How are EVGA PSUs these days?

[Charlie98]

I'm in the market for a basic 430w PSU for backup and testing purposes, I usually pick up a Corsair CX430 on sale but I haven't seen one lately... but there is this EVGA 430w unit on NE today. It's not even bronze rated, which doesn't really matter, but I'm curious... Who makes the EVGA entry-level PSU, and are they at least as good as a CX?

 

[EXCellR8]

I've used a few EVGA branded power supplies and I've got mostly good things to say about them. I haven't used that particular model, nor the CX430 as I typically require much more power... but at that price you'll get what you pay for. I'm confident it would probably be just fine for backup and testing purposes (depending on what you're using for hardware). Plus it's $20 after the rebate so you can't really go wrong.

 

[Charlie98]

Actually, I think I answered my own question... HardOCP did a review on that model, they didn't care for it very much.

I think what I'm going to do is get a nice SeaSonic to replace the CX430 in my HTPC, and then use the CX as my spare/tester.

 

[MrTeal]

Good choice.

A really good resource for finding PSU platforms are the following sites.
http://www.realhardtechx.com/index_archivos/Page2293.htm
http://www.orionpsudb.com/platforms
The realhardtech site just lists the OEM and not the platform, but it's usually really up to data and has a lot of links to review. The PSU Platform Database site is really nice since not only does it show the OEM, but also the platform and what other PSUs might use that platform. It's usually not as complete though.

That 430W EVGA is a HEC 500B model, so it's the same basic platform as the EVGA 500B as well. Jonnyguru reviewed it as well, and didn't give it too rough a treatment. It missed even the pathetic 80 Plus standard targets (though the CX430 doesn't either), but it's dirt cheap so the review is in light of what the PSU is.

 

[EXCellR8]

Good resources, thanks.

when it comes down to OEMs I think Seasonic is one of the best choices you can go with. wasn't aware that they actually manufactured some EVGA branded models.

 

[Charlie98]

Good resources, thanks.

when it comes down to OEMs I think Seasonic is one of the best choices you can go with. wasn't aware that they actually manufactured some EVGA branded models.

I lucked out when I got my first PSU... a Corsair TX750v2... a SeaSonic unit.

That's why I was asking... I've seen some recommendations for EVGA PSU's and was curious how far down the model line it went.

 

[Bearmann]

Many of the EVGA PSU's are truly superior, while others are not so great. You can get an idea of the quality of various PSU's from this tier list:

http://www.tomshardware.com/forum/id-2547993/psu-tier-list.html

 

[DigDog]

if you are using 250w, you can find the EVGA 650W G2 which has an efficiency of 90% at half load, it's eco (fan won't spin) and considering the practically irrelevant load, it should last 10 years. And you get a top notch warranty. (fully modular)
for 200W, the 550w is a better choice.

 

[Charlie98]

if you are using 250w, you can find the EVGA 650W G2 which has an efficiency of 90% at half load, it's eco (fan won't spin) and considering the practically irrelevant load, it should last 10 years. And you get a top notch warranty. (fully modular)
for 200W, the 550w is a better choice.

If I wind up swapping out the HTPC PSU, I won't even need that... It'll be lucky to hit 100w under normal load. (Come to think of it, I need to throw the Kill-a-Watt on it just to see... )

 

[DigDog]

if you have a 100w system, you need to pay special attention to the efficiency ratio of whatever psu you will be using. many will struggle at such low load. i'm sorry but i do not have experience in this matter so i can't advise.

 

[MrTeal]

The fanless Seasonic 400FL2 would make a great HTPC PSU. Not only is it (obviously) fanless, but the low load efficiency is really good. It's still over 90% at 86W DC load, and standby and low load efficiency is also very good. It's quite pricey though. The SSR-360GP would be a good alternate, you can pick it up for ~$60 shipped and it gives you all the quality to expect from Seasonic, like Japanese caps, etc.

 

[Charlie98]

I put the KoW on the HTPC yesterday, it runs at 110w at normal cruise, that is to say, streaming Netflix or stored movies, but it spikes to over 200w at boot and such. I currently have a GTX560Ti in it, so that doesn't help; once I switch it to a GTX9xx, that will knock off a few watts.

It used to have a Pentium G620, but I upgraded it to that nice i3... I don't think I've measured wattage since the Pentium. Interesting, I don't remember it running over 100w.

