Seasonic Platinum-1000 1 Kilowatt Power Supply Review @ [H]

FrgMstr

FrgMstr

Just Plain Mean
Staff member
Joined
May 18, 1997
Messages
55,701
Seasonic Platinum-1000 1 Kilowatt Power Supply Review - Seasonic breaks the Platinum efficiency barrier with its new 1 kilowatt computer power supply. We generally know if it has Seasonic branding, it is a good PSU. But we all want to know if it is a great PSU. We all know that Seasonic means "quiet" as well and believe it or not this PSU can be dead silent.
 
Thanks for the review. Seems like a very strong contender for the PSU in my Ivy Bridge build.
 
I would like it more if it had more than One 12v rail.

I understand their design reasoning, its easier to control the voltage and power quality, but for $250 I dunno I would prefer something in the 1200w class even if this model is the best.

Would highly recommend this PSU for extreme overclocking competitions for sure though. Extreme OC demands it. This unit could be the difference between setting a record and watching a system go up in smoke.
 
I enjoyed reading the review. The results were even better than I had expected.

This will be my next power supply.
 
I just happened to be looking for a supply to replace my HX750 so I could go trifire and saw this supply. I got my order in on the last day of the Newegg preorder window so I got it for $20 off and free shipping. It came last week. I installed in a HAF X without a problem. I have it running a i7-930 @ 4.2, 3x unlocked 6950 cards, 1 SSD , 2 HDD, 1 ODD, and 5 fans. I have been playing BF3 a lot and haven't had any problems. I'd definitely recommend this unit.
 
Too bad this review wasn't out earlier.. a couple of weeks ago Newegg had a pre-order promo.. $239.99 after $20 IR and free shipping

TechPowerUp has the only other review of this PSU and concluded "if you want the best 1000W unit money can buy today then SS-1000XP is the right choice".

Saw this quote over at Anandtech concerning this PSU..

jonnyguru will jizz his pants after reviewing this one
Click to expand...

The GOLD Award doesn't seem enough for this.. why not make a PLATINUM award?

Also.. Why is the PASSED logo so much bigger than the GOLD award?
 
The Load Testing section in the last page says the PSU efficiency is 0.18% under the 100% load Platinum specs for desktop PSUs (89% efficiency,) but page 4 shows a 88.13% efficiency at 100% load, which is more like 0.87% less than Platinum specs.
Am I missing something?

I was amazed to see that Platinum specs for server PSUs require 95% efficiency at all times. Is it even possible? Can the guys who manage to do it then move to designing Platinum solar cells? ^-^
 
Chimel said:
The Load Testing section in the last page says the PSU efficiency is 0.18% under the 100% load Platinum specs for desktop PSUs (89% efficiency,) but page 4 shows a 88.13% efficiency at 100% load, which is more like 0.87% less than Platinum specs.
Am I missing something?

I was amazed to see that Platinum specs for server PSUs require 95% efficiency at all times. Is it even possible? Can the guys who manage to do it then move to designing Platinum solar cells? ^-^
Click to expand...

Its because [H] uses more strenuous conditions for their standard load testing, so the efficiency goes down. If you look at page 6 they have the "80 Plus load testing" which used the 80 plus conditions for testing, which is where that 0.18% number comes from.
 
On page 6:
Since the Seasonic Platinum-1000 is advertised by Seasonic as being certified for 80Plus Platinum, meaning that it is supposed to be 90%-892%-89% efficient at 20%-50%-100% load, we went ahead and ran the 80Plus Load Tests as shown here with 80Plus' lower ambient temperature and its AC input voltage of 115v.
Click to expand...
Looks like the Seasonic just missed platinum by 800%. :p
 
Thanks @Flyinfinni, I missed that section.
And forget about the 95% efficiency for datacenter server PSUs, it's 95% power factor, not efficiency.
Still have no clue what that means, and how it's different from efficiency. I was taught P=UV, i.e. W=VA, but it's apparently not the case.
 
Chimel said:
Thanks @Flyinfinni, I missed that section.
And forget about the 95% efficiency for datacenter server PSUs, it's 95% power factor, not efficiency.
Still have no clue what that means, and how it's different from efficiency. I was taught P=UV, i.e. W=VA, but it's apparently not the case.
Click to expand...

