Actual difference from a 85% to a 90% PSU

[videobruce]

How much actual wattage drop with a given PC going from a 85% efficient PS to a 90% efficient rated around 600w? IOW' when measuring the AC current at the wall, using a 600w supply as the example, how much less wattage could I expect? 5, 10, 15 watts?

Is the extra $$ worth it?

[kalniel]

You'd expect it to drop ~5% of whatever you were using. If your system is using a typical 200W say, then a 90% efficient supply would pull 222W from the wall, and an 85% efficient supply would pull 235W, so you'd save 13W.

Worth depends on what matters to you.

[watercooled]

Also bear in mind these 80plus ratings are carried out at over 20% load, but idle power of a modern system can easily fall way below that mark, so depending on the build, a '90% efficient' PSU could in theory perform worse than a different '85% efficient' model when your system is idle. Some good review sites like JonnyGuru, Techpowerup, Hardwaresecrets, etc, test low-load efficiency to varying degrees.

[Distincthd]

One is more efficient than the other in terms of converting drawn out AC to DC (90% obviously being more efficient). You won't notice much difference short term, but in the long run you will be saving a bit more money.

[Mister]

The extra money is worth it, if your spending it on a good brand. It's all about the quality of the components.

Seasonic make a lot of the high end models for other Brands, so you always get top quality stuff from them. While e.g. Coolermaster often use re-branded cheap stuff and mark up the price.

[watercooled]

Claimed efficiency != quality.

And that's not necessarily true about Coolermaster, a lot of their stuff is made by Seasonic, CWT and Enhance. Just because it's made by one of those companies however, doesn't make it an overpriced rebrand - they're often tweaked and sometimes sell for less.

[Dottorrent]

Claimed efficiency != quality.

And that's not necessarily true about Coolermaster, a lot of their stuff is made by Seasonic, CWT and Enhance. Just because it's made by one of those companies however, doesn't make it an overpriced rebrand - they're often tweaked and sometimes sell for less.

That's a good point. XFX use Seasonic platforms and they modify and change various little components, add cables and redesign the style of the PSU, yet sell less than the OEM product it's based on. It's like VolksWagen, they own a few companies which make the same car (Skoda, Seat etc), those child companies modify the same car and sell for less.

[Saracen]

Claimed efficiency != quality.

And that's not necessarily true about Coolermaster, a lot of their stuff is made by Seasonic, CWT and Enhance. Just because it's made by one of those companies however, doesn't make it an overpriced rebrand - they're often tweaked and sometimes sell for less.

Quite right.

Just one point. Being "made by" doesn't necessarily mean made to the same standards, using the same components. That sirt of a deal may imply a contract where some product is made by, using a design from, a manufacturer, but where retail price is lower because cheaper componenrs have been used. Perhaps, for instance, instead of "belt and braces", you just get belt. Or, and this is what's hard to tell, you get a bit of old string.

As for well-known PSU names (Seasonic, for example) supplying other well-known names (like XFX) I wouldn't expect the 'old string' approach. But XFX might feel that 1% tolerance components is unnecessary and that 5% will do. Or they might just buy and rebadge.

It's hard to tell, unless you study and comoare two models, one of each brand, as some websites do.

On the 85% or 90% thing, the issues (5% of load, ish, and exactly how and how well that's measured) have already been mentioned. One other factor is the usage cycle of the PC. A PC on for an hour or two in the evening will benefit less than a business machine that's running 24/7. And, a machine used for aggressive gaming with a high-end card will draw more than a simple machine used for browsing the web.

All told, it's VERY hard to predict the outcome of 85% v 90%, because short of lots of analysis of usage, it's hard to know what load you use.

You could use a plug-in power meter and runyour PC through it for a week or two. That'll give you a rough guide to what your PC uses, and therefore costs to run. So IF you get a 5% saving, it'll put it in money terms.

[videobruce]

I decided to go with a Seasonic G-550 supply.

[Mister]

Quite right.

Just one point. Being "made by" doesn't necessarily mean made to the same standards, using the same components. That sirt of a deal may imply a contract where some product is made by, using a design from, a manufacturer, but where retail price is lower because cheaper componenrs have been used. Perhaps, for instance, instead of "belt and braces", you just get belt. Or, and this is what's hard to tell, you get a bit of old string.
...

This is almost always the case :/

You may be ok and the compromise never causes an issue. But then a week after the warranty expires it might explode when the OEM version would not have done.

If you have ever worked in a large office that uses e.g. Dell PCs, you will have seen the problem of using cheap components over a long period of time. After about 2-3 years half of them die (or get other build quality related problems) and it ends up costing you more than if you had just spent the extra money on quality stuff

[watercooled]

I decided to go with a Seasonic G-550 supply.

Decent choice.

This is almost always the case :/

You may be ok and the compromise never causes an issue. But then a week after the warranty expires it might explode when the OEM version would not have done.

I wouldn't go that far, there are plenty of MFRs using other OEMs like Delta, Enhance, Seasonic, Flextronics, etc, producing excellent quality products. Take Corsair as an example; sure some of their cheaper units are of questionable quality now, but their high-end stuff is v.good. Also Antec are producing some class-leading products OEM'd by Delta.

Loosening the tolerance on a few components doesn't mean it's going to explode. And looser tolerances doesn't necessarily mean poorer quality either, just the value of the component in question isn't guaranteed to be as close to the label value. For example, changing from a 1% to an equivalent 5% filtering/holdup cap really shouldn't make any appreciable difference to noise/stability/reliability. Neither does a 5% tolerance guarantee the component will be further from the spec than a 1% component - in fact it's quite possible the 5% cap in the 'cheaper' PSU could be .01% from spec, while the 1% cap in the 'premium' product is in fact 1% away. 'Cheaper components' doesn't mean much by itself - and it might not even be a problem depending on what exactly they are.

Seasonic are A good retail brand, but they're not the only one. Neither is stuff made by them for other MFRs necessarily inferior.

And, a lot of people overlook the fact they're powering a PC, not a transatlantic fibre node; for example, PCs shouldn't really care about super-tight voltage regulation. TBH it would be nice to see more effort put into the low-output market i.e. sub-200W, where most desktop PCs (not including high-end OFC) will fit these days, most far less than that. Using something like a 700W PSU to power such a system tends to be horribly inefficient, even if it is '80Plus Platinum' - that badge doesn't give any indication of efficiency where PCs will spend most of their time i.e. idle. /rant