Excellent review! Noted a couple of typo's in the board layout discussion. Regarding the capacitors, I think you meant "whopping". I'm sure you meant to say "coil whine" within the quotes.
There is another difference between the RMi and RMx series that the author is apparently unaware of: The RMi series uses a Fluid Dynamic Bearing fan, whilst the RMx series uses a Rifle Bearing fan, same as the RM series. The latter is supposed to not last as long, and also potentially be a little noisier. The difference is marginal though, so I think it's still very much justified to buy the RMx series if you don't need the Corsair Link. I'll definitely be looking at it for my next build :)
Although the "reviewer's guide" that I received says so as well, both of the units that I received had an NR135P fan installed. Perhaps the fan of the lower wattage models is different or Corsair ran out of NR135L's while making our samples. :)
Seems strange, as the FDB fan would be superior normally, so if they're saying it has it, and it then doesn't, that seems like false advertising even to me. I mean, it's not huge, but getting a different (and potentially inferior) fan to what you're supposed to be getting is not cool. Maybe all the production models have the FDB fan (the NR135L I guess) though in the RMi series...I'd hope so.
You're right, and having thought about it a little more I feel more strongly about it now too. This is false advertising if these are making it to customers like this. You're paying a premium for Corsair Link AND the longer lasting FDB fan, so it should damn well have it. @E.Fyll: Any chance of a follow-up with Corsair on that?
Actually... you have it backwards. Look at an old RM review and an old HXi review if you don't believe me. NR135L is the rifle bearing fan. NR135P is the FDB fan.
The NR135P is an FDB fan. The RM1000x is advertised as having a rifle bearing fan, but my sample also had a NR135P inside it (FDB).
I would assume that what happened is the exact opposite of what you are suggesting - Corsair ran out of NR135L's at some point and, instead of switching to something inferior, they chose to install a superior fan inside some of their RM1000x's.
Hmm, interesting...So then that means that buying a RMx series PSU you normally get an inferior fan to this review. I guess it should still be possible to draw conclusions from reviews of the old RM series on the fan's performance then though. Shame anyway, it robs us of a good direct comparison here. Thanks for the follow-up jonnyGURU!
Are sure that fan will not spinning until 500W load, even in longer load period? Because of i have Seasonic semi passivelly cooled PSU and fan not spinning worked only on paper, there was some temperature trigger set prettty low, so fan was spinning even in 150W,200W load after few minutes regardless of fan mode switch (PSU was Platinum 860W).
The fan uses an algorithm of load, temperature and duration. Similar to the Seasonic. The marketing represents the fan speed at different loads at 25°C. If your ambient temperatures are higher, the fan will start sooner.
Seasonic has been doing smaller footprint PSUs for a while with a 'fan only when needed' switch for a while now. For ex. a 750 Gold unit would top out at 975watts and do stay in platinum efficiency for most of the time.
Why are we getting all excited about a bearing technology 3 years after superior units jonnyGURU?
I can vouch for the quality of this product, the RMi1000W in particular. I have had one since late July when I got it for my 24/7 BOINC Crunching rig. It powers my overclocked i5 and my 2 GTX 970's along with the other components in the system, SSD, HDD, 110i GT AIO etc. The ambient temps in my home are around 22c and thus, even when my system is pulling ~600watts the fan does not turn on because the PSU itself is not heating up that much even under load. Also, the efficiency meter from the Link interface marks around 91.7%
The only knock one can give this PSU is the fact that the EPS, PCIE and ATX power cables are rather rigid and thick. It took some work to get the ATX cable into the motherboard. I brought up this complaint to Corsair. As of yesterday, Cablemod is now offering a set of EPS/PCIE/ATX individually sleeved cables for this PSU that I plan to purchase.
For anybody going to get or make custom cables for this PSU, Corsair changed the connector for the ATX 24-pin cable that plugs in to the PSU. The PSU is a 18 + 10 pin connector instead of the standard 14 + 10 pin.
They also have capacitors on the cables to aid in ripple suppression...so getting custom cables would give you different results from what is shown in the review
I've seen the 18+10 connector on a number of high end modular PSUs over the last year or so; I suspect it's the ODM's design not corsairs. Elsewhere I was told that the extra wires are used for voltage monitoring at the mobo end to actively adjust for any voltage drop.
It's not the ODM design. The extra 4-pins (actually, 3. But you can't have an odd number of pins on a connector) are for +12V sense, +5V sense and extra ground. Seasonic and Super Flower have been using this method in the past to improve voltage regulation. Corsair started doing this recently as well for the same reason.
To address the concern of using cables without sense and caps: You can do it, but you'll end up with the same level of voltage regulation and ripple suppression performance as an HXi.
But it's a cost adder and since it's only something you'll see in reviews and pretty much any decent PSU is well within Intel's (sloppy) ATX spec, it's not very common.
