Sign in with
Sign up | Sign in
Your question

80 Plus vs 80 Plus Platinum

Last response: in Components
Share
October 8, 2012 4:08:15 PM

I am planning to build a gaming rig for around $2,000 early next year and need help figuring out if I am better off going with a power supply unit rated at 80 Plus Basic, Bronze, Silver, or Platinum. I will likely have the computer on from around 8 or 9 am to 10 or 11 pm. Most of the time I will be doing web related tasks with a few hours of gaming. I plan to use this rig for two or three years. Though I am waiting until Intel Haswell is released I figure a 120 mm or 200 mm cooling fan supply unit rated between 750 and 1000 watts may be sufficient. I want to keep noise levels to a minimum and have the right energy efficiency to save on electric bills over that two or three year period. What would be the best choice? Also, would I be better off with a cooling fan or heat sink power supply unit?

More about : platinum

October 8, 2012 5:00:03 PM

By and large, the whole 80+ program is purely marketing gimmickry and has no real-world application.

I would have to double check the specifics, but the general system of the 80+ program is that the PSU is at least 80% efficient at very specific load levels. I believe they are 20%, 50%, and 100%. There is NOTHING in the 80+ specification that says the PSU must be similarly efficient at any other load level which is where your computer is probably going to be 90%+ of the time. Now I will say that there is a fairly sound line of reasoning that says if a PSU is some minimal efficiency level at certain loads, it's probably going to have something of a bleed effect to other load levels. However, how much would vary PSU to PSU, and it could be anywhere from equaling efficiency at measured load levels for the 80+ program all the way down to no added efficiency.

So I would say by and large there is no particular reason to get an 80+ PSU unless maybe there's a good deal on one or you've done the math and figured out that you will be at one of the measured load levels a significant portion of the time.
a b ) Power supply
October 8, 2012 6:24:55 PM

When you look at PSU reviews where the reviewer tests in 10% or 20% load increments, there usually is a fairly smooth progression between test points and this is simply because it is how things tend to be. The magnetic and electric properties change in a generally highly predictable and smooth way which makes it very difficult for any properly designed device to show strange curves... if a device does have erratic output, it means there is either an engineering or manufacturing failure.

While the 80+ certification only tests at certain load points, the specification does require that efficiency be at least that good between points and I would bet my shorts the certification process does test at many more points than what the specifications require.

There is no limit to how low efficiency of non-certified devices can go and some sites like Hardware Secrets or Johnnyguru do occasionally bench random no-name PSUs and have found some units with efficiency as low as 60% where a 80+ PSU would be required to provide at least 80%.

However, meeting the 80+ specification with modern components is too easy and even some cheap units manage to pass it so people who want to rely on 80+ as a general indication of build quality should look for at least 80+ Bronze.
Related resources
a c 87 ) Power supply
October 8, 2012 6:38:14 PM

jpolk1138, I'd say with a substantial budget like $2000 you probably should get a power supply with as high efficiency as possible. Although it doesn't absolutely have to have any kind of 80 Plus certification at all, as long as it is high quality. As I explain below it is unlikely that you will find a high quality PSU these days without an 80 Plus certification, so I would say aim for Silver/Gold/Platinum.

If you are looking at 750W+ PSUs then your intended system should have at least two power hungry graphics cards.
If you want your PSU to be quiet then you should also make sure that the rest of your system is quiet, otherwise you probably wouldn't hear your PSU even if it was loud.

There is a theory that more efficient PSUs are quieter than less efficient PSUs because they have to dissipate less waste heat. However it is not as black and white as that; companies can still choose to use loud fans in their 'more efficient' PSUs.

cl-scott said:
By and large, the whole 80+ program is purely marketing gimmickry and has no real-world application.

Agree that the 80 Plus 'program' is marketing and not much more.

cl-scott said:
I would have to double check the specifics, but the general system of the 80+ program is that the PSU is at least 80% efficient at very specific load levels. I believe they are 20%, 50%, and 100%. There is NOTHING in the 80+ specification that says the PSU must be similarly efficient at any other load level which is where your computer is probably going to be 90%+ of the time. Now I will say that there is a fairly sound line of reasoning that says if a PSU is some minimal efficiency level at certain loads, it's probably going to have something of a bleed effect to other load levels. However, how much would vary PSU to PSU, and it could be anywhere from equaling efficiency at measured load levels for the 80+ program all the way down to no added efficiency.

