^ Looks like $270 AR, $260 AP to me
![]()
:thumbsup: Very nice! I'm guessing AMD is trying to draw down stock in preparation for the R9 390(?) launch.
^ Looks like $270 AR, $260 AP to me
![]()
I think this is the first time the midrange build has featured an 80 Plus Platinum rated PSU.
I feel like you didn't really do your research on this one, mfenn. I can only facepalm at Antec's decision to back up a Platinum unit with only 3 year warranty. This is no doubt a result of cheaping out on the capacitors - including a CapXon 85°C on the primary side, and another CapXon (among multiple Japanese ones) on the secondary side. Remember, CapXons are third tier in JonnyGuru's book, even worse than Teapos (which aren't bad but lose to Japanese ones in reputation).
Also, for anybody wondering if there's a catch with that Asus R9 290X being so cheap... check TPU's review. It either runs quiet or cool - not both. Quiet means 94°C load temp (yes, really - it will almost throttle).
I would bite the bullet and grab an XFX R9 290X DD for $302 AR AP. Lifetime warranty. Or a Zotac 970 for $307 and without a rebate hassle. Also lifetime warranty with registration.
This argument is the sort of fear, uncertainty, and doubt line of reasoning that you see repeated over and over in PSU, but for which you never see any sort of real technical backup. Those arguments are echoes of the Taiwanese capacitor plague of 2002, but really lack any justification today.
The reason that Antec used 85C caps is because when you are working with a high efficiency, you simply don't need huge temperature ratings.

So I stand behind this recommendation 100%. There's certainly nothing wrong with the other units that you linked of course. The TP-750C in particular dropped in price last night and is definitely at good unit at $60 AR.
Characterizing the ASUS R9 290X DCu II's noise profile as out of line with other standard R9 290X's is incorrect. It's 42 dBA in the "performance" profile. Compare it to the MSI R9 290X Lightning, PowerColor PCS+ R9 290X and MSI R9 290X Gaming and it matches or beats them.
PowerColor PCS+ 63
Gigabyte WF3 OC 71
XFX DD 72
Asus DCII, perf 78
Asus DCII, silent 94
Asus DCII, silent 43.2
PowerColor PCS+ 45.9
Gigabyte WF3 OC 47
Asus DCII, perf 51.8
XFX DD 52.5
No, the reason was that Antec wanted to differentiate from others' Platinum units by offering a cheaper alternative built from cheaper parts and from a sub-par OEM.
If as you point out, they're trying to differentiate as a platinum unit for less cost, then that 3 years of extra warranty (which has to be included in the price of every product) is eating into the margins on being a lower-cost platinum alternative.Why not show customers that you have confidence in your own product by backing it up with proper 5-7 year warranty?
JG has sort changed on this recently also. See recent reviews of the CWT's new build of the Corsair RM850 and Andyson's N700 TitaniumNot that I don't disagree with you about the capacitors - I'd rather defer to JonnyGuru than you on matters of PSU components
Think about that number. Seriously. Fifteen years is a boatload of time... longer than the lifespan of most electrolytic capacitors, regardless of whose factory they came out of. The big Chinese names have had time now to iron the wrinkles out, and technology has improved. The whole thing started with a formula that got copied down wrong, and that formula likely hasn't been used in ages by the major Chinese brands. Ltec, Teapo, OST, Capxon... when was the last time any of you saw one of these fail? Wait, let me answer that. The last ones I saw failed were the OSTs on one of my Asrock 939Dual-VSTA motherboards. Those came out in 2005. Ten years ago, when I was just getting into my thirties. Folks, you don't stay in business by not addressing issues like this.
...
I am no longer scoring against Teapo electrolytics, let alone their polymers, so I have no complaints here. Seriously, Teapo has to be the widest used Chinese brand out there in these things. And I'm not just talking about second string OEMs like CWT, either. No, I've seen their stuff in Win-Tact, Etasis, Zippy, and Delta units as well. If those four don't have a problem using Teapo, neither do I.
If as you point out, they're trying to differentiate as a platinum unit for less cost, then that 3 years of extra warranty (which has to be included in the price of every product) is eating into the margins on being a lower-cost platinum alternative.
It's highly speculative to map directly from length-of-warranty to quality-of-product in the commodity market.
JG has sort changed on this recently also. See recent reviews of the CWT's new build of the Corsair RM850 and Andyson's N700 Titanium
For example:
That's interesting, thanks. Does this mean JonnyGuru also stops scoring against CapXons?