Well, Power = Voltage(V) * Current(I), so you are correct. However, in a power supply or whatever you are working with, some power is lost in the system due to the current and voltage not being perfectly in phase with each other. So what you are looking at is you have "apparent power" which really is = V*I, but REAL power is = V*I*PF so with a PF of 95%, and an apparent power of 1000W (12V, 83.3A), you end up with 1000W * 0.95 = 950 W REAL power.

Wikipedia's explanation is actually pretty decent.
http://en.wikipedia.org/wiki/Power_factor
 
I read the Wikipedia definition, and one from another site, and ended up even more confused.
Thanks for clarifying. Your formula P=V*I*PF makes more sense, I don't understand why they teach only the simplified version at school. Hope this has changed. Stupid French education system... ^-^

On another note, all these efficiency gains will be dwarfed by the latest ARM CPUs for servers that use 5W, 1.5W at idle...
 
Chimel said:
I read the Wikipedia definition, and one from another site, and ended up even more confused.
Thanks for clarifying. Your formula P=V*I*PF makes more sense, I don't understand why they teach only the simplified version at school. Hope this has changed. Stupid French education system... ^-^

On another note, all these efficiency gains will be dwarfed by the latest ARM CPUs for servers that use 5W, 1.5W at idle...
Click to expand...

The only issue there is that, yeah they may only be 5W processors, but how does their processing power compare to a sandy bridge based Xeon or Bulldozer based Opteron? How many of those 5W chips do they need to match a single Xeon or Opteron? Power efficiency is only 1 factor in the usefulness of a server (granted one of the biggest right now) but footprint is also a big factor. If you can fit 100 Xeon cores in a 1000 sq. ft. room and it takes 10kW to run it at load (I'm making up numbers;-), it makes a big difference if it takes 1000 of these ARM chips at 5kW to run, but takes a 2000 sq. ft. room.
 
Crud. Now I'm thinking of replacing my 850w hairdryer (PCP&C 850 SSI) with this just to keep noise down.
 
Flyinfinni said:
The only issue there is that, yeah they may only be 5W processors, but how does their processing power compare to a sandy bridge based Xeon or Bulldozer based Opteron? How many of those 5W chips do they need to match a single Xeon or Opteron? Power efficiency is only 1 factor in the usefulness of a server (granted one of the biggest right now) but footprint is also a big factor. If you can fit 100 Xeon cores in a 1000 sq. ft. room and it takes 10kW to run it at load (I'm making up numbers;-), it makes a big difference if it takes 1000 of these ARM chips at 5kW to run, but takes a 2000 sq. ft. room.
Click to expand...
The thing is that they don't have to match Xeons. They are apparently suited for specific tasks, such as file serving, not tasks that require high computing power.
They don't need a whole CPU, GPU or ATX PSU if all they manage is HDDs or SSDs.
Hell, maybe we'll see servers with redundant motherboards and CPUs at this rate! ^-^

About space, the company mentioned savings in the order of 90%, HP says 2,800 servers in a single rack.
Anyway, the CPU is one thing, but all the paraphernalia, price and industry adoption is another.
HP plans to use these CPUs and Intel Atoms for his new range of low-energy servers.
When it rolls out, the entry cost for a new datacenter will be like 15% of new ones created today. And probably more like 5% of datacenters created a year or two ago... [/off-topic]
 
Chimel said:
I read the Wikipedia definition, and one from another site, and ended up even more confused.
Thanks for clarifying. Your formula P=V*I*PF makes more sense, I don't understand why they teach only the simplified version at school. Hope this has changed. Stupid French education system... ^-^

On another note, all these efficiency gains will be dwarfed by the latest ARM CPUs for servers that use 5W, 1.5W at idle...
Click to expand...

As you may know, AC power is a Sin wave that travels between a positive and negative voltage. For example, standard US power is +-120Vrms at 60Hz. I think Europe is +-240Vrms at 50Hz. However, you cannot simply multiply 120V by the current drawn. AC is actually a voltage wave and a current wave. Ideally, these waves overlap perfectly. In reality, due to the capacitance or inductance of the load, the voltage wave and current wave do not overlap perfectly. One wave will "lead" or "lag" the other. Their distance apart, measured in degrees, dictates how much real, useful power you may actually use. The relationship is PF=Cos(theta). An ideal PF is 1, at worst it is 0. The higher it is, the more useful power you have.
 
Great review. Hopefully JG does one soon and confirms. Would be a great upgrade from my X650. Better start saving those pennies...
 