As far as I'm concerned the only reason to doubt the efficiency claims made in an 80 Plus report is if the manufacturer/brand selling the PSU sent a cherry-picked or otherwise unrealistic unit to be tested by Ecova, in other words if the brand/manufacturer is not trustworthy or trying to get a competitive advantage it hasn't truly/honestly earnt.
Plenty of reviews show that good units, which have been 80 Plus certified, have an efficiency curve which is pretty straight; they don't very often, as you seem to suggest, fluctuate/dip between 20% and 50% or 50% and 100% loads.
I would be concerned with the efficiency below the 20% mark, because as a lot of people oversize their PSUs and spend most of their time in idle they are using less than 20% of the unit's capacity '90%+ of the time'.
However the efficiency aspect is not what is important to highlight IMO, it is the fact that it is only about efficiency, it is not about quality, or performance. Which are obviously more important factors in deciding on which PSU to buy.

cl-scott said:
So I would say by and large there is no particular reason to get an 80+ PSU unless maybe there's a good deal on one or you've done the math and figured out that you will be at one of the measured load levels a significant portion of the time.

Apart from the fact that it is quite hard to find a high quality PSU these days that isn't 80 Plus certified.
a b ) Power supply
October 8, 2012 6:58:16 PM

I don't see why you would think a PSU that is 80% efficient at 20%, 50% and 100% load levels would arbitrarily vary between those levels (to the effect of you implying the certificate isn't of use). That's against any physical intuition I have.
a b ) Power supply
October 8, 2012 7:08:03 PM

Silvune said:
Apart from the fact that it is quite hard to find a high quality PSU these days that isn't 80 Plus certified.

And once you get into 80+ Gold territory (the highest I might bother with unless prices drop closer to 80+ Bronze), there shouldn't be much left to quibble about quality-wise.
a b ) Power supply
October 8, 2012 7:16:57 PM

FinneousPJ said:
I don't see why you would think a PSU that is 80% efficient at 20%, 50% and 100% load levels would arbitrarily vary between those levels (to the effect of you implying the certificate isn't of use). That's against any physical intuition I have.

It can happen if the manufacturer lets magnetics enter into unplanned saturation or other problems like that... but then that would be a critical and amateurish engineering failure.
October 8, 2012 7:54:54 PM

FinneousPJ said:
I don't see why you would think a PSU that is 80% efficient at 20%, 50% and 100% load levels would arbitrarily vary between those levels (to the effect of you implying the certificate isn't of use). That's against any physical intuition I have.


That would of course be a logical fallacy. I'm not saying that as an insult, I'm saying that just because something is 80% efficient at three specific load points does NOT necessarily mean it will be equally efficient at any other load point. I'm not exactly sure how one would do it, but at least hypothetically it could happen that you have a PSU that is 80% efficient ONLY at those specific load levels, and is horribly inefficient at every other level.
a c 126 ) Power supply
October 8, 2012 7:57:09 PM

A psu will draw only the power demanded of it regardless of it's maximum capability.

It is unlikely that a more efficient psu will save you enough over 3 years to pay the premium for gold or whatever rating.

I suppose you could do a calculation.
the difference in efficiency might be as much as 10%
That might amount to 20w over an hour, or 200w per day.
For 1000 days, that is 200kw hours, and at 10c per kw hour, you are looking at $20 savings.

One good reason to go gold is that such a psu will likely be quieter since the fan will not need to spin up for cooling.

I would select a psu first on quality, then on power(amps/watts) capability.. A bit over minimum requirements will be best so the psu operates mostly in the more efficient middle third of it's range.
a c 260 ) Power supply
October 8, 2012 8:54:14 PM

FinneousPJ said:
I don't see why you would think a PSU that is 80% efficient at 20%, 50% and 100% load levels would arbitrarily vary between those levels (to the effect of you implying the certificate isn't of use). That's against any physical intuition I have.


