a PSU that you have to mod when Seasonic/Corsair/etc are quiet in stock form?

Agreed. Very mediocre psu from Lian-Li as was to be expected.

I would also like to see a review on the new Nexus NX-8050 modular 80+ efficiency 500W model, with 17.6 dba ISO-chamber claim.

A note on the article: the link to Corsair HX520 & HX620 reviews has a typo.

jaganath wrote:

~El~Jefe~ wrote:Is spcr going to venture once again into fan swapping/controlling?

I remember how spcr people used to show how all one needed was an efficient % psu and then a swap of the fan would rectify much for low expense.

I think the Lian-Li's fan is good, it's just got a poor fan controller and poor efficiency. Any PSU can be modded with a quieter fan, but with low(er) efficiency PSUs heat can be a problem; and then the question becomes why buy a PSU that you have to mod when Seasonic/Corsair/etc are quiet in stock form?

I concur, spcr liked the fan when it was at minimum operating speed (under 150w), just the ramp up was insanely fast due to the fan controller/very poor efficiency. Sounds like this fan would have been great in a Seasonic, sadly the 600W version leaves much to be desired. I get the feeling the oem behind the PSU isn't exactly great at high efficiency devices yet. (similar to how channel well had no 80+ until very recently).

Lian- Li should stick to manufacturing aluminium chassis…period. That aside; a review on the Nexus 500W would be of great interest as to shed some light on their claims. Claims, which seem to be a tad bit on the overly optimistic side.

~El~Jefe~ wrote:Is spcr going to venture once again into fan swapping/controlling?

As a side note for this... I don't think a fan swap would be all that great of an idea for this PSU. The actual cooling hardware seemed below par for the rated power output (~30C rise at full load), so I'm not sure how well it would fare with a low speed fan at all times.

Was I the only one expecting the PSU housing to be made of aluminum?

Aluminium housing might have been expected if Lian-Li had been the manufacturer, which is not the case.

Shaman wrote:Was I the only one expecting the PSU housing to be made of aluminum?

I was also expecting a 300$ price tag, because design is cool...

walle wrote:Aluminium housing might have been expected if Lian-Li had been the manufacturer, which is not the case.

no pun intended?

I don't understand why you would need two power supplies together when one of these is more than enough.

Techreport has made numerous reviews checking power consumption. They've done it for 8800GTX in SLI with Core 2 Extreme X6800 (not the most power hungry processor family, but commonly used nowdays)
http://www.techreport.com/reviews/2007q ... ex.x?pg=14
Their system consumes 373W from the wall. Assuming 80% efficiency will leave you with a system drawing no more than 300W. So, one of the most powerful systems you can build today (except from using Core 2 Quad) will draw no more than that. Why do people keep moving from 600W to 800W, and from 800W to 1000W and on and on.

I find it hard that anyone would reach 500W even with quite heavy OC. 1000W? I dare anyone to show me a system that will overload a PSU that powerful! It would be rather interesting.

Ackelind wrote:I don't understand why you would need two power supplies together when one of these is more than enough.
...

It's not about real power needs, it's about "i/we have a bigger e-peen".
Haven't you heard something like "My computer is sooo powerfull, sooo overclocked, has sooo many light/fans/whatever that a 400/500/600/7zillion W PSU is not enough" (add some 1337 speech if u wish) million times?

Ackelind wrote:I find it hard that anyone would reach 500W even with quite heavy OC. 1000W? I dare anyone to show me a system that will overload a PSU that powerful! It would be rather interesting.

it's about marketing as well as e-penis

nvidia lists all these 750-1000w power supplies that are certified for SLI

http://www.slizone.com/object/slizone_build_psu.html

I'm having a hard time convincing people about realistic power draws at other forums. Maybe I just should give up and let "nature" have it's way. It's not my money that is spent in vain, but the result is that less cheap and good low-power PSUs are available.

Ackelind wrote:I don't understand why you would need two power supplies together when one of these is more than enough.

Techreport has made numerous reviews checking power consumption. They've done it for 8800GTX in SLI with Core 2 Extreme X6800 (not the most power hungry processor family, but commonly used nowdays)
http://www.techreport.com/reviews/2007q ... ex.x?pg=14
Their system consumes 373W from the wall. Assuming 80% efficiency will leave you with a system drawing no more than 300W. So, one of the most powerful systems you can build today (except from using Core 2 Quad) will draw no more than that. Why do people keep moving from 600W to 800W, and from 800W to 1000W and on and on.

I find it hard that anyone would reach 500W even with quite heavy OC. 1000W? I dare anyone to show me a system that will overload a PSU that powerful! It would be rather interesting.

A file-server with 20-30 disks would draw quite a bit from 12 V (at least on spin-up), In such a setup i think i would get a 700+ PSU.

(i now have only 10 disks, and on spin-up it draws 350 watts from the outlet)

AtW

I just found one setup that actually WOULD need a PSU like this.

AMDs new 4x4 gaming platform.
http://techreport.com/reviews/2007q1/qu ... dex.x?pg=1
Just read that review. The power consumption is horrible and still they only use one 7900GTX. With a pair of 8800GTX you will end up using just over 500W.

Adding hard drives won't increase the power consumption dramaticly if you are calculating using their spin-up power consumption. How often do you end up having your hard drives spin up (i.e. boot) while loading both your GPUs at the same time?

Ackelind wrote:I just found one setup that actually WOULD need a PSU like this.

AMDs new 4x4 gaming platform.
http://techreport.com/reviews/2007q1/qu ... dex.x?pg=1
Just read that review. The power consumption is horrible and still they only use one 7900GTX. With a pair of 8800GTX you will end up using just over 500W.

Adding hard drives won't increase the power consumption dramaticly if you are calculating using their spin-up power consumption. How often do you end up having your hard drives spin up (i.e. boot) while loading both your GPUs at the same time?

The fact is still that a modest machine (with the cheapest vga-card one can get) I have with 10 disks still draws 350 watts at its worst, which is quite a lot. 10-20 more disk would cause quite a big draw I think.

AtW