"In fact, the efficiency is above 80% over almost the entire range which can not only save money on electrical costs but is also generally a good indicator of power supply quality. "
"This isn't an ideal result, but at the same time PFC isn't necessarily one of the most critical factors in determining power supply quality."
As far as I know, a PFC of ~0.92 means 8% more will actually be billed to me by the provider. Thus, the good efficiency of over 80% does not help at all.
Meeting the 80% efficiency does still help, but you are pointing out one of the interesting developments these days, how PSU manufacturers are tweaking to arrive at higher efficiency and that within the context of expectations of how the industry (reviewers et al) will review, particularly when it's a retail product. Robbing Peter to pay Paul is ideally bad, but at any given moment and price, can be subjectively more or less important depending on your needs.
The important part is that this information was revealed so you can decide for youself if this unit meets those needs or if you'd rather some other compromise. No PSU is perfect in every way including price.
I'd like to see where the x-pro stands to see if I should upgrade or not :) got an ETA on that? All over the forums for the past 5 months that PSU has been on banners :P
Great review, though perhaps Anandtech should consider gearing a small percentage of articles to those who aren't technomaniacs. BTW - how do you pronouce "Anandtech"?
I'll take the second part - I think there might be some less-techy articles on occasion, but the PSU stuff definitely doesn't qualify.
Anyway, for those interested, it's pronounced Ahn-Ahnd-Tech. Or "On Ond Tech". So if you pronounce it with a nice southern twang and an "A as in apple" sound, Anand might make weird faces at you. That, or I need to check my hearing and make sure Anand isn't saying, "Hi guys, it's '&n - &n(d)" as opposed to "Hi guys, it's 'än - änd." (When did http://en.wikipedia.org/wiki/International_Phoneti...">phonetic symbols get so stinking complex?)
If you take a DMM and measure the power drop on the actual molex connectors, and not take the results directly from the Chroma how does it look then?
I suspect you have an exponential increase in resistance which causes the Chroma to display incorrect voltage values... (Because of the cable length from the PSU's connectors and up to your load, including interface boards)
Yea, but the voltage resistance issue is something that has been the same for all PSU reviews here at Anand
When you have this problem with all PSU's you need to realize there is something wrong with your testing equipment, sorry for being so blunt... (Especially since none of the other 2 big sites report the same)
And yes, some scope readings for this DEER PSU sure would have been interesting (just to make sure to beat the dead horse a bit more)
Strange though that the Zippy G1 has nearly the whole time close to to 12.0v. I have seen reports from the companies and they look similar (also the high efficiency) and thus I don't think the resistance will be a big issue.
Are they using a Chroma as well for those test reports? Or are they reading directly at the connector without another interface like the spec calls for?
Christoph Katzer; The issue is of course not that you are using the Croma, it's a great unit; however, the way you use it will most likely result in incorrect voltage readings
Due to the fact that the resistance in the cables loading the unit will most likely result in a exponential increase in resistance, therefore the results shown by the Chroma will be incorrect, and more incorrect as the load increases and the resistance exponentially increases...
Just putting a multimeter on an unloaded Molex connector, or, directly on the molex connector you are loading (and not further down where the chroma reads the voltage) would quickly prove or disprove my theory
You are correct, that a high enough current on too low a wire gauge does cause significant voltage drop, I have observed it many times myself particularly with some of the poorer PSU using 12V connectors with less than 18 ga. wires.
However, similarly we can't just take the reading from an unused molex connector instead, as a PSU is spec'd to provide it's voltages at the load through the existing wiring harness. It is not necessary to try to evenly distribute that load across all the wire pairs in that harness as that is a practically impossible scenario for implementation running a system, so a bit of a derating factor is needed to appoximate the typical expected loads. IMO, a good start would be loading each supply wire at about 6A (not counting ground returns) up until the rating per rail is met, leaving some supply wires per rail unused when (sum of rail wires * 6A) > rated current per rail. Obviously some connectors and leads are more robust and necessary than others, for example a floppy connector should just be ignored while the 2 x n 12V CPU connector should always be used.
But wouldn't regular PC components drawing power from a PSU result in the same sort of increase in resistance? (Note: I'm not at all an electrician, so I could be wrong. Just asking a question.)
Yes and no. The additional interface and the extreme load a load tester puts accross a single connector magnifies the problem beyond what a system would see happen which is what people were asking about.
to see some Seasonic PSU tests. Not sure *why* but I cannot remember seeing any Seasonic PSU test from you, or any other sites I frequent, and they are supposedly the manufactuers of several of the PSUs for other companies that have good PSUs.
