Sign in with
Sign up | Sign in
Your question

a psu discussion for veterans great watts vs amps debate

Last response: in Systems
Share
March 31, 2013 1:28:23 AM

atomicwar and I have been talking about recommendations that we give to people needing upgrades. whats been happening is we agree on reliability, wattage and headroom. so the OP will ask "is my 450w coolermaster ok for a 650ti upgrade? and not wanting the OP to experience a comp meltdown wed recomend to upgrade to at least 550w and a better psu. but I guy will come along and say " eh, youll be fine with the coolermaster 450w..its got 28A on the 12v for your 650ti upgrade." but we like to play it safe with the suggested 500w psu. and the OP will toddle off knowing he doesn't have to upgrade. meanwhile we are recommending 550w xfx for headroom etc... we have missed many best answers because of this and also when the OP fires up his i7/650ti his CM will go poof. what we would like to know is where we all stand on this issue.? id like to hear from ko888 here also.
a b ) Power supply
March 31, 2013 1:33:08 AM

you know where i stand! better safe then sorry.
m
0
l
March 31, 2013 1:38:49 AM

I also think what happens is the OP Wants to hear what he wants to hear. like "omg I don't wanna spend more $" and then some random guy says hell be fine and walks away knowing that that huge decision just saved him 100$
m
0
l
a b ) Power supply
March 31, 2013 1:45:02 AM

ARICH5 said:
I also think what happens is the OP Wants to hear what he wants to hear. like "omg I don't wanna spend more $" and then some random guy says hell be fine and walks away knowing that that huge decision just saved him 100$


i agree people want to hear their stuff works...even if it doesn't or at least shouldn't. which is no reason to tell them its ok. personally i think anyone making recommendations that don't fit with the manufacturers recommendations should put a very clear disclaimer that its a bad idea ;)  they make those recommendations for a reason. its not just to get you to spend more money on PSU as i so commonly hear in PSU discussions. people forget about capacitor wear....the fact that 80% efficient is just that...500watts rated is actually 400W....and don't even get me going on sustained draw ratings which are usually even less.
m
0
l
March 31, 2013 1:47:36 AM

atomicWAR said:
ARICH5 said:
I also think what happens is the OP Wants to hear what he wants to hear. like "omg I don't wanna spend more $" and then some random guy says hell be fine and walks away knowing that that huge decision just saved him 100$


i agree people want to hear their stuff works...even if it doesn't or at least shouldn't. which is no reason to tell them its ok. personally i think anyone making recommendations that don't fit with the manufacturers recommendations should put a very clear disclaimer that its a bad idea ;)  they make those recommendations for a reason. its not just to get you to spend more money on PSU as i so commonly hear in PSU discussions. people forget about capacitor wear....the fact that 80% efficient is just that...500watts rated is actually 400W....and don't even get me going on sustained draw ratings which are usually even less.


or the houses "quality" power and ripple.

m
0
l
March 31, 2013 1:50:23 AM

lol rmbr that guy that turned his comp on by his light power switch? I thing he had his main power from his comp going to one of those dual light socket/220 plug things.
m
0
l
a c 84 ) Power supply
March 31, 2013 2:03:13 AM

Quote:
they make those recommendations for a reason.


They do, to make up for the Diablotek's in this world. The PSU recommendations are for the poor quality PSUs that are out there. Look at what they suggest for a card and then look up how many watts a system with that card REALLY take. These numbers usually aren't even close. They HAVE to say more then you really need because of so many low quality PSUs are out there.

The thing that gets me is when people see a "high end" GPU like a 660TI and think you need a 550W PSU just to run it. Whats the power draw on a 660TI? It's around a 6870, ~175W, I think even a bit less. You don't need a 550-600W PSU to run that. Heck, a good 600W with adapters could run two 660TIs.

I believe in enough power, but not to extremes. Running a ~300W draw rig on a 650W PSU is a waste. You spend more money then needed to just to buy the PSU. And by running it less then 50% load its not running at max efficiency. (though the difference between 82% and 80% isn't very large.) What does matter is if you spent the extra money to get a 650W on a higher rated 450W. Now my 450W silver or gold PSU that cost the same is running at 85% over your 80%. I'll end up paying less over the long run and I still have enough PSU for my needs. Keep in mind that in modern systems a ~300W system can be a 7870 or GTX660TI. Depending on the card even 325W can be a 7950.
m
0
l
March 31, 2013 2:47:46 AM

this is exactly what im talking about. are you suggesting a 325w psu can power a comp with a 7950? an sli 660ti on a 600w? 300w on a 7870? i digress and respect your opinion. i disagree though.
m
0
l
a c 84 ) Power supply
March 31, 2013 4:23:58 AM

Which part did you not understand?

