Power supply requirements for my choice of three

[jonnyGURU]

Originally posted by: orangat

Originally posted by: jonnyGURU

Originally posted by: orangat
There is a reason it sold for $10.

You think it's selling for $10 because it's not a good power supply? It's $10 after mail in rebate at one vendor. So you know little about PSU's and you know little about retail. Two for two.

Radio Shack is running the promo because Radio Shack is hurting in the B&M and are trying to stir up revenue and traffic to their website. I'm sensing they might pull a Montgomery Wards (file bankruptcy and then go online only.) The PSU sells for at least $60 anywhere else.

The 500W X-Finity will work fine if you're going to run a single Radeon X850xt, 7600GT or 7900GT. You could also use it for Crossfire X850xt and it's even approved by ATI to do so. But I'm not sure it would like 7600GT's or 7900GT's in SLI.
..............

My my what a hothead you are.
Your own review of the 600W model showed that the 12V rail is at the maximum 5% deviation at typical 3.3/5V loads. Oh well I'm sure its good enough and a complete bargain at $10.

Yeah.... I'm a hot head. And yes, I showed a pretty good deviation in voltage in my review. But my review is absolute torture. Case and point: You see my BFG review? That thing got SLAUGHTERED. But it's still a good PSU and worth more than $10. It's not something I would recommend for the money. There's better choices, but it's not like what you make it out to be. If that thing was selling for $50, I'd be pimping that thing left and right... as long as you didn't have SLI/Crossfire. 😉

That Antec would completley burst into flames in my load tests, but that doesn't mean the OP shouldn't try to use it in his rig. Open your eyes and be objective. I realize it's not like you said the PCP&C 1kW was crap and only worth $10, but the OP doesn't even have SLI/Crossfire! 😉

 

[orangat]

The x-finity series are defintely better than x-connect models but the voltage stability isn't that good. In the 500W model the xbit review shows that the 3.3V rail is barely adequate across most of the typical range where the 3.3V load would be at. The 600W model is better but has poor crossloading.

 

[jonnyGURU]

Originally posted by: orangat
The x-finity series are defintely better than x-connect models but the voltage stability isn't that good. In the 500W model the xbit review shows that the 3.3V rail is barely adequate across most of the typical range where the 3.3V load would be at. The 600W model is better but has poor crossloading.

X-bit's 3.3V is NOT where it typically would be. That's why the only crossload I do is the low 3.3V+5V (inspired by your comment in a post you made in another thread, BTW. 😉)

I was going to do no crossload at all, but then I saw the results of a real-life test that put Crossfire and a Pentium 4 CPU in a build with only two sticks of RAM and one hard drive and the 12V dropped below spec... yet the PC didn't hang, lock, etc. Very odd. I had to investigate... And I did. Enquiring minds wanted to know! 😉 By the way, I calculated that the actual 3.3V+5V combined load would be less than 30W. Use a calc like the one on Zippy's site (a blatent rip off of Takaman's I might add) and you'll come to the same conclusion. Any way... it's just like when I started out this hobby and used lightbulbs to load the PSU. Rails crossload and drop out of spec.

Fortunately, Oleg does state that his crossload tests are exaggerated. People are not designing power supplies to pass artificial crossload tests and adhere to the Intel ATX polygon (anymore.) They're designing them to work in regular PC's. Unfortunately, nobody reads anymore. They skim.

 

[Smartazz]

Originally posted by: orangat

Originally posted by: Harvey
This ad for Ultra Products is in one of the forum banner ads. The giveaways are nice, but they also show link to a special @ Radio Shack for an Ultra 500 W supply that comes in at $9.99 after rebates.

I don't know much about Ultra, but if you can use a 380 W supply, one rated at 500 W should probably hold up for your needs. 🙂

The Ultra is not a good brand. There is a reason it sold for $10.

Normally they're pretty expensive.

 

[pkme2]

One should check the prices on these Seasonic S-12s.
http://www.xoxide.com/seasonic.html?gclid=CILB1rXOqoUCFTV7JAodDwtd3Q
The prices are between $110 and $145, July 4th.

 

[orangat]

Originally posted by: jonnyGURU

Originally posted by: orangat
The x-finity series are defintely better than x-connect models but the voltage stability isn't that good. In the 500W model the xbit review shows that the 3.3V rail is barely adequate across most of the typical range where the 3.3V load would be at. The 600W model is better but has poor crossloading.