Also, it would be very nice to have an article once in a while that explained where the different PSUs were manuafactuered, who actually makes them, and what parts each company typically uses for various models. Anyone can write a review, but no one seems todo this. Be the first ! I *could* probably scour the web to find this information, but if it comes from your guys, I could probably trust the information ;)
A Seasonic 500w PSU review was done very recently by another ]H[ard-core review site. You might want to check it out. They and JonnyGURU as well are very clear about who makes PSU's, capacitor brands in them, etc.
Seasonic makes PSU's for a number of companies --the Antec Neo HE line is one example, but they do some others as well.
As for the Silver Power, if Anandtech's weblink is correct to MWave (it is partially broken) and that PSU is available for $69, it's a great price.
The web links are just pulled from the engine automatically (automagically even) based on some search text. I put in "Tagan" and that's what came up. Still, the 530W Tagan is probably pretty decent for $69. This PSU is apparently $88 shipped from http://www.newegg.com/Product/Product.aspx?Item=N8...">Newegg (see above comments).
heh Jarred, read the first user review on newegg ... not exactly a good example of a PSU 'review'. I probably would take it with a grain of salt *if* the overall user review % was not what it is.
Yeah, Newegg reviews are basically random people spouting off. I've had numerous good quality PSUs fail over the years, and I've had "crappy" PSUs that are still kicking after four years. Without some clear details about the load the PSU was under and the operating environment - let's be honest, overclocking can kill a lot of PSUs if you push it too far - there's any number of reasons a PSU can fail. $90 for a >80% 600W PSU is really pretty good.
It appears that a log of people on newegg have this PSU fail after a few months. Now before we get into newegg consumer feedback 'reliability' :) there are a few people who've had these fail, not just one or two.
I'm ganna give it a shot anyway. Should work well with my DFI Infinity 975 board, being EPS12v and all.
Knowing it is an Apex-built power supply took most of the wind out of its sails. Apex, also known as L&C and Deer, has made some of the most unreliable power supplies ever.
This seems to be one of their better efforts, but reliability will always be a concern with this company.
SMPS tech is reasonably mature beyond present tweaking a bit for higher efficiency, more 12V current, and these without cost rising out of control.
Point being, Apex (actually better known as Foxconn or Hon Hai) can easily build quality PSU, it is not inability it is the choice of product placement and construction cost that results in some of the junk we've all seen.
As for reliability, there is a problem as always that you hear of someone with a failed PSU but no autopsy most of the time. No disrespect meant to reviewers, but over the years I can't could how many times a product seemed good at first glance, and second glance (a review period), but later a fault compromised the lifespan. A review of one unit can't take forever, it is going to be inherantly limited in scope, but still must be seen as a way to disqualify products more than qualify them for long term use.
We’ve updated our terms. By continuing to use the site and/or by logging into your account, you agree to the Site’s updated Terms of Use and Privacy Policy.
Facebook
Twitter
RSS
33 Comments
Back to Article
meyergru - Tuesday, August 28, 2007 - link
...how do those two statements add up?"In fact, the efficiency is above 80% over almost the entire range which can not only save money on electrical costs but is also generally a good indicator of power supply quality. "
"This isn't an ideal result, but at the same time PFC isn't necessarily one of the most critical factors in determining power supply quality."
As far as I know, a PFC of ~0.92 means 8% more will actually be billed to me by the provider. Thus, the good efficiency of over 80% does not help at all.
13Gigatons - Tuesday, September 4, 2007 - link
PFC intended purpose is to turn a complex load into a simple one.http://en.wikipedia.org/wiki/Power_factor_correcti...">http://en.wikipedia.org/wiki/Power_factor_correcti...
mindless1 - Thursday, August 30, 2007 - link
Meeting the 80% efficiency does still help, but you are pointing out one of the interesting developments these days, how PSU manufacturers are tweaking to arrive at higher efficiency and that within the context of expectations of how the industry (reviewers et al) will review, particularly when it's a retail product. Robbing Peter to pay Paul is ideally bad, but at any given moment and price, can be subjectively more or less important depending on your needs.The important part is that this information was revealed so you can decide for youself if this unit meets those needs or if you'd rather some other compromise. No PSU is perfect in every way including price.