Seems a bit overpriced for what it is. Voltage regulation is its best feature, efficiency is decent considering its platinum and the rest is just average. I just don't see the $250/260 value in this.
 
Hope the 860W version is just as good. Supposed to be out Dec. 860W is just about fine for my needs.
 
It's a Seasonic so I wouldn't expect anything less than a Gold. Thanks for the review!!
 
xIronCrossx said:
Seems a bit overpriced for what it is. Voltage regulation is its best feature, efficiency is decent considering its platinum and the rest is just average. I just don't see the $250/260 value in this.
Click to expand...

As an overclocker, I'll jump for this power supply when I'm building my Ivy Bridge system. Sure there's an up front cost, but I know it's extremely unlikely to crap out and take any compoents with it. Also, it'll probably work perfectly five years down the road.

Plus it's quiet.
 
xIronCrossx said:
Seems a bit overpriced for what it is. Voltage regulation is its best feature, efficiency is decent considering its platinum and the rest is just average. I just don't see the $250/260 value in this.
Click to expand...

So don't buy it. But for the rest of us it'll be a nice upgrade. There are plenty of lesser 1000W rated PSU's selling in the $220-$250 range, so the Seaosonic isn't exactly outrageously priced.

I'll take Paul and Kyle's word for it.

"The Seasonic Platinum-1000 is quite literally THE power supply on the market today."
"The good news is that in that intervening time period Seasonic clearly did something right as this is perhaps the absolute best overall power supply we have ever seen."
 
Last edited:
Thanks, @Mad Cat. It looks like most decent PSUs reviewed by [H] have a pretty good PF of 0.99, way above 80Plus specs.
As I understood, it doesn't really affect residential customers who are billed in kWH, only industries like datacenters who are billed in VA.
Still, the more efficient, the greener.
 
PiSquared said:
As an overclocker, I'll jump for this power supply when I'm building my Ivy Bridge system. Sure there's an up front cost, but I know it's extremely unlikely to crap out and take any compoents with it. Also, it'll probably work perfectly five years down the road.

Plus it's quiet.
Click to expand...

Corsair AX1200 supplies more power, costs only $10 more and has amazing performance. I just find the AX1200 is priced too closely to make the Seasonic worth the money.
 
Corsair AX1200 is an OEM Flextronics, and not in the same league as the Seasonic. Not to mention the AX1200 is a LOT bigger. That makes the Seasonic compatible with a wider range of smaller cases.
 
Chimel said:
Thanks @Flyinfinni, I missed that section.
And forget about the 95% efficiency for datacenter server PSUs, it's 95% power factor, not efficiency.
Still have no clue what that means, and how it's different from efficiency. I was taught P=UV, i.e. W=VA, but it's apparently not the case.
Click to expand...
Reality is a bit more complex than that. In particular the issue comes with how exactly you define P, U and V?

The underlying physical equation is.

"instantaneous power"=""instantaneous voltage"*"instantaneous current"

However in an AC system the "instantaneous voltage", "instantaneous current" and the "instantaneous power" are all constantly changing.

If current is proportional to voltage then we can say

"mean power"="RMS current"*"RMS voltage".

It turns out that the above is the best case. In cases where voltage and current are not proportional the "mean power" (known to electrical engineers as the "real power") will be lower than the value from the above equation (known to electrical engineers as "apparent power").

The ratio of real power to apparent power is known as the power factor.

There is a convention of using the unit symbol "W" for real power and the unit symbol "VA" for apparent power.

Mad Cat said:
In reality, due to the capacitance or inductance of the load, the voltage wave and current wave do not overlap perfectly. One wave will "lead" or "lag" the other. Their distance apart, measured in degrees, dictates how much real, useful power you may actually use. The relationship is PF=Cos(theta).
Click to expand...
The above is true for capacitive and inductive loads. However with electronic loads things get more complex since the voltage and current waveforms will not only be offset in time they will also be different in shape.
 
Last edited:
I used to have a PC P&C 750w Seasonic built unit that developed coil whine under load with an SLI configuration. I loved the power supply performance, but the whine got to me. Does anyone know if any of the Seasonic branded units like this one or the X-Series have that issue? I am assuming they have taken care of this by now. I would love to pick this new one up for the serious voltage regulation!
 
Flyinfinni said:
Well, Power = Voltage(V) * Current(I), so you are correct. However, in a power supply or whatever you are working with, some power is lost in the system due to the current and voltage not being perfectly in phase with each other. So what you are looking at is you have "apparent power" which really is = V*I, but REAL power is = V*I*PF so with a PF of 95%, and an apparent power of 1000W (12V, 83.3A), you end up with 1000W * 0.95 = 950 W REAL power.

