• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

How stable is stable?

8

89MustangGX

n00b
Joined
Dec 18, 2005
Messages
33
Hello,
First time I really took a look at a power supply (I've been reading up on here) and I have a question about the stability.

I just put a new PS (CoolMax 600w Modular) in a new build (A64 3700+, Radeon x800XL) and I have been watching it through software (I understand this can be misleading).

The 5.5v and 3v I never really see move -- they are rock solid. The 12v seems to move a little. It will move from about 12.04 to 11.904 to 11.84. Now this is all within the 5% range specified, so I don't think I need to be worried about that (do I?), but I'm wondering about how often it moves. I haven't noticed a real difference between loaded and idle numbers, but it does seem to move around a little more during load.

I am posting a screenshot of the monitoring program I have been using with a chart, this was taking while under load -- can anyone offer any thoughts on this? I am not afraid to send this PS back, I want to be comfortable with my new setup.



Adam

BTW - I did have a 400w before this one and I thought it might be time to upgrade, but I do know that I don't need the 600w setup.
 
It seems fine. As long as everything is running stable (e.g. pc crashing, shutting down byitself, etc) you shouldn't be worried.
 
enyceexdanny said:
It seems fine. As long as everything is running stable (e.g. pc crashing, shutting down byitself, etc) you shouldn't be worried.
Click to expand...

Thanks for the reply.

Yes -- everything has been rock solid.

So, just to reiterate, basic feeling is that as long as the chart doesn't look like a 9.8 earthquake, numbers within 5%, and no operating problems we think it's A-Ok?

Adam
 
The 12V rail is supplying the CPU with power so a little droop is normal, it doesn't even mean the PSU is drooping, the motherboard could be getting a sag between where the 12V is reported from and where it's getting the draw against it.

Get a multimeter and check the voltages with it and I think you'll find that the fluctuations are a lot less than you're seeing.
 
madmat said:
The 12V rail is supplying the CPU with power so a little drop is normal, it doesn't even mean the PSU is drooping, the motherboard could be getting a sag between where the 12V is reported from and where it's getting the draw against it.

Get a multimeter and check the voltages with it and I think you'll find that the fluctuations are a lot less than you're seeing.
Click to expand...

I think that is true since many of the software monitoring do not have the resolution. They jump in rather large numbers in changes of the reading. A multi meter will give a better indications of fluctuations as well being more actuate.
 
Thanks for the replies. I think I'll stick with this one then. BTW - been paying more attention to it and it's pretty much solid until I put a full load on it, that's when I get the picture like you saw above. Guess that qualifies as normal.
 
89MustangGX said:
BTW - I did have a 400w before this one and I thought it might be time to upgrade, but I do know that I don't need the 600w setup.
Click to expand...


I would not be so sure of that
CoolMax is not known for producing "high quality supplies"
and that rating is very likely at 21 > 25C
meaning that in comparision to say a Seasonic its nowhere near 600W either peak or sustained under a reasonable operating temperature
(Seasonic for instance rates at 40C)

while you certainly dont "need" 600W or even 400W its a mistake to assume that PSUs are directly comparable without derating curves for each supply
(for instance FSP group derates their supplies 2W for each degree over 25C)
the vast majority of the rated power is simply unavailable, because its there for backward compatibility, its entirely possible to be short on one rail with gobs to spare on another
thus why old monster supplies cant run some hungry modern configs regardless of their wattage rating

its all about amps per rail, not watts

it seems reasonably stable with the current load your putting on it (whatever that is)
but keep the temperature variable in mind as well this summer ;)

the other variable would be stability under a given range of crossload which will fundementally change with different components
so keeping a close eye on the power with serious changes to either the load placed on it or environmental changes is a good idea ;)
 
Thanks for the additional reply.

I did do some reading and understand the importance of amps, but watts seem to be the universal language that most speak in (why?).