Christoph Katzer - Tuesday, August 28, 2007 - link
What was the relation of the both again?Your provider charges reactive power?
swtethan - Tuesday, August 28, 2007 - link
I'd like to see where the x-pro stands to see if I should upgrade or not :) got an ETA on that? All over the forums for the past 5 months that PSU has been on banners :PMissPriss - Monday, August 27, 2007 - link
Great review, though perhaps Anandtech should consider gearing a small percentage of articles to those who aren't technomaniacs. BTW - how do you pronouce "Anandtech"?JarredWalton - Monday, August 27, 2007 - link
I'll take the second part - I think there might be some less-techy articles on occasion, but the PSU stuff definitely doesn't qualify.Anyway, for those interested, it's pronounced Ahn-Ahnd-Tech. Or "On Ond Tech". So if you pronounce it with a nice southern twang and an "A as in apple" sound, Anand might make weird faces at you. That, or I need to check my hearing and make sure Anand isn't saying, "Hi guys, it's '&n - &n(d)" as opposed to "Hi guys, it's 'än - änd." (When did http://en.wikipedia.org/wiki/International_Phoneti...">phonetic symbols get so stinking complex?)
SemiCharmed - Monday, August 27, 2007 - link
I agree with MissPriss. It could be called "NotsoAnandtech"DividedweFall - Monday, August 27, 2007 - link
Hoorah for MissPriss! I don't send comments in fear of being rejected by the eleet technomanic crowd.Samus - Monday, August 27, 2007 - link
We readers at Anandtech are loyal to this place BECAUSE the reviews are for technomaniacs. They're among the most thorough reviews of hardware around.Go look at any other site's PSU reviews and you'll see what I mean.
Per Hansson - Monday, August 27, 2007 - link
If you take a DMM and measure the power drop on the actual molex connectors, and not take the results directly from the Chroma how does it look then?I suspect you have an exponential increase in resistance which causes the Chroma to display incorrect voltage values... (Because of the cable length from the PSU's connectors and up to your load, including interface boards)
Sincerely - Per Hansson
MrOblivious - Monday, August 27, 2007 - link
Well that and if this really is a Solytech (Deer) it could just be a flaming hunk of crap.Per Hansson - Monday, August 27, 2007 - link
Yea, but the voltage resistance issue is something that has been the same for all PSU reviews here at AnandWhen you have this problem with all PSU's you need to realize there is something wrong with your testing equipment, sorry for being so blunt... (Especially since none of the other 2 big sites report the same)
And yes, some scope readings for this DEER PSU sure would have been interesting (just to make sure to beat the dead horse a bit more)
Christoph Katzer - Tuesday, August 28, 2007 - link
Strange though that the Zippy G1 has nearly the whole time close to to 12.0v. I have seen reports from the companies and they look similar (also the high efficiency) and thus I don't think the resistance will be a big issue.MrOblivious - Tuesday, August 28, 2007 - link
Are they using a Chroma as well for those test reports? Or are they reading directly at the connector without another interface like the spec calls for?Christoph Katzer - Tuesday, August 28, 2007 - link
"Every" company in Taiwan uses Chroma for their own evaluations.Per Hansson - Tuesday, August 28, 2007 - link
Christoph Katzer; The issue is of course not that you are using the Croma, it's a great unit; however, the way you use it will most likely result in incorrect voltage readingsDue to the fact that the resistance in the cables loading the unit will most likely result in a exponential increase in resistance, therefore the results shown by the Chroma will be incorrect, and more incorrect as the load increases and the resistance exponentially increases...
Just putting a multimeter on an unloaded Molex connector, or, directly on the molex connector you are loading (and not further down where the chroma reads the voltage) would quickly prove or disprove my theory
mindless1 - Thursday, August 30, 2007 - link
You are correct, that a high enough current on too low a wire gauge does cause significant voltage drop, I have observed it many times myself particularly with some of the poorer PSU using 12V connectors with less than 18 ga. wires.However, similarly we can't just take the reading from an unused molex connector instead, as a PSU is spec'd to provide it's voltages at the load through the existing wiring harness. It is not necessary to try to evenly distribute that load across all the wire pairs in that harness as that is a practically impossible scenario for implementation running a system, so a bit of a derating factor is needed to appoximate the typical expected loads. IMO, a good start would be loading each supply wire at about 6A (not counting ground returns) up until the rating per rail is met, leaving some supply wires per rail unused when (sum of rail wires * 6A) > rated current per rail. Obviously some connectors and leads are more robust and necessary than others, for example a floppy connector should just be ignored while the 2 x n 12V CPU connector should always be used.