No, a 325W PSU can't power a 7950 System. 350W PSUs don't have the two PCIe 6pin plugs. You also need SOME headroom to cover things. (Extra drives, higher heat, etc.) A QUALITY 450W PSU can do a 7950 system though. I'm not sure how many 450W PSUs have the plugs you need so you still might be looking at 500W minimum just to have the plugs you need. There is nothing wrong with running a QUALITY PSU at 100%. The worst part about it is you'll be at the lower efficiency of what it can do. Running a PSU around 50-80% of output is perfectly fine. This means for this supposed 7950 system that needs 325W of power a 650-407W PSU will do fine. What I was trying to say is why spend the $$$ on a 650W PSU when a 450-500W PSU will do fine? You can either get the 650W bronze unit, or probably get a 500W silver unit that will cost you less in the long run. I'd rather get 85% efficiency from a 500W then 82% efficiency from a 650W assuming both PSUs cost me the same.

Doing some more reading in this thread...

Quote:
the fact that 80% efficient is just that...500watts rated is actually 400W


Nope. Not the way it works. Efficiency has to do with how well it turns the AC electricity at the wall into the DC electricity your PC uses. If a QUALITY 500W PSU is 80% efficient then it will pull 625W (625 * .8 = what? ;)  ) of AC electricity to provide that 500W of DC power. A 90% efficient PSU will need to pull only 556W to provide the same 500W. This matters because you pay for the electricity you pull from the wall. So the more efficient things are the less you pay over time.

Not meaning to be rude at all here. It's good you are asking questions. But you two seem to be friends thinking you are right when obviously you don't understand the terms being used. Keep asking questions and tell me/us what you don't understand. Please don't just say "i disagree though."
m
0
l
a b ) Power supply
March 31, 2013 4:48:47 AM

4745454b said:
Which part did you not understand?

No, a 325W PSU can't power a 7950 System. 350W PSUs don't have the two PCIe 6pin plugs. You also need SOME headroom to cover things. (Extra drives, higher heat, etc.) A QUALITY 450W PSU can do a 7950 system though. I'm not sure how many 450W PSUs have the plugs you need so you still might be looking at 500W minimum just to have the plugs you need. There is nothing wrong with running a QUALITY PSU at 100%. The worst part about it is you'll be at the lower efficiency of what it can do. Running a PSU around 50-80% of output is perfectly fine. This means for this supposed 7950 system that needs 325W of power a 650-407W PSU will do fine. What I was trying to say is why spend the $$$ on a 650W PSU when a 450-500W PSU will do fine? You can either get the 650W bronze unit, or probably get a 500W silver unit that will cost you less in the long run. I'd rather get 85% efficiency from a 500W then 82% efficiency from a 650W assuming both PSUs cost me the same.

Doing some more reading in this thread...

Quote:
the fact that 80% efficient is just that...500watts rated is actually 400W


Nope. Not the way it works. Efficiency has to do with how well it turns the AC electricity at the wall into the DC electricity your PC uses. If a QUALITY 500W PSU is 80% efficient then it will pull 625W (625 * .8 = what? ;)  ) of AC electricity to provide that 500W of DC power. A 90% efficient PSU will need to pull only 556W to provide the same 500W. This matters because you pay for the electricity you pull from the wall. So the more efficient things are the less you pay over time.

Not meaning to be rude at all here. It's good you are asking questions. But you two seem to be friends thinking you are right when obviously you don't understand the terms being used. Keep asking questions and tell me/us what you don't understand. Please don't just say "i disagree though."


actually no you have the effiency backwards i suggest you read this
http://fkagchel.hubpages.com/hub/Everything-you-need-to...
quote from article
"For instance you have a 600 watt PSU with 70 percent efficiency that would mean it supplies 420 watts of power to your PC, with a wastage of 180 watts given off as heat. So the higher the efficiency percentage the less wastage your powersupply will have."

i know exactly what i mean when i talk about power efficiency rating. i served in the navy as an electronics tech and studied as a EE in college, had my A+ and C+ certifications (now expired as i retired do to disability)...point is efficiency rating is gives you actual wattage delivered from a PSU. 500W @ 80% effeciency means 400 watts is delivered and 100 watts is lost as heat.
m
0
l
a c 84 ) Power supply
March 31, 2013 5:11:36 AM

Wow. Can't believe he could do so well and then botch efficiency. Lets look at what's considered probably the best PSU testing site out there.

http://www.jonnyguru.com/modules.php?name=NDReviews&op=...

Scroll down to the third chart. (doesn't seem to be a pic...) Look at the middle column. DC Watts/AC watts. You can do the math yourself. AC watts times the efficiency number found two columns to the right = the DC watts. Trust me, the guy in that site you linked is wrong. I would think with that background you would know that.