This is the PS I got: http://www.newegg.com/Product/Product.asp?Item=N82E16817159049

And this is what it says on the box (no temp rating):
DC O/P Load Max 600W
+3.3V 30A
+5V 55A
+12V1 19A
+12V2 19A
-12V 1A
+5VSB 2.5A
+3.3V & +5V Max 320W
+12V1 & +12V2 Max 430W

Now, can you tell me if this is adequate or roughly where this puts me in the range of power supplies from great to junk? Also, if it turns out this isn't a good PS at all, can you point me in the right direction -- a brand that you like who makes a modular model (I really like the cabling).

Thanks,

Adam
 
as to why "watts" is the lingua franca of power supplies:

once apon a time far far away, most computers had very similar power requirements
thus the amps needed per rail where generally in a set ratio, the +3.3V and +5V rail being what powered most everything that didnt spin, and those rails where in fact one rail as far as the power available was concerned, more draw on one decreased the power available on the other. So a supply rated in watts was a fairly good indicator of whqat to expect.

When CPUs and GPUs started to switch to the +12V rail that all changed

L0s7 4 Lyf3 said:
CPU power question...

Yep, you read the title I have a question about powering CPU's...

What reasons did CPU makers have to switch power draw over to the +12V Rail? What advantages does it have over +3.3V or +5V?
Click to expand...

gee said:
lots of reasons.

The main one is current - if you have a CPU pulling 100W on a motherboard, the Vcore regulator will pull >20 amps from the 5V rail. This is a lot of current to move around... at 12V, the current required is >8.3A - less than half. >20A requires thick PCB traces, heavy cabling, and so on.

You can more efficiently make a power supply that produces low current at high voltage, than a power supply that produces high current at low voltage. So your computer's PSU runs a bit cooler and draws less power from the wall.

And load regulation became an issue - lots of stuff in your computer (PCI cards, hard drive controllers, USB devices) run on 5 volts, and the 5V rail has to be stable. In your average linux box with a Prescott, your computer's CPU changes its power demand from 10W to 100W hundreds of times a second! and this causes immense current swings on the 5V rail, causing it to "jump around", possibily making your computer unstable. Few things in a computer require that the 12V rail is steady - normally it's just used for electromechanical things (fans, hard drive spindles, etc) which don't care if they're getting 10V or 14V. So moving the CPU regulator there made sense, since it keeps crap off the 5V rail.
Click to expand...
(gee is our resident electronics engineer and has actually designed switch mode power supplies SMP which is what we are discussing as opposed to linear, and the above is a very elegant answer to that question)

as far as that supply goes
I cant definatively answer what its quality is especially in relation to your specific power needs
its specs are certainly up to snuff, (read basic compliance but not exceeding any spec)
but what its actual performance parameters are I cant say.
I would avoid overtaxing it but you seem in no danger of that,

putting a supply to the test is far far harder than 99% of the PSU reviewers on the net seem to think, and requires tons of equipment (only a slight exaggeration)

considering your OK currently and without some compelling reason to get a supply that exceeds spec I think your fine, but thats not the same passing up this opportunity to pontificate a little :p

next time your in need of a supply youll have a little more understanding under your belt ;)
 
Hahaha..you paid 130 for that! You could go out and buy a Expensive Seasonic or Powersteam (I think) for that price! EDIT* yeah, Shows reading the [H] is a good thing before buying based on egg reviews. Sorry dude, I almost pissed my pants when I saw the price high amps on 12v or not..
 
dchrsf said:
Hahaha..you paid 130 for that! You could go out and buy a Expensive Seasonic or Powersteam (I think) for that price! EDIT* yeah, Shows reading the [H] is a good thing before buying based on egg reviews. Sorry dude, I almost pissed my pants when I saw the price high amps on 12v or not..
Click to expand...

Maybe instead of trying to make an ass out of me you could offer me reasons to go buy an expensive "Seasonic" or "Powersteam," what do you say?
 
Ice Czar said:
considering your OK currently and without some compelling reason to get a supply that exceeds spec I think your fine, but thats not the same passing up this opportunity to pontificate a little :p

next time your in need of a supply youll have a little more understanding under your belt ;)
Click to expand...