Wikipedia's explanation is actually pretty decent.
http://en.wikipedia.org/wiki/Power_factor
Click to expand...
Except the difference between real and apparent power isn't consumed, it's transmitted back through the power lines. The only thing it really affects is the power company's transmission losses, not your own energy costs.
 
Zero82z said:
Except the difference between real and apparent power isn't consumed, it's transmitted back through the power lines. The only thing it really affects is the power company's transmission losses, not your own energy costs.
Click to expand...

It won't affect residential users who have a high PF. However, some industries have a terrible PF due to their use of large electric motors and such, and will be billed extra by the power company.
 
plugwash said:
Reality is a bit more complex than that. In particular the issue comes with how exactly you define P, U and V?

The underlying physical equation is.

"instantaneous power"=""instantaneous voltage"*"instantaneous current"

However in an AC system the "instantaneous voltage", "instantaneous current" and the "instantaneous power" are all constantly changing.

If current is proportional to voltage then we can say

"mean power"="RMS current"*"RMS voltage".

It turns out that the above is the best case. In cases where voltage and current are not proportional the "mean power" (known to electrical engineers as the "real power" will be lower than the value from the above equation (known to electrical engineers as "apparent power").

The ratio of real power to apparent power is known as the power factor.

There is a convention of using the unit symbol "W" for real power and the unit symbol "VA" for apparent power.


The above is true for capacitive and inductive loads. However with electronic loads things get more complex since the voltage and current waveforms will not only be offset in time they will also be different in shape.
Click to expand...

Zero82z said:
Except the difference between real and apparent power isn't consumed, it's transmitted back through the power lines. The only thing it really affects is the power company's transmission losses, not your own energy costs.
Click to expand...

Right- I was trying to simplify it as much as I could:)
 
Mad Cat said:
It won't affect residential users who have a high PF. However, some industries have a terrible PF due to their use of large electric motors and such, and will be billed extra by the power company.
Click to expand...
In that case, I don't think the difference between a PSU with a PF of 0.9 and one with a PF of 0.95 would be of particular concern to them.
 
Corsair AX1200 supplies more power, costs only $10 more
Click to expand...

The CORSAIR Gold AX1200 is $300 but OOS at Newegg..$297.27 at AVADirect.. $317.15 at Amazon and $280 at TigerDirect while the Seasonic Platinum 1000 is $260 at Newegg.. $321.17 at AVADirect but N/A at Amazon and TD. Caeteris paribus, it seems as if you will pay $20-40 more to get 200 additional power although at not the same efficiency but very close.

Corsair AX1200 is an OEM Flextronics, and not in the same league as the Seasonic. Not to mention the AX1200 is a LOT bigger.
Click to expand...

During OklahomaWolf's review of the AX1200, he wrote "The soldering on this unit is very high quality, among the best I've seen" and had no complaints about the quality of workmanship or components used. OW eventually gave the AX1200 he reviewed an almost perfect score of 9.8.

As for size, the AX1200 is 150mm(W) x 86mm(H) x 200mm(L) while the Platinum-1000 is 190mm(W) x 86mm(H) x 150mm(L) making the Platinum-1000 about 95% the volume of a AX1200.. the two PSUs just have their volume shaped differently.. the AX1200's length certainly is a lot longer than the Platinum-1000.. but one would think anyone using a 1000-1200 PSU is using a fairly large case and probably either PSU would fit with no problem with respect to other components.. or at least consider how much room they had for a PSU before buying it.

Both seem like excellent PSUs.. might be splitting hairs on which one is really better.. probably depends more on a person's needs/desires/brand loyalty/other than on which PSU is actually "better".
 
Insula Gilliganis said:
As for size, the AX1200 is 150mm(W) x 86mm(H) x 200mm(L) while the Platinum-1000 is 190mm(W) x 86mm(H) x 150mm(L) making the Platinum-1000 about 95% the volume of a AX1200.. the two PSUs just have their volume shaped differently.. the AX1200's length certainly is a lot longer than the Platinum-1000.. but one would think anyone using a 1000-1200 PSU is using a fairly large case and probably either PSU would fit with no problem with respect to other components.. or at least consider how much room they had for a PSU before buying it.
Click to expand...
The AX1200 is 10mm longer. The 150mm*86mm dimension is the size of the ATX PSU mount, which is standard.
 
xIronCrossx said:
Seems a bit overpriced for what it is. Voltage regulation is its best feature, efficiency is decent considering its platinum and the rest is just average. I just don't see the $250/260 value in this.
Click to expand...