No, the reason was that Antec wanted to differentiate from others' Platinum units by offering a cheaper alternative built from cheaper parts and from a sub-par OEM. You get what you pay for, though - as we've seen, an equally inexpensive Gold unit built by a better OEM from high end parts is actually more efficient and also better in every other way. Antec tried to paint a picture that Platinum doesn't have to mean expensive, but what they forgot to paint was the part where Platinum doesn't have to mean quality either.
Also, if a Platinum unit is perfectly fine with 85°C caps, then why have only 3 year warranty? Why not show customers that you have confidence in your own product by backing it up with proper 5-7 year warranty?![]()
I didn't characterize its noise profile out of line, I was talking about thermal efficiency which is better on almost all other aftermarket 290X's. This allows all other 290X's to achieve both quiet and cool operation, whereas Asus 290X DCII is one or the other. Cards like Lightning, Gaming, PCS+ and XFX DD don't have to heat up to reference 94°C to acheive quiet operation.
If you're intending to maximize value for money with this $1000 build, EA-650 Platinum is not the best unit you could've recommended. However, I can understand why you'd be disinclined to make changes because you view it as "good enough".
That's not how warranty expenses work, at least not in publicly traded companies that follow GAAP. The length of a warranty is not a direct statement about how much "confidence" a company has in a product, nor does it make any direct claim about the quality of the components contained therein.
The way warranty expenses work is that you take the estimated rate of failure on a per unit time and multiply that by the length of the term and the cost of repair. On a unit such as a PSU which is likely to be replaced rather than repaired, your cost to repair is essentially your manufacturing cost plus shipping. That expense is then accrued whenever you sell the unit, and as such must be included in your margin calculation.
Obviously, by the equation for the warranty expense, there are 3 ways to lower warranty cost:
1. Reduce the failure rate estimate. This is hard on a new product based on a new platform because your estimates are essentially just guesses. It's also hard to adjust this estimate after the fact for a given product without taking write-downs.
2. Reduce the manufacturing cost. Reducing the raw materials costs is one way to do this, but obviously that doesn't necessarily have a direct correlation to component quality because there are many ways to reduce component cost.
3. Reduce the warranty term. This is straightforward and easy to predict the effect of.
So you can see how if you're a product manager who wants to hit a certain price point, reducing warranty length can be a valuable tool because it is adjustable all the way up to product launch, even if you have a ton of inventory already sitting in a warehouse. That doesn't necessarily mean that the product is of lower quality.
This is simply a factually untrue statement based on the data presented in the source that you cited in reply #1055 (TPU). Here's the data:
ASUS DCuII (performance) 78 C / 42 dBA = 1.86 C/dBA
MSI Lightning 71 C / 42 dBA = 1.69 C/dBA
PowerColor PCS+ (performance) 68 C / 49 dBA = 1.38 C/dBA
MSI Gaming 79 C / 41 dBA = 1.92 C/dBA
The DCuII isn't the best, but it's not the worst either, which is why I take exception to your claim in reply #1055 that it's somehow a card with a "catch".
Let me make one thing abundantly clear: I do not make the claim that my weekly build maximizes value in a mathematical sense (i.e. is the most efficient build given perfect market information). However, I do make the claim that it is a good build with solid parts, and that you won't end up with a bad machine if you following. In other words, it the best build that I can find at the time that I make it. I am human, and I do miss some deals and prices will change for people looking at the market after the fact.
That's why I don't objection to people offering alternatives (in fact I welcome it). I do take objection when people make claims, such as the ones made in replies #1054 and #1055, that the parts I recommended are somehow bad or should be avoided.
In the part you quoted, I wasn't talking about how warranty expenses work.
All I'm saying is - as long as there's a correlation between warranty and quality in power supplies on the whole, the customer can look at the warranty length and expect, on average, lower quality from a product with short warranty than from a product with long warranty. It doesn't matter to the customer exactly why the company chooses 3 year instead of 5 year warranty - the fact the they did choose 3 year is enough to raise doubt. And when doubt has been raised, the customer reads a review to see if said doubt is unfounded, and lo and behold, in a 3 year warranty unit he finds 85°C CapXons instead of 105°C Nippon Chemi-Cons - which implies, but does not prove, that in this particular case there is a direct relation between length of warranty and quality of product.
