Upgrading Intel E6750 Stock Air Cooler (LGA 755)

[87dtna]

Yeah well the HEC you call junk I've used to power an I7 920 at 4.2ghz and gtx460 at 850 core clock for several months, then after that it's been powering my I5 750 rig for a year (several different video cards but same I5 at 4.2ghz).

So your guys opinions on what you'd consider junk is pretty irrelevent to me.

You'd probably also think that no 850w PSU would ever handle an X5667 (intel hex core I7 970 equivalent 12 threads) at 5ghz (yes 5ghz no typo) and 5850 tri fire overclocked to 1000 core clock. blah blah blah.

I've also powered the I5 750 at 4.2ghz with a gtx470 with a 450w PSU, I'm sure you wouldn't ''recommend'' that either.

 

[StrangleHold]

So in other words. Your saying you play Russian Roulette and havent blown out your brains yet. So it makes it irrelevent to you if I think its dangerous and you would recommend others to do it. You can run anything you want, but I am still not going to suggest it.

 

[Okedokey]

Of course we wouldn't recommend that. If you want to take the (stupid) chance with your own gear, then go for it, however this forum is for people who want proper advice - not rediculous suggestions to save what, a few bucks? False economics.

 

[87dtna]

Just because YOU consider it russian roulette doesn't make it so. Thats the whole point here. You have opinion, I have backed up FACTS based on my own experiences.

This also isn't a matter of saving a few bucks. He's not asking ''should I get this PSU or spend a few more bucks and get this one''. He already has the rosewill PSU, and I seriously seriously doubt going to a small 3.2ghz overclock is going blow it up. He's gonna be pulling maybe 10-20 more watts....OMG it's gonna blow

 

[Okedokey]

Of course when you scream your point (caps), you know one is desperately holding on to a losing argument. Your anecdotal evidence is not factual, its based on luck, and you are not the first to believe such things, however the probablity of a system damaged by such a low quality PSU even in its current stock state, is much higher.

You are also not the only one that may read this in future. So of course I have provide actual facts relating to the engineering principles outlined in my above posts, backed up by real facts.

You can do what you like, doesn't make it justifiable, nor responsible. If you want to take your view and claim that running a PSU at its very limit, especially with such low quality, go for it, but do it with your own gear, don't recommend it to the uninitiated.

Over and out.

 

[FuryRosewood]

*facepaws at this dicsussion*

 

[87dtna]

Of course when you scream your point (caps), you know one is desperately holding on to a losing argument. Your anecdotal evidence is not factual, its based on luck, and you are not the first to believe such things, however the probablity of a system damaged by such a low quality PSU even in its current stock state, is much higher.

You are also not the only one that may read this in future. So of course I have provide actual facts relating to the engineering principles outlined in my above posts, backed up by real facts.

You can do what you like, doesn't make it justifiable, nor responsible. If you want to take your view and claim that running a PSU at its very limit, especially with such low quality, go for it, but do it with your own gear, don't recommend it to the uninitiated.

Over and out.

LOL, you are the one spewing out bogus numbers (such as 30%, where's the scientific factual evidence of this 30%) and I'm the one losing the argument? You are just a legend in your own mind buddy.

*facepaws at this dicsussion*

^fail

*facepalm*

 

[StrangleHold]

LOL, you are the one spewing out bogus numbers (such as 30%, where's the scientific factual evidence of this 30%) and I'm the one losing the argument? You are just a legend in your own mind buddy.

Scientific factual evidence. It pretty common knowledge. Most quality power supplies are rated at around 40c. Which is what they under normal circumstances run at, even under load. The higher the temp, the faster the wattage and efficiency drops.

I'll give you a example, the seasonic I linked too is a 520W and in reality has as many watts and amps on the 12V rail as either of the other 650W.
The seasonic is rated at 40c with 83% efficiency and the Roadkill is rated at 25c with 72% efficiency. Now I hope we can agree the seasonic is better quality.

On the Seasonic model if you increase the temp to 50c the wattage and efficiency will drop by 20%. Now the Roadkill is rated at 25c and even at 25c is has a lower efficiency. Now we both know that a power supply isnt going to be running at 25c, what is that 77f. You would have to put it in a refrigerator. Guess why thay rate it at 25c, if you increased it to 40c. The efficiency would probably drop to around 60% and the wattage to around 500W and alot of that is on the 5V rail. So yes, 30% drop is pretty close to the real number without testing it.