Thanks -- I understand a little more now than before I bought. Could you still throw me some brand names or models I could look at that would fit my needs and be considered a good all around unit -- something that would fall in the middle-high end of the spectrum?
 
Didnt mean to make it sound that way, sorry Mustang I was just in a rare mood and knowing there are threads all over saying this sort of stuff, then you purchasing a $130 PSU that is worth about 50 is sort of dissapointing. I hope you learned a lesson though. At least its not THAT bad of a unit like a deer or powmax.

Basically the Seasonics are super silent and are for the person wanting a near silent PC while providing stable voltages that rarely flucuate under load etc. Everything good in a PSU but personally I shy away from them since they dont have sleeved cables. The powerstreams are rated one of the best price/performance units out there and are very popular. Nothin more to say about them really.

I would suggest looking around google's results of Powerstream 520w review, and Seasonic 600w Reviews to get a sort of feeling how they perform. Also check out http://www.jonnyguru.com/ and look at the PSU reviews comparing the bad units to quality ones.

Hope this helps and sorry being such a ass, It wasent intended that way.
 
Yeah. At that price, just send it back and get a Seasonic and you won't have to ask this question. You will just know that your psu is rock solid. I just use software monitoring also and my 12v rail says 12.1 and never budges. And its just so darn quiet.
 
the seasonic's newegg review is only 4 out of 5 stars on their 600 watt model.

Some of the brands youfun guys make of have 5 out of 5 stars on their user reviews....

Including coolmax - they have 5 out of 5 reviews on several of their models.
 
Thanks for the replies.

I still have the same model Coolmax 400w that I have had for a while now and I don't think I ever had any voltage fluctuations or anything with it. In fact, I'm going to stick it back in and watch it now to see.

I don't want to say that you are all wrong and Coolmax is making the best product in the world or anything like that -- I'm definitely sending the 600w back -- but I might just stick with the 400w until I find one I like better or really have a reason to change it out. This was really just done for the heck of it anyway -- I never noticed any difference.

Thanks for the info guys, much appreciated.

Adam
 
Archaea said:
the seasonic's newegg review is only 4 out of 5 stars on their 600 watt model.

Some of the brands youfun guys make of have 5 out of 5 stars on their user reviews....

Including coolmax - they have 5 out of 5 reviews on several of their models.
Click to expand...
I doe not pay as much attention to reviews until it gets 3 out of 5 with over 50 replies. Some of the coolamax models do have 3 stars as well.
 
Newegg reviews are not what you should be reading for informed opinions on power supplies.
 
Ok, I put my original 400w Coolmax back in (again - same model but 400w), and here is what the voltage charts look like with that one.



Looks to me like all of the voltages are slightly higher (I know they might not be accurate, but they are higher) and the 12v rail is more stable, though the 3.3v is not as stable. From my recent education, it would seem that this is a better supply than the 600w.

So, looks like I will stick with this one until I find another one I like and send back the other.

Feel free to add any comments, make recommendations, or throw anything else in here. I certainly learned something about power supplies today.

Adam
 
well PSUs more than most components suffer from users perceptions
often subjectivly based perceptions, of course the manufacturers have created an environment that has fostered that.

continuing our story, there have always been manufacturers that have designed very stable and high quality supplies from reliable components, but Switch Mode Power Supplies (SMPS) are to say the least not easy to get right and have several disadvantages over a linear power supply, but because they are both more efficient and can be much smaller for their rating are what has been traditionally employed in computers. The main drawbacks of SMPS is they cant be as tightly regulated (voltage) as a linear without one hell of alot of additional circuitry, and they are "noisy" meaning that AC power spikes leak into the converted DC power as Periodic and Random Deviation (PARD) otherwise known as AC Ripple, and that requires more circuitry\components to filter it out. Thus when a manufacturer makes a "budget" supply the quality of the power isnt going to be the same even though its claimed capacity might be, that isnt immdiately evident to an end user.