Correct, it doesn't follow necessarily. Yet as I said above length of warranty and quality tend to correlate in the PSU world. The three parameters you listed can be set in various ways for a given warranty cost. From the customer's point of view, the best case scenario is option 1, i.e. when the failure rate estimate is reduced without undue increase in manufacturing cost. This results in a maximally reliable unit that's covered by long warranty (since costs didn't need to be cut by reducing the warranty term), and competitively priced against units with shorter warranty.
Techpowerup tests on an open test bench, not in a closed case. In Tom's Hardware's tests (also linked earlier), Asus DCII and MSI Gaming are roughly on par on the test bench, just as in Techpowerup's tests, with MSI winning by a small dBa margin. But when tested inside a case, Asus runs 9°C hotter and 0.4 dBa louder. There's the catch again. No, I'm not saying Asus 290X is a bad card per se, I'm saying if there's an XFX or a PowerColor PCS+ or a Tri-X or even an MSI Gaming for a similar price, then one of those is a better purchase.
All of this is quite obvious and completely fine by me. edit: actually, the part about not maximizing value wasn't obvious, the rest was.
Yet given better market information than what you had when you pieced the build together, those parts are bad purchases for what they cost and they should be avoided. No, they are not bad parts per se, they are bad parts per dollar compared to what else is available, and that is the only basis for my criticisms. But given what you say about not trying to maximize value here, I take back what I said in #1054 about changing this week's PSU.
I still recommend against buying the EA-650 Platinum and the 290X DCII for reasons stated, because else it wouldn't make any sense for me to recommend alternatives. Arguing against X is the flip side of arguing in favor of Y when X and Y are direct competitors.
Making decisions based on uncertainty and doubt is not something that I have ever recommended. Take the warranty for what it's worth at face value, which is a length of time in which the product will be replaced
but don't try to draw conclusions which simply cannot be drawn from a warranty period.
Look at it this way: if somebody posted a build with an i3 4330 and a NH-D15, you would probably tell them that they were spending way too much on their cooling system. Likewise, spending extra for 105C capacitors on a unit which doesn't touch 60C in 40C ambient is probably a waste of money. That doesn't mean that the 85 C capacitors are lower quality or more likely to fail.
The DCuII was the best value R9 290X at the time of making my list. It is not a bad card with a "catch" as was claimed in #1055. We can argue sources back and forth for days, but the fact is that the conclusions drawn in #1055 simply did not follow from the source provided. I hope that you can understand how that could grate on me one post after being accused of not doing the research.
Certainly there are cards with more efficient cooling systems, but they would cost you more based on the prices at the time that I made the post. Since the DCuII isn't somehow a lemon with a horrible cooler, but is instead in line with standard cards, it is the best value for an all-around system.
I think that you're misunderstanding me here. I did not say, "I don't try to maximize value". I said that I could notguarantee that my build meets the mathematical definition of maximum value. I absolutely do try to maximize value, but I am human and I can't see everything.
This is a bit of a false dichotomy. There is a difference between:
(a) offering a (better) alternative due to shifting prices, a different use case, or something that I just plain old missed
and
(b) making claims that the parts are bad recommendations derived from lack of research and which should not be purchased even at sale prices as were made in #1054 and #1055.
I welcome (a) but will vehemently defend my recommendations in the face of (b). That's not becasue I want to win value arguments or that I'm "disinclined to make changes,"
because I do put effort into researching and recommending solid parts and I do believe that someone could take my build every week and be happy with the result.
Bad analogy. Any cooler that is not Intel stock cooler is overkill. It is a well established fact of the past decades of Intel overclocking that it is voltage, not temperature, which kills processors.
Who cares what kills processors in the longterm? The reason we recommend aftermarket coolers at all is that they allow you to not-throttle while applying more voltage to stabilize your overclock, so it was a perfectly good analogy. If you're not overvolting to stabilize an overclock, and why would you with a locked-down i3, your CPU is going to run pretty cool unless you're AVX'ing the hell out of it or double your computer case as a cat-den.
No, I cannot understand that. The TPU link #1055 was only intended as a source for the claim that the DCII runs either quiet or cool. In calling it a "catch", there was implicit the assumption that other cards did not suffer from this dichotomy, an assumption I've since proven using other sources. In so doing I also disproved your assertion that the DCII "matches or beats" Lightning, PCS+ or even Gaming. It's worse than all of those, and it's worse than the XFX that I recommended in its place.
It is clearly not in line with "standard cards" such as the ones we've compared it to in this thread, PCS+ and Tri-X in particular, but MSI Gaming, XFX DD and Gigabyte WF3 as well. I don't intend to go over the previously posted detailed comparisons all over again.