 

[87dtna]

Yes I agree the seasonic is better quality, and I agree that the 600w is bloated and overrated. But even a quality 450w PSU would power his rig, it's not some crazy power hungry rig.
So, taking into account that 30% would be what it would derate by, and 9800gtx+'s typically consume 10-12a fully loaded (even take the high end 12 since his *may* be an OC edition).

So we have 12a max being consumed on the one 14a derated rail, and 5-6 amps from the cpu on the other 14a derated rail. Where's the danger in overclocking here? The absolute most he would pull from the cpu is 8 amps overclocked. Factor in the other 12v accessories and we are still no where near even the derated 14a figure.

So again to sum it up and reiterate, even accepting your 30% drop figure it's still not even close to maxxed out. I see no danger in a small overclock.

 

[Okedokey]

The 30% I refer to is simply a temperature derating curve based on the fact that all powersupplies will derate proportional to temperature. Low quality PSUs rated at 25oC can derate between 1W-4W/degree above 20oC. This means that a standard operating temperature of 50oC means that it could actually derate by between 2.5A - 10A in summer. This means that your 19A rail is now actually only able to deliver a MAX of between 16.5A and 9A. Average of that is 12.75A. As you can see, running a 9800GTX+ (requiring up to 13A according to nvidia), on a low quality PSU such is this is veyr dangerous. And this is just the graphics card, not the rest of the 12V system (minus the CPU).

What you have also forgotton is the synergistic effect of the cross loading temperature. We now have a rail that when used to provide power to the 9800GTX+ during a nice bout of summer gaming is drawing close to maximum. This produces even more heat. De-rating the ability of the other rail to provide proper energy. The 12V+1 rail is dedicated to the CPU, so it is completely out of the equation. However its ability to provide stable power is linked to GPU (and other) rail, as it is nearing 100% max load regardless. If this PSU was of higher quality with EPS2.91 cert, and 2 19A rails, then yes, you'd be right, but in this case, its simply stupidity to load this PSU further, regardless of rail tolerances.

You have also forgotten capacitor ageing.

Actually I cannot be bothered. If you want to recommend that this is a good idea go ahead. I certainly wouldn't do it.

 

[StrangleHold]

I never said anything about his CPU rail, or if he could overclock or not. Its his system rail thats at its limit. Under normal operating temps. he has around 14/15 amps on the rail. The average system with a 9800 GTX+ under load is around 275W. Say we -100W for the CPU. That leaves 175W being pulled off the second rail, thats around 14/15 amps being pulled off it. The rail wont trip till it hits 19 amps. But it only has around 15 max to pull. Eveytime he loads it while gaming, the rail is being pulled at its max. If your running it at its max, what happens. It heats up and its efficiency starts to go down even more. My point is, this thing will eventually burn up.

The problem with these cheap fake dual rail power supplies is the system rail. Everything from the 24 pin power connector that powers the Chipsets/USB/Lan/PCI and PCIe slots/Sound/CPU fan/Memory, anything that is powered from the motherboard. Then any thing hooked up to a PCIe or molex connector. All pulls off that second rail.

 

[87dtna]

Well, a 9800gtx+ is rated at the same TDP as a 9800gtx but draws substantially less due to being 55nm. So although it's rated TDP is ~12a, actual full load is around 10a. Now thats 100% load too.

I understand your point strangle, but he said he was running a 500w rosewill on this rig for 3 years before the fan crapped out. Now he's got a 600w. :-/

Anyway, since the CPU rail is totally seperate, an overclock shouldn't affect anything anyway since right now it's not even drawing 6a.

 

[Okedokey]

The GTX9800+ has a TDP of 141W (11.75A). Overclocked this would be higher of course. The GTX9800+ is now branded as a GTS250 (same thing) and that has a TDP of around 150W. So if you are going to make a guess about how the 55nm architecture would perform regarding power, its wise to use what nvidia have claimed - not some guess. 12A at max would be conservative, but hey, lets call it 11A.

Now where you have a rail that, due to derating, may struggle to provide 10A, things are not looking good.

Sceondly, the second rail isn't 'completely seperate', its a seperate current limitation only. It will suffer from heat, stability, ripple etc from the overloaded system rail - why, because it is 1 transformer coil. The 'rails' are simply voltage dividers that are current limited. Yes, they're seperate in regards to load, but not for all else. This is why, an overloaded rail on one will cause issues on the other.

In fact this could be a cause of the heating issues in the OP.