I 2nd that thought!

Tom
a b ) Power supply
October 8, 2012 10:26:10 PM

cl-scott said:
I'm saying that just because something is 80% efficient at three specific load points does NOT necessarily mean it will be equally efficient at any other load point.

But apart from a major engineering oversight or manufacturing/material fault, there is absolutely no reason for this to happen. Also as I said before, thile the 80+ spec only provides hard numbers to beat at three specific load points, it is a common understanding that the lowest allowable efficiency between two points would be the least of the two points, there is no need to spell that out in the spec.

The reason why PSUs have lower efficiency at low load is because gate drive and other ancillary loads account for a larger chunk of incoming power. As power draw increases, this overhead accounts for an increasingly small percentage of input power. At very low loads like 10W, most of the power ends up in the gate-drive circuitry which uses almost constant power regardless of load as soon as the PSU is turned on.

The reason why PSUs have lower efficiency at high load is because losses in the magnetic core(s), copper windings, MOSFETs, diodes, etc. are causing them to heat up and become more lossy/leaky.

Both sets of phenomenons and everything in-between is very predictable and linear by nature. Producing a legitimate efficiency roller-coaster like what you alluded to would be a much bigger engineering challenge than producing a proper PSU.
October 8, 2012 11:42:33 PM

InvalidError said:
But apart from a major engineering oversight or manufacturing/material fault, there is absolutely no reason for this to happen. Also as I said before, thile the 80+ spec only provides hard numbers to beat at three specific load points, it is a common understanding that the lowest allowable efficiency between two points would be the least of the two points, there is no need to spell that out in the spec.

The reason why PSUs have lower efficiency at low load is because gate drive and other ancillary loads account for a larger chunk of incoming power. As power draw increases, this overhead accounts for an increasingly small percentage of input power. At very low loads like 10W, most of the power ends up in the gate-drive circuitry which uses almost constant power regardless of load as soon as the PSU is turned on.

The reason why PSUs have lower efficiency at high load is because losses in the magnetic core(s), copper windings, MOSFETs, diodes, etc. are causing them to heat up and become more lossy/leaky.

Both sets of phenomenons and everything in-between is very predictable and linear by nature. Producing a legitimate efficiency roller-coaster like what you alluded to would be a much bigger engineering challenge than producing a proper PSU.


I'm not disagreeing, I'm merely pointing out that there is nothing about 80+ PSUs that stipulates across the board efficiency levels. If people want to use the 80+ "badge" as an indicator that, all other things being equal, that PSU is more likely to be of better quality... I don't have a problem with that. But a lot of people are going to look at 80+ and think that means the PSU is 80% efficient across the board (I was one of them until I looked into a bit more) and that's not the case.
a b ) Power supply
October 9, 2012 12:53:09 AM

cl-scott said:
I'm not disagreeing, I'm merely pointing out that there is nothing about 80+ PSUs that stipulates across the board efficiency levels.

From 20% to 100%, the 80+ certifications are pretty clear about what the minimum efficiency has to be and with the Titanium certification, the lower bound goes down to 10%.

Trying to push certification range down to almost 0% would be a little silly since meeting higher efficiency goals at very low power levels would start to drastically increase costs and likely compromise greater efficiency gains across the rest of the range.

There isn't much of a point in pursuing higher efficiency if it comes at the expense of drastically higher prices that will never be recovered by power savings and also greater environmental impact of manufacturing those more efficient and likely much more complex devices.
a b ) Power supply
October 9, 2012 7:11:08 AM

cl-scott said:
That would of course be a logical fallacy. I'm not saying that as an insult, I'm saying that just because something is 80% efficient at three specific load points does NOT necessarily mean it will be equally efficient at any other load point. I'm not exactly sure how one would do it, but at least hypothetically it could happen that you have a PSU that is 80% efficient ONLY at those specific load levels, and is horribly inefficient at every other level.

I would love to see some actual proof for this argument.
!