X-bit's 3.3V is NOT where it typically would be. That's why the only crossload I do is the low 3.3V+5V (inspired by your comment in a post you made in another thread, BTW. 😉)

I was going to do no crossload at all, but then I saw the results of a real-life test that put Crossfire and a Pentium 4 CPU in a build with only two sticks of RAM and one hard drive and the 12V dropped below spec... yet the PC didn't hang, lock, etc. Very odd. I had to investigate... And I did. Enquiring minds wanted to know! 😉 By the way, I calculated that the actual 3.3V+5V combined load would be less than 30W. Use a calc like the one on Zippy's site (a blatent rip off of Takaman's I might add) and you'll come to the same conclusion. Any way... it's just like when I started out this hobby and used lightbulbs to load the PSU. Rails crossload and drop out of spec.

Fortunately, Oleg does state that his crossload tests are exaggerated. People are not designing power supplies to pass artificial crossload tests and adhere to the Intel ATX polygon (anymore.) They're designing them to work in regular PC's. Unfortunately, nobody reads anymore. They skim.

Why do you state that the 'bad area' in xbit's 3.3V crossloading test is not typically in normal use?
http://www.xbitlabs.com/articles/other/display/atx-psu4_10.html - 500W
From the look of the crossloading graph of the 500W model, the red(5% deviation) area is right along the bottom below 50W which is where the typical power draw is at.

http://www.xbitlabs.com/articles/other/display/atx-psu4_12.html - 600W
And from the look of both graphs, both the 500/600W models have the best 12V stability when the power draw of the 3.3/5V rail is increased in linear fashion along with the 12V rail - which is not typical of normal use.

Where does Oleg state that his crossloading graphs are exaggerated?

 

[orangat]

ttt for jonny.

 

[jonnyGURU]

Originally posted by: orangat
ttt for jonny.

Gee thanks. :roll:

I just don't think they represent realistic crossloads. I said my peace and got out.

Oleg didn't state his graphs were exaggerated. I did.

Look... His load tests don't make sense. He shows the 3.3V as 5% off the entire time the 3.3V and 5V have ANY kind of load on them going anywhere from 0 to 50W, as long as there was ANY load on the 12V rail going all of the way up to 300W. I can easily duplicate that and NOT see any deviation in the 3.3V or 5V rails unless the 12V was at or above 75% of it's capability, asssuming we kept the 3.3V and/or 5V below that 50W threshhold.

It's like his graph is reversed. Something is either wrong with the PSU, the data or the testing methodology.

Think about it. Since most computers idle at 100W with only about a 30 to 50W load on the 3.3V and 5V combined, and will go as high as 300W when under duress. That would mean ANY test of that PSU anywhere on the WWW would also exhibit a 3.3V rail fluctuation of -/+ 5% when tested in a regular PC. But we don't see that, do we? In real life builds with real life loads, we hardly ever see these crossloads unless the build only has one hard drive, two sticks of RAM and the 12V is overloaded with multiple CPU cores and multiple GPU's.

The crossload you should see is a 12V that drops out of spec, not a 3.3V or 5V that drops out of spec. Not unless you do a crossload that consists of an unusually high 3.3V and 5V and a low 12V (like putting an ATX12V PSU in an ATX build) And if a low 3.3V and 5V crossload condition can't be overcome by slapping a 10 or 20W resistor on a 5V lead, theoretically or realistically, then there's something wrong with the PSU. That said, if someone is showing a 5% deviation in the 3.3V rail no matter what the 12V load is and no matter what the 3.3V load is as long as it's at or under 50W, then there's something wrong.

 

[orangat]

Originally posted by: jonnyGURU

Originally posted by: orangat
ttt for jonny.

Gee thanks. :roll:
I just don't think they represent realistic crossloads. I said my peace and got out.
Oleg didn't state his graphs were exaggerated. I did.

Look... His load tests don't make sense. He shows the 3.3V as 5% off the entire time the 3.3V and 5V have ANY kind of load on them going anywhere from 0 to 50W, as long as there was ANY load on the 12V rail going all of the way up to 300W. I can easily duplicate that and NOT see any deviation in the 3.3V or 5V rails unless the 12V was at or above 75% of it's capability, asssuming we kept the 3.3V and/or 5V below that 50W threshhold.

It's like his graph is reversed. Something is either wrong with the PSU, the data or the testing methodology.

Think about it. Since most computers idle at 100W with only about a 30 to 50W load on the 3.3V and 5V combined, and will go as high as 300W when under duress. That would mean ANY test of that PSU anywhere on the WWW would also exhibit a 3.3V rail fluctuation of -/+ 5% when tested in a regular PC. But we don't see that, do we? In real life builds with real life loads, we hardly ever see these crossloads unless the build only has one hard drive, two sticks of RAM and the 12V is overloaded with multiple CPU cores and multiple GPU's.