mindless1 - Thursday, August 30, 2007 - link
I meant, high enough current on too high, too small a wire gauge.JarredWalton - Tuesday, August 28, 2007 - link
But wouldn't regular PC components drawing power from a PSU result in the same sort of increase in resistance? (Note: I'm not at all an electrician, so I could be wrong. Just asking a question.)MrOblivious - Tuesday, August 28, 2007 - link
Yes and no. The additional interface and the extreme load a load tester puts accross a single connector magnifies the problem beyond what a system would see happen which is what people were asking about.I am sure Chris is looking into it ;)
yyrkoon - Monday, August 27, 2007 - link
to see some Seasonic PSU tests. Not sure *why* but I cannot remember seeing any Seasonic PSU test from you, or any other sites I frequent, and they are supposedly the manufactuers of several of the PSUs for other companies that have good PSUs.Also, it would be very nice to have an article once in a while that explained where the different PSUs were manuafactuered, who actually makes them, and what parts each company typically uses for various models. Anyone can write a review, but no one seems todo this. Be the first ! I *could* probably scour the web to find this information, but if it comes from your guys, I could probably trust the information ;)
Axbattler - Thursday, August 30, 2007 - link
Silent PC Review has made some reviews of Seasonic (and the Corsair too if I remember right).LoneWolf15 - Monday, August 27, 2007 - link
A Seasonic 500w PSU review was done very recently by another ]H[ard-core review site. You might want to check it out. They and JonnyGURU as well are very clear about who makes PSU's, capacitor brands in them, etc.Seasonic makes PSU's for a number of companies --the Antec Neo HE line is one example, but they do some others as well.
As for the Silver Power, if Anandtech's weblink is correct to MWave (it is partially broken) and that PSU is available for $69, it's a great price.
JarredWalton - Monday, August 27, 2007 - link
The web links are just pulled from the engine automatically (automagically even) based on some search text. I put in "Tagan" and that's what came up. Still, the 530W Tagan is probably pretty decent for $69. This PSU is apparently $88 shipped from http://www.newegg.com/Product/Product.aspx?Item=N8...">Newegg (see above comments).yyrkoon - Monday, August 27, 2007 - link
heh Jarred, read the first user review on newegg ... not exactly a good example of a PSU 'review'. I probably would take it with a grain of salt *if* the overall user review % was not what it is.JarredWalton - Monday, August 27, 2007 - link
Yeah, Newegg reviews are basically random people spouting off. I've had numerous good quality PSUs fail over the years, and I've had "crappy" PSUs that are still kicking after four years. Without some clear details about the load the PSU was under and the operating environment - let's be honest, overclocking can kill a lot of PSUs if you push it too far - there's any number of reasons a PSU can fail. $90 for a >80% 600W PSU is really pretty good.Christoph Katzer - Monday, August 27, 2007 - link
Seasonic comes up this week.Don't worry about the rest, that'll come as well...
Samus - Monday, August 27, 2007 - link
Looks identical. Newegg has it for $80.Chris, can you confirm whether this is the same power supply as the silverpower?
Christoph Katzer - Monday, August 27, 2007 - link
Thanks for the info. It is indeed the same PSU.Samus - Monday, August 27, 2007 - link
It appears that a log of people on newegg have this PSU fail after a few months. Now before we get into newegg consumer feedback 'reliability' :) there are a few people who've had these fail, not just one or two.I'm ganna give it a shot anyway. Should work well with my DFI Infinity 975 board, being EPS12v and all.
Slaimus - Wednesday, August 29, 2007 - link
Knowing it is an Apex-built power supply took most of the wind out of its sails. Apex, also known as L&C and Deer, has made some of the most unreliable power supplies ever.This seems to be one of their better efforts, but reliability will always be a concern with this company.
mindless1 - Thursday, August 30, 2007 - link
SMPS tech is reasonably mature beyond present tweaking a bit for higher efficiency, more 12V current, and these without cost rising out of control.Point being, Apex (actually better known as Foxconn or Hon Hai) can easily build quality PSU, it is not inability it is the choice of product placement and construction cost that results in some of the junk we've all seen.
As for reliability, there is a problem as always that you hear of someone with a failed PSU but no autopsy most of the time. No disrespect meant to reviewers, but over the years I can't could how many times a product seemed good at first glance, and second glance (a review period), but later a fault compromised the lifespan. A review of one unit can't take forever, it is going to be inherantly limited in scope, but still must be seen as a way to disqualify products more than qualify them for long term use.