However, I would like some clarification on the part "I can't see everything". Are you admitting here, that EA-650 Platinum was not the best value at the time you posted - that had you noticed the XFX XTR or HX650, for instance, you would've recommended those instead? If so, why so vehemently defend the EA-650 Platinum?
As you say, you can't overclock an i3, thus your reason for recommending aftermarket cooling does not apply. Therefore, any aftermarket cooling is overkill, and singling out Noctua D14 as being overkill makes no sense, therefore bad analogy. It was implicit in mfenn's analogy that other aftermarket would not be overkill for an i3-4330, when that is clearly untrue. EDIT: Just to be clear, the use of "overkill" here only has to do with reliability, stability and lifetime - the same metrics by which capacitors in power supplies are judged.
Please refer to the figures posted in #1057 and #1062 and see if you can understand the basis for my claim that it is in line with standard R9 290X's. Note that I am not trying to convince you to accept my reasoning here, it is clear that I will not be able to do that. I only want you to understand the basis for mine.
I've already stated in three different posts that I am happy to accept that there are good alternatives to my choices and I welcome you to post them. That includes all of the parts that you have listed. I am defending the EA-650 Platinum because you made the claim that is a bad power supply, even when on sale. I disagree 100% with that assertion for the reasons stated.
My analogy was regarding spending more on a part which has little added benefit versus an alternative. The primary benefit of aftermarket cooling on an i3 is to get rid of the high load noise level of the stock Intel cooler (65W version). Many aftermarket coolers achieve this at a much lower price point than an NH-D15.
Overall. it seems to me that we have very different fundamental perspectives on things like warranties, the meaning of the measurements posted in #1057, #1058, #1062, #1063, and the way to provide constructive criticism.
It's extremely unlikely I am going to convince you, and you are likewise not likely to convince me.
All I see in #1057 is your misinterpretation of my post, which I then cleared up in #1058.
#1062 - sure, given what you had researched about the card at the time, I can understand why you'd want to defend it against the claim that it's not in line with other 290X's.
Excuse me but you have no basis to say that I've claimed EA-650 Platinum is a bad power supply. This is what I've actually said regarding the EA-650 Platinum:
"I honestly don't recommend people buy this, even at a sale price."
Does not say: EA-650 Platinum is bad or unreliable.
Does say: Its sale price simply does not make it good enough value to be worth considering with respect to the competition (= the units posted in #1054). Specifically, I am talking about the current sale price in the current market, not some general sale price in a general market. Context.
"only 3 year warranty [...] no doubt a result of cheaping out on the capacitors"
Does not say: EA-650 Platinum on the whole is bad, or even that the capacitors are bad.
Does say: In a 5 year warranty unit, you'd probably see higher rated capacitors from a more expensive OEM.
"Antec tried to paint a picture that Platinum doesn't have to mean expensive, but what they forgot to paint was the part where Platinum doesn't have to mean quality either. "
Does not say: EA-650 Platinum is a bad quality unit.
Does say: EA-650 Platinum uses lower quality parts, is from a generally lower quality OEM, and/or lacks other quality-related features (modularity, long warranty, zero RPM fan) when compared to many units whose wattage and price it competes against.
I'm not sure if these perspectives are in any way fundamental, they seem more like value-for-money related opinions or habits which we've acquired from experience and research. Opinions can change, and should change.
I'm not very happy that you assume an attitude that's near-immune to being convinced. I'm even less happy that you assume me to have a similarly limited attitude, despite all the patience, attention to detail and interest in evidence and reasoning that I've shown here.
mfenn said:The specific phrase that made me think that you were claiming the EA-650 was a bad unit was "I honestly don't recommend people buy this, even at a sale price." (Emphasis mine). The use of the indefinite article (a) instead of the definite article (the) lead me to believe that you were talking about general market conditions versus this specific market, which would typically be indicated with the definite article.
Although mfenn is the forum moderator, he always takes the position that saving $10 on two sticks of memory is more important than purchasing equivalent specced but lower voltage memory. Or similar kinds of purchase recommendations (ignoring the power requirement differences and related heat output) when it's nVidia vs. AMD video cards. Which seems muddle-headed, to me at least.
But: he's making the big bucks employed as forum moderator, so...
You cannot reference another member's moderator status when they are posting as a member. This is both unfair and irrelevant to any member discussion at hand.
Perknose
Forum Director
Mfenn should be making dimes for sure