The crossload you should see is a 12V that drops out of spec, not a 3.3V or 5V that drops out of spec. Not unless you do a crossload that consists of an unusually high 3.3V and 5V and a low 12V (like putting an ATX12V PSU in an ATX build) And if a low 3.3V and 5V crossload condition can't be overcome by slapping a 10 or 20W resistor on a 5V lead, theoretically or realistically, then there's something wrong with the PSU. That said, if someone is showing a 5% deviation in the 3.3V rail no matter what the 12V load is and no matter what the 3.3V load is as long as it's at or under 50W, then there's something wrong.

Said your peace?
You said people don't read but simply skim and making up a statement where Oleg himself stating his graphs were exaggerated - which you then had to retract.

I can understand if you simply want to dismiss xbit x-infinity testing as faulty but I still don't get what you are saying in your 2nd (bolded) comment.

Why is the crossloading/stability test not representative of real world conditions? Why would we not see those crossloading conditions - low 3.3/5V, high 12V which the x-infinity fares poorly in voltage stability?

 

[jonnyGURU]

No offense, but you're either confused or trolling. Maybe I'm just hard to follow and I'm sure that's my bad. But maybe I'm hard to follow because I'm flustered because you're trolling. 😛

Oleg DOES say his graphs reproduce an exaggerated load. My claim is that they're just plain WRONG. No retraction. I misspoke. My angle is just different looking at the whole situation since Oleg would never admit that he's wrong. 😉

If you're going to hold my misspeak of the word "exaggerated" vs. "wrong" then you are just trolling me... either that or I'm flip-flopping worse than John Kerry. 😉

Yes, they are exaggerated, sure. But IMHO a PSU should be able to handle exaggerated crossloads too. And there are some realistic crossloads in the "rainbow" as well. Look at my reviews and you'll see I've done some exaggerated crossloads and have even had PSU manufacturers ask me, "why the hell would you ever crossload a PSU like that?"

That's not my point.

I'm not going to say there's anything wrong with an exaggerated crossload as long as someone looks at it and says it's exaggerated. I'm focusing on the one point you're making using the 3.3V results as an example and I'm just flat out telling you it's wrong.

My point is that if his graph was representive of any real world testing on the same PSU anywhere on the web, why do we not see the same crossload anywhere else on the web? THAT is what I mean in my second bolded statement. I thought you'd understand that and that's why I bolded it. Even if the crossload isn't exaggerated, it would still show a 3.3V that is off by 5% and yet we never see that anywhere else on the web.

A 3.3V and 5V load of less than 50W is very typical. Oleg's chart shows the 3.3V as 5% off as long as the 3.3V and 5V load are under 50W. A very TYPICAL scenario, agreed? So if that's a very typical scenario why don't we see the 3.3V off by as much as 5% in every other review of that power supply on the web? Why can't I reproduce it in the lab? Why? Because there's something wrong with the PSU or the testing methodology.

Unfortunately, I've seen this before in Oleg's tests and not just an X-Finity but a number of other PSU's that I've also load tested, which is why I think the problem is in the methodology. Not until Oleg can post what the load on each rail is and what the voltage on each rail is during each load, I can't give his load charts any creedence. Because he's showing "numbers" (I wish he actually showed numbers, but hopefully you know what I mean) that don't coincide with anything anywhere else in the industry or the web, then his methodology has to be put into question.

 

[orangat]

Trolling? This is what you originally said - "Fortunately, Oleg does state that his crossload tests are exaggerated. " and you insinuated some people only skim and don't read.
Take that how you will.

And portions of Oleg's graphs are testing exaggerated crossloads but I was only refering to the bottom edge (from the very beginning) which only shows low 3.3V/5V crossloads.

You were a little hard to follow because you were long-winded. Now I understand you were simply saying that Olegs graphs are uncorroborated by others.

And did you actually test the 500W model? Your own results mirror Olegs own result in his graph - 3.3V/5V rails good but 12V sags under high load. Some other brands like Hiper also exhibit poor 3.3/5V stability in Oleg's graphs under typical real world conditions.

 

[jonnyGURU]

My results don't mirror Oleg's. My 3.3V didn't drop when I crossloaded. The 12V did. And the 12V didn't drop significantly unless the PSU was crossloaded. And when the 3.3V and 5V was only 30W, it only took taking them up to 45W for everything to come back together again.

I'm done here. If I'm so difficult to understand. I guess I should just quit altogether.

 

[Tsuwamono]

heres something to calculate... so you dont buy 600 watts when you dont have to http://www.extreme.outervision.com/psucalculator.jsp

 

[orangat]

Originally posted by: jonnyGURU
My results don't mirror Oleg's. My 3.3V didn't drop when I crossloaded. The 12V did. And the 12V didn't drop significantly unless the PSU was crossloaded. And when the 3.3V and 5V was only 30W, it only took taking them up to 45W for everything to come back together again.

I'm done here. If I'm so difficult to understand. I guess I should just quit altogether.

I was referring to YOUR review was for the 600W model NOT 500W which is model in the link for $10 AR.