Sorry, but I can't give any credence to a review of a PSU that was specially built by the manufacturer for the reviewer. How bout buying a PSU from the Egg or walking into Best Buy and getting one, the same PSU we minions would buy, and test that? I can't help but wonder how "hand-picked" the PSU for this review was. This review is worthless.
[quote]We are still waiting for our 15 hard-drives to arrive, but we have talked with at least one AnandTech reader that successfully powered up 18 hard-drives with a single-rail Silencer 750W where all multi-rail power supplies failed.[/quote]
Am I missing something here? I've been running a 16-drive RAID from a cheap 500W Fortron PSU for over one year with no issues. The drives don't even draw 200 watts while working and, even if they weren't set to staggered spin-up, probably wouldn't hit more than 400 during boot.
Stick to Zippy, Fortron and PCP&C, and you'll be fine. Nearly everything else is either a rebranded Zippy / FPS or a piece of crap.
[quote]As always, PCP&C provides a 7-year warranty for their power supply. We're not sure how that works with the 200,000 MTBF, as even running 24/7 for ten years wouldn't reach half that figure, but that's apparently a testament of their PSU quality.[/quote]
This is a disgusting admission: if the writers at Anandtech don't understand how MTBF is calculated, what chance is there for the people who come to Anandtech to get technical information?
(sigh)
The way it is calculated is:
The manufacturer decides what a "normal" lifespan is for the product, and then runs a large number of models for that length of time, and sees how frequently they fail.
So, for example, they might decide that a "normal" lifespan for a power supply is 5 years. If the MTBF is 200,000 hours, that means that if you run 1000 power supplies continuously for 5 years, on average one of them will fail every 200 hours within the first five years. The figure tells you nothing whatever about what the failiure rate is after the first five years (although you can be sure it will be higher) and of course there is also no indication about what the manufacturer thinks is a "normal" lifespan for the product: maybe it's 5 years, maybe it's just 1.
This is naturally spectacularly misleading to the consumer who thinks it means that just one power supply will tyically run for 200,000 hours before failing. Sadly, he is wrong.
You're right, and I inserted the comment mostly tongue-in-cheek. The reality is that theoretical MTBF is only vaguely interesting. 1 million is better than 500K, and 500K is better than 250K... at least in theory. In practice, there's an element of "luck of the draw" - plus it depends on how the testing is conducted.
So start with 1000 power supplies (or hard drives, or some other component). Make sure all of them function -- dead on arrival units don't count towards MTBF! Now start running all of them with a stopwatch and see how long can it takes before a failure occurs. In the case of PCP&C, it appears that out of 1000 units, the first failure occurred after 200 hours.
How many times do they repeat this sort of test? They probably do it once for each new model. Or maybe they run all of the power supplies until they have two failures? Probably not. The main point is that MTBF isn't a super useful figure, unless you're talking about operational MTBF -- which is only available after the fact, when a device has been on the market for a while. I consider the seven year warranty to be a better indication of quality than the MTBF.
I'm actually running 10 internal sata drives on the 750W silencer along with a quad core, and an 8800gts 512, PCPower&Cooling definitely makes the best power supplies and their method of sticking with the single rail i hope isn't something they will change in the future.
"If you're buying a PCP&C power supply, you should not plan on silent computing - there just isn't any power supply in the lineup that can qualify as silent."
may be misleading, as my 610W Silencer operates silently in my dual Xeon workstation.
You might consider qualifying your statement more thoroughly.
Regards,
jtleon
"Quiet" means different things to different people. I consider most computers to be noisy, but then again I've been "tainted" by SPCR. As for PCP&C PSUs, no, they're not silent IMO. They're not even particularly quiet. However, that's just IMO, just as you thinking your unit operates silently is in your opinion. The best way to quantify "quiet/silence" may be if the PSU on its own (not being drowned out by other fans) is noisier than ambient sounds in a quiet room.
I have a Silencer 470 and it is indeed quieter than the ambient noise, even when all the other fans are off.
I do believe that the sound depends on the heat level and the power you are drawing, the most I ever pulled was a couple hard drives, optical drives and an overclocked x2 65watt and 7900GS @ 700mhz core. My PC's usually have some 120mm fans keeping the interior cool (better some slow-moving large fans then one really fast small one.)