EDIT: Woah! Running Valley pumps it up to 370w! Yow! Better stick with the 430w...

 

[MrTeal]

I put the KoW on the HTPC yesterday, it runs at 110w at normal cruise, that is to say, streaming Netflix or stored movies, but it spikes to over 200w at boot and such. I currently have a GTX560Ti in it, so that doesn't help; once I switch it to a GTX9xx, that will knock off a few watts.

It used to have a Pentium G620, but I upgraded it to that nice i3... I don't think I've measured wattage since the Pentium. Interesting, I don't remember it running over 100w.

EDIT: Woah! Running Valley pumps it up to 370w! Yow! Better stick with the 430w...

Yeah, that's the 560Ti. At those power levels the CX430 is a hair over 80% efficient, so you're looking at a DC load of ~300W. Techpowerup measures the actual DC into the card and they listed a max of 202W for a stock unit. Given the efficiency of the CX430, that card could suck back ~250W at the wall. Under the same conditions, a GTX 960 only pulled 129W, so it's a substantial difference.

 

[Charlie98]

Yeah, that's the 560Ti. At those power levels the CX430 is a hair over 80% efficient, so you're looking at a DC load of ~300W. Techpowerup measures the actual DC into the card and they listed a max of 202W for a stock unit. Given the efficiency of the CX430, that card could suck back ~250W at the wall. Under the same conditions, a GTX 960 only pulled 129W, so it's a substantial difference.

I really want a 950 or 960, but they are such a terrible deal right now... :'(

 

[MrTeal]

I really want a 950 or 960, but they are such a terrible deal right now... :'(

LOL... I didn't even notice in your sig that the 560Ti is the 448 core GF110 version. That beastie can pull 263W DC under Furmark in TPU's testing. That's more than twice as much as a 960. It is too bad that they're so expensive right now, but it might be worth it just to get your HTPC power and noise down.

 

[lehtv]

Yeah, that's the 560Ti. At those power levels the CX430 is a hair over 80% efficient, so you're looking at a DC load of ~300W. Techpowerup measures the actual DC into the card and they listed a max of 202W for a stock unit. Given the efficiency of the CX430, that card could suck back ~250W at the wall. Under the same conditions, a GTX 960 only pulled 129W, so it's a substantial difference.

A couple of things wrong with that.

1. You're comparing 560 Ti's Furmark power consumption to GTX 960's Metro: LL power consumption:

NVIDIA 560 Ti (Techpowerup)
Asus 960 Strix (techpowerup)

2. You're comparing a stock 560 Ti vs a factory overclocked GTX 960. Always compare stock vs stock or OC vs OC if you want to shed light on generational differences. Here's the Asus 560 Ti DC II review which shows a bit higher readings than the stock NVIDIA one, and as such is a better point of comparison to Asus 960 Strix.

3. Furmark creates an artificially high load that doesn't represent reality in any way. Also, newer generations of graphics cards have stricter and smarter TDP limits in place which restrict their power use in Furmark type tests. This makes Furmark useless for comparing actual power consumption as it makes older cards look much worse than they actually are.

Techpowerup Maximum number means the single highest reading during Furmark, while the Peak number means the single highest reading during a game or gaming benchmark.

Peak vs peak:
Asus 560 Ti (3Dmark03) - 157W
Asus 960 (Metro:LL) - 129W
-> 28W difference

Maximum vs maximum (Furmark)
Asus 560 Ti - 228W
Asus 960 - 147W
-> 81W difference, or a +189% error compared to real world use

4. Usually we're not interested in peak wattage, but average load wattage, when talking about from-the-wall power consumption. AC watts have to do with running costs, not with reliability. So to find out how much more power a 560 Ti is drawing from the wall compared to GTX 960, with CX430 V2 as the power supply, let's use the Average chart in Techpowerup's reviews:

Asus 560 Ti (3Dmark03) - 137W
Asus 960 (Metro:LL) - 114W

Efficiency is 83%.

137W / 0.83 = 165W
114W / 0.83 = 137W
-> 28W difference from the wall.

Of course, none of this directly translates to advice for Charlie98 since he's using a 560 Ti 448 Cores which is closer to GTX 570 level power consumption.

 

[Charlie98]

I was hoping you would be along, Lehtv...