Of course those manufactuers want to sell supplies and know that power quality isnt really a primary concern to most of the end users provided it actually can run a computer. So they incorporate bells and whistles like finish, displays, modular cables, windows, sleeving, UV reactive coatings ect. Inorder to attract end users, but those things cost money as well, so guess where the "value engineering" occurs most of the time? In the quality of the components or the complexity of the design that ensures better load regulation (how tightly the voltage is maintained) the transient response (how quickly the PSU reacts to a a change in the draw on the supply) and the AC Ripple. Now those parameters are set forth in the Spec, (ATX12V) for load regulation its 5%, and for AC Ripple its 50mV > 120mV p-p (peak to peak) depending on which rail (voltage) is being measured

So inorder to look like they are comparative the books need to get cooked
one way to do that is to artificially limit the acceptable VAC the power coming from your wall being Voltage Alternating Current. Your utility spikes and brownsout your power more than most appreciate, further simply having a heavy motor on the same house circuit can do the same (say a washer dryer or refrigerator) if a supply has a wide VAC rating it means that its able to maintain load regulation even when the VAC sags. But many manufacturers will limit that spec to a much tighter range so that they look comparative.

Next temperature plays a big role in two "end user" perceived triggers
the first is watts, the total capacity of a power supply is actually a determined when it all said and done by the temperature, its what is known as a derating curve, as the temperature increases the maximum power its capable of decreases, related to that is the second "trigger" noise. Fan noise is directly tied alot of the time these days to the operating temperature, so

by testing a supply at an artifically low temperature (typically the open workbench ambient of 21 or 25C per Intel certification specs) you get an unrealistic rating, why is it unrealistic? Because when that cert was originally assembled we didnt have the same firebreathing monsters we do today, and the PSU has traditionally exhausted the whole case, so now you have PSU more often than not directly exhausting the CPU heatsink at temperature (depending on how hot the initial air put into the heatsink is) upweard of 40>50C

How does a manufacturer develop a bad rep?
Simply stated they promise the stars and when some poor slob actually beleives them and adds far too much load to a supply for a given range of temperature things go very very wrong.
Supposedly supplies have "safety systems" that keep the PSU from allowing a catastrophic surge through the compoinents when they blow, but as I mentioned above designing a SMPS isnt nearly as easy as it would seem, they are prone to all sorts of instabilities and value engineering components especially say midway through a production run to save costs leads to problems, as does keeping up quality control and assembly standards.

In short provided your able to do a little mental windage when looking at the features and the specs
you get what you pay for

Testing:
the problem here is that crap supplies are pretty easy to spot, but once you get past those its very difficult without some pretty expensive components to actually stress test a PSU upto what could reasonable occur to an end user, say a hot summer room with summer induced power surges and brownouts from the utility and a marathon gaming session

thus the best that can be said about end user reviews, and just as importantly 99% of enthusiasts reviews is that it didnt blow up their computer that day

there are some notable exceptions, but to fully test a PSU takes about (no joke) $100,000 stack of electronics test equipment. The reviewers that at least know thier ass from a hole in the ground have assembled the most important components or reasonable facimilies of them to test the more important chracteristics of a PSU.

The leading reviewers are

Oleg Artamonov @ Xbit Labs
Lee RoboTech Garbutt @ Various
Mike Chin @ SPCR
our own jonnyGURU @ SLCentral

and even they dont "publish" stress testing though they do sometimes do it
the victims often being anonymous so they dont piss a manufacturer off considering thier "test" was "atypical and unfair"
but that is exactly what it would take to seperate out the wheat from the chaff in the upper end of supplies.

My recommendation is always the same, invest in infrastructure, get a good UPS it will make any PSU a stronger link in the chain, and get as good a PSU as you can afford, compared to other components even the most expensive are cheap

Id add that these days if your going to keep on the bleeding edge, get a PSU that is adaptable between the ATX12V and EPS12V spec for compatibility (compatibility is connections, while the number of +12V rails ect is compliance to a spec, strick compliance isnt always necessary) the rumor is that ATX12V will be adopting the 8 pin +12V auxillary connection form the EPS12V spec, much like it did the split +12V rail and the 24pin connector in the last major revison
 
You must log in or register to reply here.
Back
Top