As far as "silencer", I think it is apt, depending on the rest of your rig, even in the dead of night the Silencer fan is better than my fairly quiet hard drive, and it has a pleasing tone quality (not whiny or windy.)
That said, I am fully satisfied with my PCPower and I am thinking that "only" a 470watt was overkill, especially since the 360 silencer is $25 cheaper for 5A less on the 12volt rail ($55 direct!!), I can't recommend another PSU for mainstream overclocking (~3-4 Ghz, 1-2 drives, no SLi, or mainstream SLi, 7600/8600).
I too read SPCR, and I wish I had the money to do my own entirely passive, "heat-tunnel" rig (all the heat passes through one or two vertical pipes, and convection moves the air). Since I can't afford that, I like the PCPower supplies. And it is a proven fact that fanless systems use more electricity, simply due to reduced conductivity as heat levels rise.
Way off topic. This is a "Turbo-Cool" model, so I would expect loud noises, but then again if my system used this PSU I would sequester it in a sound-proof room.
Come to think of it, what do you actually interact with on a PC that requires it to sit next to you? (I personally program microcontrollers and interface directly with home-made circuits, and of course overclocking requires you to make adjustments at times.) I hardly ever even load a CD/DVD anymore, and I could use an external unit on my desk. Video and input are Digital (DVI and USB), so why is the system sitting around the desk? 25' of cable should put the system well away from your sensitive ears. In fact it is only lack of space, or I would do it, we simply don't always have the space to closet the system.
I love that PC Power is getting the attention it deserves, so often people ridicule your PSU if it doesn't have a mirror finish and a glowing fan, and a silly mainstream " 'clockers 'tude ".
I don't think it is misleading. PCP&C does not make any silent power supplies. You have a quiet power supply, but that doesn't make it silent. Even passively cooled PSUs are subject to coil whine and other related noises that make them not silent, however, they are very quiet. Silence is an absolute.
Rather than use the word "silent", the description should be "cannot be heard".
Logic would dictate that if you are using an 860W power supply, those components you are powering would also not be "silent".
However, most likely in such a system, the psu may not be audible (i.e. cannot be heard) above the components summed noise level (HDD's, CPU & VGA cooling fans, case cooling fans, etc.).
That considers that the space in which the system is operating is completely silent - which is almost impossible to have 0dB background noise.
Thus, the term "silent" is often interchangeable with "cannot be heard" over the background/component noise level. For example, in the outback of Australia during winter, the background noise level does not fall below 28-32dBA.
Most offices/dwellings have background noise levels at or above 35dBA thanks to HVAC noise / appliance noise / exterior traffic noise / etc. Therefore the pursuit of true silence is practically impossible in the REAL world. PSU manufacturers should not be held to such unrealistic targets in any event, as we the customer must pay extraordinary prices as a result.
Anandtech has a history of offering practical, realistic test results - my only objective of this posting is to help preserve that legacy.
We’ve updated our terms. By continuing to use the site and/or by logging into your account, you agree to the Site’s updated Terms of Use and Privacy Policy.
20 Comments
Back to Article
jamstan - Wednesday, July 2, 2008 - link
Sorry, but I can't give any credence to a review of a PSU that was specially built by the manufacturer for the reviewer. How bout buying a PSU from the Egg or walking into Best Buy and getting one, the same PSU we minions would buy, and test that? I can't help but wonder how "hand-picked" the PSU for this review was. This review is worthless.dawurz - Monday, February 11, 2008 - link
"While the fan runs is rated at the same speed as before"nitromullet - Friday, February 8, 2008 - link
Is the 860W a "true" TurboCool PSU with WIN-TACT internals or is it like the the Silencer line with Seasonic internals?Christoph Katzer - Monday, February 11, 2008 - link
Yeah 100%. Not Seasonic.nitromullet - Tuesday, February 12, 2008 - link
Awesome. Thx for the reply.Justin Case - Thursday, February 7, 2008 - link
[quote]We are still waiting for our 15 hard-drives to arrive, but we have talked with at least one AnandTech reader that successfully powered up 18 hard-drives with a single-rail Silencer 750W where all multi-rail power supplies failed.[/quote]Am I missing something here? I've been running a 16-drive RAID from a cheap 500W Fortron PSU for over one year with no issues. The drives don't even draw 200 watts while working and, even if they weren't set to staggered spin-up, probably wouldn't hit more than 400 during boot.