Edit: Think of it this way if it helps. If it can only output 400W due to 80% efficiency then its LYING about being a 500W. Notice how I've been typing QUALITY in caps? That's because a QUALITY PSU can output the power that it claims to. (It also uses quality parts and all of them. It doesn't skimp when it comes to parts inside the unit.)

Can anyone link a forum sticky that talks about all this?
m
0
l
a c 84 ) Power supply
March 31, 2013 5:27:12 AM

Try reading this. Not sure he did any better then I did.

http://www.jonnyguru.com/forums/showthread.php?t=1036

Quote:
The calculation for efficiency is DC Output divided by AC Input.

When a power supply is more efficient, it will use less power from the wall than one that is less efficient even if it produces the same amount of DC power.


Edit: While I'm thinking about this... Did you see where I wrote above about cheaper PSUs "skimp when it comes to parts inside the unit"? Here is an example. Sort of.

http://www.jonnyguru.com/modules.php?name=NDReviews&op=...

Quote:
More line filtering out front. Two Y caps, one X, one coil. No MOV or TVS diode for surge suppression - as always, Superflower counts on the APFC parts to provide that functionality. I'm still a bit dubious about that practice, but I've stopped scoring against it.


Ever hear about cheap PSUs weigh less then quality ones? That's because they are LITERALLY missing parts they should have. The MOV is often one of the first things that gets removed. Superflower did it from this PSU, I know FSP often removes it from theirs as well. They also start to remove caps or even the coil from the line filtering. This is what I meant by skimping on the parts inside a unit. I thought I'd point it out if that wasn't clear.
m
0
l
a b ) Power supply
March 31, 2013 5:58:36 AM

your are correct about lower quality supplies (which i never recommend)...they are defective, unsafe or incompete...or all of the above...


while we disagree on efficiency though your not totally wrong nor am i, let me explain... as you know the formula for efficiency E=Pin/Pout the problem is not all manufactures rate the power in and out the same....quality supplies use AC watts from the wall as power in as you stated and DC power out (correct use of efficiency formula)...less scrupulous manufactures take the DC power the PSU is capable (peek load) of producing as power in (ie 500W capable =>400 scenerio) and then divide what the rail(s) actually DC output to the cables at any given moment under load. now the reason i always go with the latter is its the safe bet....as you pointed out there are some bad manufacturers out there that could care less if someone system fries...which is why one reason our sources differ on outputs.

is the practice right? of course not....does it happen? most certainly i read about it all the time. so as much as i was being short with you...we are both right depending on scenario...i apologize if i was overly aggressive. i am just tired of peeps saying 300w supply is plenty for gtx690 or some other ridiculousness.
m
0
l
a b ) Power supply
March 31, 2013 6:06:05 AM

anyway i hope that made more sense for you...i am a bit tired as it is 3 am here...i tried little more clear the second go around.
m
0
l
a c 84 ) Power supply
March 31, 2013 10:17:31 AM

If they are counting like that they shouldn't. Those are the guys lying about their PSU, and I never suggest people buy them. I'm so anti lying that if the label on the PSU isn't correct you shouldn't buy it. Meaning if it doesn't say what the "four" 12V rails can output together then don't buy it. If you are honest about your PSU then I'll consider it. There are now to many good PSUs out there for me to bother with the junk.

Quote:
i am just tired of peeps saying 300w supply is plenty for gtx690 or some other ridiculousness.


LOL, and I'm tired of the opposite. I was just in a thread a few hours ago where a guy was getting a seasonic built 850W PSU for his GTX670. Uses what, 180W? Whole system would use 300-325W. So even under load he'll be using less then 50% of his PSU. Worse, when he is just surfing and doing nothing he's system will probably use less then 150W. That's less then 20% of his PSU. Will be HORRIBLY inefficient. But that's the PSU he wanted I guess.
m
0
l
a b ) Power supply
March 31, 2013 10:29:39 AM

4745454b said:
If they are counting like that they shouldn't. Those are the guys lying about their PSU, and I never suggest people buy them. I'm so anti lying that if the label on the PSU isn't correct you shouldn't buy it. Meaning if it doesn't say what the "four" 12V rails can output together then don't buy it. If you are honest about your PSU then I'll consider it. There are now to many good PSUs out there for me to bother with the junk.

Quote:
i am just tired of peeps saying 300w supply is plenty for gtx690 or some other ridiculousness.