I've never fooled with Furmark, maybe I'll take a look at it, and I know there is a difference in wall draw vs what's actually leaving the PSU. I know the GPU makes the biggest difference... thinking back, when I had my Pentium in I also has a passively cooled HD6450, and it only drew about 40w (at the wall) during use.

TBH, I expected the 9xx series cards to be much more efficient than a 3rd Gen older 560... and, yes, it's a 448. One of these days, when it's raining and I'm holed up in the house, I'll have to bench all 3 of my GTXs in my GAME rig (apples to apples to apples, as much as I can...)

 

[MrTeal]

A couple of things wrong with that.

1. You're comparing 560 Ti's Furmark power consumption to GTX 960's Metro: LL power consumption:

NVIDIA 560 Ti (Techpowerup)
Asus 960 Strix (techpowerup)

2. You're comparing a stock 560 Ti vs a factory overclocked GTX 960. Always compare stock vs stock or OC vs OC if you want to shed light on generational differences. Here's the Asus 560 Ti DC II review which shows a bit higher readings than the stock NVIDIA one, and as such is a better point of comparison to Asus 960 Strix.

3. Furmark creates an artificially high load that doesn't represent reality in any way. Also, newer generations of graphics cards have stricter and smarter TDP limits in place which restrict their power use in Furmark type tests. This makes Furmark useless for comparing actual power consumption as it makes older cards look much worse than they actually are.

Techpowerup Maximum number means the single highest reading during Furmark, while the Peak number means the single highest reading during a game or gaming benchmark.

Peak vs peak:
Asus 560 Ti (3Dmark03) - 157W
Asus 960 (Metro:LL) - 129W
-> 28W difference

Maximum vs maximum (Furmark)
Asus 560 Ti - 228W
Asus 960 - 147W
-> 81W difference, or a +189% error compared to real world use

4. Usually we're not interested in peak wattage, but average load wattage, when talking about from-the-wall power consumption. AC watts have to do with running costs, not with reliability. So to find out how much more power a 560 Ti is drawing from the wall compared to GTX 960, with CX430 V2 as the power supply, let's use the Average chart in Techpowerup's reviews:

Asus 560 Ti (3Dmark03) - 137W
Asus 960 (Metro:LL) - 114W

Efficiency is 83%.

137W / 0.83 = 165W
114W / 0.83 = 137W
-> 28W difference from the wall.

Of course, none of this directly translates to advice for Charlie98 since he's using a 560 Ti 448 Cores which is closer to GTX 570 level power consumption.

1. No, I'm not. You linked the Strix overclocked version. I used the reference 960 values found in the 950 review. Not sure why they included the reference 960 peak consumption in the Strix review but not the max.

2. No, I didn't.

3. True, this is a little more questionable. While I'd normally agree with you, the Peak test changed between the two reviews while they continued to use the same max test between the two.

4. He measured 370W at the wall. JG measured 82% at 313W AC and 81% at 423W AC. I split the difference and used 81.5%.

 

[lehtv]

1 & 2. Fair enough. However, the incredibly low Maximum consumption of the reference GTX 960 just reinforces the point about how Furmark overstates the differences in real world power consumption between new and old cards.

3. Your agreeing with the point about Furmark misleading shouldn't depend on imagined differences in power consumption between 3Dmark03 and Metro:LL. Sure, there might be differences between those, but they both still represent real world power consumption pretty well, and it's still true that Furmark creates loads that are just not realistic.

4. Sure, I wasn't criticizing that. I just wanted to get a somewhat precise estimate for the difference in actual (AC) power use between GTX 560 Ti and GTX 960. Your post made it seem like there's a massive difference when really there isn't - although as you later said, you didn't notice Charlie98 had a 560 Ti 448 cores which is on another level in terms of its power requirements.

 

[Charlie98]

Although I'm enjoying the conversation, and, please, keep it up... I learn something new every day...

I ordered the 600w EVGA 600B this morning, $35 (AR.) After reading the reviews on the EVGA 430w, I didn't have a good feeling about it. As it turned out, the NE ad this AM had both the $20 (AR) CX430 and this 600B... so I spent the extra money on the EVGA, figuring it could sub for any of my PSU's if a problem came up. The one review I read on it looked pretty good for a budget PSU, and better than the 500B.

 

[ElFenix]

how is the 7950 130 watts lower than the 7970? that doesn't make sense.

 

[lehtv]

@ElFenix

It's Furmark, don't expect it to make sense 7950 seems to have a much tighter TDP limiter in place.