Stick to Zippy, Fortron and PCP&C, and you'll be fine. Nearly everything else is either a rebranded Zippy / FPS or a piece of crap.
nrb - Thursday, February 7, 2008 - link
[quote]As always, PCP&C provides a 7-year warranty for their power supply. We're not sure how that works with the 200,000 MTBF, as even running 24/7 for ten years wouldn't reach half that figure, but that's apparently a testament of their PSU quality.[/quote]This is a disgusting admission: if the writers at Anandtech don't understand how MTBF is calculated, what chance is there for the people who come to Anandtech to get technical information?
(sigh)
The way it is calculated is:
The manufacturer decides what a "normal" lifespan is for the product, and then runs a large number of models for that length of time, and sees how frequently they fail.
So, for example, they might decide that a "normal" lifespan for a power supply is 5 years. If the MTBF is 200,000 hours, that means that if you run 1000 power supplies continuously for 5 years, on average one of them will fail every 200 hours within the first five years. The figure tells you nothing whatever about what the failiure rate is after the first five years (although you can be sure it will be higher) and of course there is also no indication about what the manufacturer thinks is a "normal" lifespan for the product: maybe it's 5 years, maybe it's just 1.
This is naturally spectacularly misleading to the consumer who thinks it means that just one power supply will tyically run for 200,000 hours before failing. Sadly, he is wrong.
JarredWalton - Monday, February 11, 2008 - link
You're right, and I inserted the comment mostly tongue-in-cheek. The reality is that theoretical MTBF is only vaguely interesting. 1 million is better than 500K, and 500K is better than 250K... at least in theory. In practice, there's an element of "luck of the draw" - plus it depends on how the testing is conducted.So start with 1000 power supplies (or hard drives, or some other component). Make sure all of them function -- dead on arrival units don't count towards MTBF! Now start running all of them with a stopwatch and see how long can it takes before a failure occurs. In the case of PCP&C, it appears that out of 1000 units, the first failure occurred after 200 hours.
How many times do they repeat this sort of test? They probably do it once for each new model. Or maybe they run all of the power supplies until they have two failures? Probably not. The main point is that MTBF isn't a super useful figure, unless you're talking about operational MTBF -- which is only available after the fact, when a device has been on the market for a while. I consider the seven year warranty to be a better indication of quality than the MTBF.
Sorry for the confusion.
JEDIYoda - Monday, February 4, 2008 - link
Silence is not an absolute.Now absolute silence we all can agree is silent!
The word Silence means many things to many people.
I for one have PC Power & Cooling PSU`s in all my riggs.
I have no issues with what some would call the noise and others would call a lack of silence.
Over all a very nice review!!
pauldovi - Monday, February 4, 2008 - link
It seems to me that Anandtech is in bed with PC Power and Cooling.Can we get more love for Silverstone. Their OP / DA series are lovely.
HOOfan 1 - Tuesday, February 5, 2008 - link
You mean like thishttp://www.anandtech.com/casecoolingpsus/showdoc.a...">http://www.anandtech.com/casecoolingpsus/showdoc.a...
or this
http://www.anandtech.com/casecoolingpsus/showdoc.a...">http://www.anandtech.com/casecoolingpsus/showdoc.a...
I see a total of 2 Silverstone revies and 2 PC Power and Cooling Reviews.
JEDIYoda - Monday, February 4, 2008 - link
if they are(which they are not) they did a very accurate review as I am sure they would also do with a silverstone PSU!!KaosDrem - Monday, February 4, 2008 - link
I'm actually running 10 internal sata drives on the 750W silencer along with a quad core, and an 8800gts 512, PCPower&Cooling definitely makes the best power supplies and their method of sticking with the single rail i hope isn't something they will change in the future.yonzie - Monday, February 4, 2008 - link
Let me guess as to the purpose of those 15 harddrives and custom-made case & PSU...It's for Anand's media server!
jtleon - Monday, February 4, 2008 - link
Your Page 9 comment:"If you're buying a PCP&C power supply, you should not plan on silent computing - there just isn't any power supply in the lineup that can qualify as silent."
may be misleading, as my 610W Silencer operates silently in my dual Xeon workstation.