LOL, and I'm tired of the opposite. I was just in a thread a few hours ago where a guy was getting a seasonic built 850W PSU for his GTX670. Uses what, 180W? Whole system would use 300-325W. So even under load he'll be using less then 50% of his PSU. Worse, when he is just surfing and doing nothing he's system will probably use less then 150W. That's less then 20% of his PSU. Will be HORRIBLY inefficient. But that's the PSU he wanted I guess.


yeah either extreme is no good...i think that what this thread is all about ;)  and your right lying manufactures stink...they just exsist is all. i like to play it safe so someone doesn't fry a nice system. better to overspend on a psu then underspend and buy a whole new pc !
m
0
l
a c 84 ) Power supply
March 31, 2013 10:56:31 AM

Quote:
i like to play it safe


So the question now becomes what "safe" means. One of the things I like about HardOCP and their testing is they put a PSU under an 80% load for 8 hrs. The thing is I've never seen a PSU pass that and fail the 100% load. Meaning if a PSU passes the 100% load test it can pass the 80%/8hr test as well. Modern quality PSUs can handle 100% output. You don't want to buy a 350W PSU for your 300W load system obviously. But there isn't much of a need to go over that. Some people seem to think that "safe" means buying twice as much PSU as you need so you are running at 50% of output. And as above I strongly feel this isn't right either. The answer will depend on each person and the amount of "risk" they are willing to take. For me, an extra 50-100W is fine. As long as you have "native plugs" for what you want. (Meaning no power adapters.)
m
0
l
a b ) Power supply
March 31, 2013 12:09:54 PM

agree on the native plugs ( i did run an AGP era psu without pcie plugs for my 7850 for a while. It was a pretty nice quality one though :p )

best thing we should do is explain well to the OP the risk he is taking when he dosen't have a quality psu that minimally meets his needs.

about the efficency thing, it should be wall draw vs output. Its related to how much power a psu needs to input to ouput it's rated amount
m
0
l
March 31, 2013 5:15:19 PM

i enjoyed your arguement 4745.:)  yes, we should all play it safe when recomending a psu. and save the "power draw from the wall/100% at load watts" for the quality psu's. id say silver and above.
m
0
l
April 1, 2013 10:27:28 AM

another discussion id like to have is the age old argument..." high wattage and low amps? or high amperage and low watts? case in point. I have a seasonic 520 with 43 amps. relitivley low wattage but very high amps. http://www.tweaktown.com/reviews/5191/seasonic-platinum.... now I would have no problem with putting a gtx670 on that psu. next scenario: I have a coolermaster with 725 watts but with 28 amps. would you put a 670 on it?
lately I noice OP's not stating their brand of psu but just their watts. "I have a 600w psu..can I run this on it?" case in point: http://www.tomshardware.com/answers/id-1640307/card-buy... ...now the other "experts" on that thread has this guy bolting out the door to get his new 670. that's not an isolated thread. its happened many times where the experts are not informed and the OP is unwilling to give info needed to make a smart recommendation. on that note, WE all need to make informed recommendations. and try to talk the everyday OP from making bad desions. try to help the few, but save the many. :) 
m
0
l
a c 84 ) Power supply
April 1, 2013 6:01:56 PM

I almost feel like I'm hijacking this thread...

Quote:
I have a seasonic 520 with 43 amps. relitivley low wattage but very high amps. http://www.tweaktown.com/reviews/5191/seasonic-platinum.... now I would have no problem with putting a gtx670 on that psu. next scenario: I have a coolermaster with 725 watts but with 28 amps. would you put a 670 on it?


Watts is more or less meaningless for me. When systems moved to the 12V rail(s) it's the amps on that rail that matter. In your first example that PSU is a platinum unit that can do ALL 520W on the 12V rail if needed. (516W but who's counting.) The second unit I probably wouldn't even touch. "725W" but only 336W available on the 12V rail(s)? That's a joke by today's standards. I'm guessing that PSU is an older one that wasn't designed for todays parts.

As you correctly pointed out in that thread sometimes/often we need to question the rest of the system before giving advice. You might think 600/725W is plenty of power but it could be some older unit like in the above example. 336W is barely enough for a GTX670, but I wouldn't risk it on an older PSU with aged caps. In nearly every thread we get either not enough info, or some DX report that fills the page with useless junk. Is it really so hard to give us info so we know the system? Apparently it is...
m
0
l
a b ) Power supply
April 2, 2013 2:01:44 PM

amps on a rail usually help speak to the componants inside the unit.

I 100% agree that if you have the money for a new power supply then go ahead and get one. I also believe that 50-100 watts is the bare minimum for overhead wattage to have for a system at load. this puts him at running 80% of capacity at full load (350/400-450 watts). Considering that this seasonic unit runs at 80% efficiency till 450 watts I think that is an acceptable number. It is kinda surprising that they have not labled it that way.
the above is from another thread ehere arich5 and i were chatting
http://www.tomshardware.com/answers/id-1641693/7870-400...
m
0
l
a c 84 ) Power supply
April 7, 2013 6:32:43 AM

That is a long horrible read. What were we supposed to take away from that?
m
0
l
!