Anyway, to answer the question of the OP, you can overclock whatever you like, however I wouldn't recommend it. In fact I wouldn't recommend that PSU at all.

Get yourself a multimeter and test the 12V rail stability (molex connector). I suspect massive voltage droop which can cause temperature issues amongst other things. You are running this system on the edge of what is considered smart, you have been warned - however I hope it doesn't go badly for you.

 

[87dtna]

So, according to your philosophy, the TDP on a card is how much it WILL draw and then the specs of a PSU are crap and overrated. Nice philosophy.

You may know PSU's but you don't know jack about cards other than what you can read on a paper or website. The TDP is overrated, and thats not a guess I've owned a gts250, a 9800gtx+, and many other cards....and yes I've tested load wattage differences and it's more like 120-130w....and again thats 100% solid load with a benchmark, something most games won't do to a card.

And the TDP on a gts250 is 145w.

So now the rail is struggling to produce 10a. What happened to 30% loss, now we are up to nearly 50% loss. The more you talk the more I laugh. You don't even know what the name of the card is. ''gtx9800+''. LOL

 

[Okedokey]

So, according to your philosophy, the TDP on a card is how much it WILL draw and then the specs of a PSU are crap and overrated. Nice philosophy.

Derating would make the card draw more, not less. And yes, derating a PSU is standard practice.

You may know PSU's but you don't know jack about cards other than what you can read on a paper or website. The TDP is overrated, and thats not a guess I've owned a gts250, a 9800gtx+, and many other cards....and yes I've tested load wattage differences and it's more like 120-130w....and again thats 100% solid load with a benchmark, something most games won't do to a card.

So now the rail is struggling to produce 10a. What happened to 30% loss, now we are up to nearly 50% loss. The more you talk the more I laugh. You don't even know what the name of the card is. ''gtx9800+''. LOL

I said "at least" 30%. And backed that up with a derating curve, and calculations. You clearly didn't read it.

Even if it is 120W as you claim, thats 10A. Derating of such a low quality PSU could be 4W/degree over 20oC. So at 50oC thats 30 x 4 = 120W. Or 10A derated. On a 19A (max) rail that leaves 10A (max). So worst case scenario, you will have (even on your calcs) an overloaded system.

And the TDP on a gts250 is 145w.

Incorrect, its 150W.

http://www.nvidia.com/object/product_geforce_gts_250_us.html

 

[87dtna]

Yeah, well, the 500w rosewill he was running for 3 years was only rated at 16a on the 12v rail. It's obviously less than 30% not ''at least'' 30%. Kinda blows your whole theory right outta the water.

OK well my evga gts250 was 145w.

 

[Okedokey]

Doesn't blow anything out of the water - why, because he got lucky. The fan failed though indicating a higher work load due to heat, which is a symptom of overload. Who knows what other issues were caused.

What you've give above is a circular argument. It worked, because it works so it'll work. Any first year science student ca explain to you why that's wrong, and most certainly not a way to judge technical situations or risk.

At the end if the day, the PSU is junk, on the limits of what it is designed for and failure can cause loss of whole system. Risk assessment 101 says you don't push it, u less you're happy with the risk of damagIng the system. That's the whole point of this thread!

 

[87dtna]

Oh so now the fan died from high work load + heat and not because it's simply a cheap piece of crap fan. Talk about a circular argument. 3 years is a helluva long time for a cheap fan to work under any conditions, so it doesn't really seem thats its life was terminated short at all.

My point was, your 30% or more argument is obviously false for these PSU's. If it held up to a 16a rail for 3 years, it should be even better with a 19a rail. Common sense no?

You call it luck, but these PSU manufacturers would be long out of business if the power derating is as much as you claim. They'd have a million small claims cases from PSU's blowing up taking out their hardware.

You remind me of Ben stiller in Along Came Polly.....a guy that lives in his own little bubble assessing risk for a living therefore making sure everything is overkill and skewed toward the negative aspect to make absolutely sure there is no risk whatsoever.

Do you not walk over sewer grates because there's a 1 in 40,000 chance of falling in? LOL

 

[Okedokey]

I think this thread is over - the OP should have enough information to make an informed decision.

 

[87dtna]

He never even considered getting a different PSU, that was not the thread topic, and you did not change his mind. I would laugh at you too like I'm sure he is if I was running a 500w PSU for 3 years, bought 600w to replace it and then you tell me it's not a good enough PSU for my system and it might blow up.

I'm the only one that provided him with any kind of information he desired.