You might consider qualifying your statement more thoroughly.
Regards,
jtleon
Zap - Monday, February 4, 2008 - link
"Quiet" means different things to different people. I consider most computers to be noisy, but then again I've been "tainted" by SPCR. As for PCP&C PSUs, no, they're not silent IMO. They're not even particularly quiet. However, that's just IMO, just as you thinking your unit operates silently is in your opinion. The best way to quantify "quiet/silence" may be if the PSU on its own (not being drowned out by other fans) is noisier than ambient sounds in a quiet room.nubie - Monday, February 4, 2008 - link
I have a Silencer 470 and it is indeed quieter than the ambient noise, even when all the other fans are off.I do believe that the sound depends on the heat level and the power you are drawing, the most I ever pulled was a couple hard drives, optical drives and an overclocked x2 65watt and 7900GS @ 700mhz core. My PC's usually have some 120mm fans keeping the interior cool (better some slow-moving large fans then one really fast small one.)
As far as "silencer", I think it is apt, depending on the rest of your rig, even in the dead of night the Silencer fan is better than my fairly quiet hard drive, and it has a pleasing tone quality (not whiny or windy.)
That said, I am fully satisfied with my PCPower and I am thinking that "only" a 470watt was overkill, especially since the 360 silencer is $25 cheaper for 5A less on the 12volt rail ($55 direct!!), I can't recommend another PSU for mainstream overclocking (~3-4 Ghz, 1-2 drives, no SLi, or mainstream SLi, 7600/8600).
I too read SPCR, and I wish I had the money to do my own entirely passive, "heat-tunnel" rig (all the heat passes through one or two vertical pipes, and convection moves the air). Since I can't afford that, I like the PCPower supplies. And it is a proven fact that fanless systems use more electricity, simply due to reduced conductivity as heat levels rise.
Way off topic. This is a "Turbo-Cool" model, so I would expect loud noises, but then again if my system used this PSU I would sequester it in a sound-proof room.
Come to think of it, what do you actually interact with on a PC that requires it to sit next to you? (I personally program microcontrollers and interface directly with home-made circuits, and of course overclocking requires you to make adjustments at times.) I hardly ever even load a CD/DVD anymore, and I could use an external unit on my desk. Video and input are Digital (DVI and USB), so why is the system sitting around the desk? 25' of cable should put the system well away from your sensitive ears. In fact it is only lack of space, or I would do it, we simply don't always have the space to closet the system.
I love that PC Power is getting the attention it deserves, so often people ridicule your PSU if it doesn't have a mirror finish and a glowing fan, and a silly mainstream " 'clockers 'tude ".
Tiamat - Monday, February 4, 2008 - link
I don't think it is misleading. PCP&C does not make any silent power supplies. You have a quiet power supply, but that doesn't make it silent. Even passively cooled PSUs are subject to coil whine and other related noises that make them not silent, however, they are very quiet. Silence is an absolute.jtleon - Monday, February 4, 2008 - link
Hmmmm...You make an interesting argument.
Rather than use the word "silent", the description should be "cannot be heard".
Logic would dictate that if you are using an 860W power supply, those components you are powering would also not be "silent".
However, most likely in such a system, the psu may not be audible (i.e. cannot be heard) above the components summed noise level (HDD's, CPU & VGA cooling fans, case cooling fans, etc.).
That considers that the space in which the system is operating is completely silent - which is almost impossible to have 0dB background noise.
Thus, the term "silent" is often interchangeable with "cannot be heard" over the background/component noise level. For example, in the outback of Australia during winter, the background noise level does not fall below 28-32dBA.
Most offices/dwellings have background noise levels at or above 35dBA thanks to HVAC noise / appliance noise / exterior traffic noise / etc. Therefore the pursuit of true silence is practically impossible in the REAL world. PSU manufacturers should not be held to such unrealistic targets in any event, as we the customer must pay extraordinary prices as a result.
Anandtech has a history of offering practical, realistic test results - my only objective of this posting is to help preserve that legacy.
Regards,
jtleon
DigitalFreak - Monday, February 4, 2008 - link
*yawn*