You may find X-1250 a bit more worthy, since it's almost equally priced, a bit less efficient, about as stable, delivers 250W more power and carries the same feature set as its Platinum brethren.
 
xIronCrossx said:
Seems a bit overpriced for what it is. Voltage regulation is its best feature, efficiency is decent considering its platinum and the rest is just average. I just don't see the $250/260 value in this.
Click to expand...

xIronCrossx said:
Corsair AX1200 supplies more power, costs only $10 more and has amazing performance. I just find the AX1200 is priced too closely to make the Seasonic worth the money.
Click to expand...

The Seasonic is a higher quality unit than the Corsair. Specifically, the fan in the Seasonic is a Sanyo Denki unit that probably retails for about $25 (based on the price of the 9S1212F401; I can't find the F404, but I'm assuming it's basically identical minus a tach output). The fan in the Corsair is a piece of garbage from Yate Loon. It's ball bearing, but I very seriously doubt that it's in anything like the same league, quality-wise, as the Sanyo Denki in the Seasonic.

This may or may not matter to you, but it's a feature that I'm willing to pay for. All of my CAD builds, and all of my servers that don't require redundancy get Seasonic X-Series or the lower-wattage Corsair AX (which are based on the X-Series) power supplies instead of something of basically equal quality specifically because those Sanyo Denki fans are typically rated for 60,000hrs of operation at 60*C with a 90% survival rate.

Fan failure is one of the two leading causes of power supply failure in my experience, and all other things being equal, a PSU that's got a fan from a reputable manufacturer (i.e. Delta, Sanyo Denki, Nidec or EBM Papst) is worth $50 more to me than a PSU that doesn't.

And if that's not important to you I'd bet money that the Seasonic's fan has a much less offensive noise profile than the AX1200's does.
 
Siorus said:
The Seasonic is a higher quality unit than the Corsair. Specifically, the fan in the Seasonic is a Sanyo Denki unit that probably retails for about $25 (based on the price of the 9S1212F401; I can't find the F404, but I'm assuming it's basically identical minus a tach output). The fan in the Corsair is a piece of garbage from Yate Loon. It's ball bearing, but I very seriously doubt that it's in anything like the same league, quality-wise, as the Sanyo Denki in the Seasonic.

This may or may not matter to you, but it's a feature that I'm willing to pay for. All of my CAD builds, and all of my servers that don't require redundancy get Seasonic X-Series or the lower-wattage Corsair AX (which are based on the X-Series) power supplies instead of something of basically equal quality specifically because those Sanyo Denki fans are typically rated for 60,000hrs of operation at 60*C with a 90% survival rate.

Fan failure is one of the two leading causes of power supply failure in my experience, and all other things being equal, a PSU that's got a fan from a reputable manufacturer (i.e. Delta, Sanyo Denki, Nidec or EBM Papst) is worth $50 more to me than a PSU that doesn't.

And if that's not important to you I'd bet money that the Seasonic's fan has a much less offensive noise profile than the AX1200's does.
Click to expand...

Here's the page for the 9S1212F401 fan. The difference between the 401 and 404 is that the 404 doesn't have a tach. I bought a few of these last year to compare to Gentle Typhoons. They are strong performers and focus airflow straighter than the Gentle Typhoon series, but are louder per rpm with a characteristic that is more noticable.
 
FaRKle0079 said:
Here's the page for the 9S1212F401 fan. The difference between the 401 and 404 is that the 404 doesn't have a tach. I bought a few of these last year to compare to Gentle Typhoons. They are strong performers and focus airflow straighter than the Gentle Typhoon series, but are louder per rpm with a characteristic that is more noticable.
Click to expand...

Not to take this too far off topic, but I'd agree with that assessment word-for-word. I switch back and forth between the Gentle Typhoons and the L/M/F401s on most of my builds depending on what's available at the time and how badly I want to crimp molex connectors. In terms of assembly and material quality I think they're pretty much a wash, although if my memory is correct the Sanyo Denkis have a lower starting voltage. It's been a couple years since I compared them, though, so I could be mistaken about that.
 
You must log in or register to reply here.
Back
Top