I do remember seeing that this was accounted for in the Cablemod custom cables and they do indeed have the proper cabling design to uphold the ripple suppression seen in the stock cables.
I am sorry, I read the table quickly and just saw the line EPS-8pin empty. However, I still don't get what sort of PSU I need to power, for example, an ASUS X99-E WS, which needs (I quote from the ASUS website): 2 x 8-pin EATX 12 V Power connector 1 x 6-pin EATX 12 V_1 Power connector(s) 1 x 24-pin EATX Power connector(s) 1 x 8-pin ATX 12V Power connector(s)
Looking at the manual (as opposed to the speclist) or pictures of the board it only has 1x24, 2x8, and 1x6 pin power connectors. The two 8 pin ones are at the top of the board and presumably are 8 pin EPS connectors. The 6pin is between the CPU and 1st PCIe slot, I assume it's a PCIe power connector. The two 8 pin and 1 6 pin connectors have opposite polarities (the 8pins have the 12V on the side with the clip, the 6pin has the ground on the clip side); which fits my recollection of differences between the two types of 12V connectors but I don't recall which is which. No clue where the supposed 3rd 8pin connector came from...
They do, generally in 4+4 type (eg the two EPS 4+4 Pin from the feature table here). If you have both halves of the 4+4 together it's an 8pin EPS connector. If you pop the plug apart you get a 4 pin "P4" connector suitable for use with older/low power systems or as supplemental power on some enthusiast boards. (My LGA 1151 board has both an 8 pin EPS and 4 pin P4 power connector, with the latter to provide more power for overclocking - not sure why they didn't just do 2x 8pin connectors honestly, even if they didn't need it for the CPU might be enough to remove the need for an optional PCIe6 connector around the expansion card area to power 3 way GPU setups.)
Probably because they couldn't ship that much current from above the cpu socket where that connector most likely was down to the pcie slots without significant voltage drop, making it ~pointless. With really high current on a PCB you want that connector as close to the point of load as possible. I mean even with 2 oz copper you are looking at a pretty fat trace, and boards these days are pretty packed ... So you are talking about possibly adding in 2 entire extra layers...
I wish that Anandtech made more reviews for PSU in the range of 500/750W.
Even entusiast PCs are maxed at that power. Even with multiple GPU.
It does not make sense to use more PSUs of higher wattage than needed, because efficiency drops, and efficiency on idle or light use (like browsing Internet or using office tools), is much worse with a high wattage PSU.
There is about a 2% difference in efficiency between these 1000W PSUs, and a 650W 80 plus Platinum PSU at 100W of DC usage. So, we are talking about using an extra 2 Watts...big whoop.
1- Those numbers are made up by you. 2- You are ignoring different vendors and models. 3- A modern, powerful rig may consume like 50W when browsing or doing office work. 4- There is nothing wrong in saving 1% of power by using a more appropriate PSU, and surely a cheaper one.
1- No they are made by actually reading reviews and looking at the numbers. Go find Kitguru's review of the RM1000x. They get 87.32% efficiency from the RM1000x at 100W. Go look at the Jonnyguru.com review of the EVGA P2 650 he gets 87.6% efficiency at 65W and 90.7% at 142W. So you are looking at around 88%-89% efficiency at 100W. Guess what, that is a 2%-3% difference. You haven't provided ANYTHING 2- This article was about the RM000i/RM000x...so why should I talk about different vendors? 3- At 50W, a 5% difference in efficiencies is 2.5W....so the lower the power draw, even with a difference in efficiency, you are still wasting less power than you would be at higher loads. 4- There is also nothing wrong with using an over rated PSU which can provide a lot more power before the fan kicks on. 5- These sites publish what brings in readers. Reviewer after reviewer has said that the high power PSU reviews bring in more viewers.
Or you could do a more apples to apples comparison from the Kitguru, since they use the same testing equipment and AC source
Corsair RM1000x 87.3% efficient at 100W DC draw Be Quiet Dark Power Pro 11 89.5% efficient at 110W DC draw Coolermaster V 550 88.6% efficient at 110W DC draw
So the 80 plus platinum 550W unit is only 2.2% more efficient than the 1000W 80 plus gold and the 80 plus gold 550W unit is only 1.3% more efficient than the 1000W 80 plus gold
In case you didn't read, I mean :"I wish that Anandtech made more reviews for PSU in the range of 500/750W", which overkills any powerful single GPU rig.
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.
47 Comments
Back to Article
HOOfan 1 - Wednesday, September 30, 2015 - link
Looks like you may have been getting 80 Plus Platinum efficiency from itHOOfan 1 - Wednesday, September 30, 2015 - link
Nevermind, I see you were using a 230V input, but it still exceeded 80 Plus Gold by a decent amountBearmann - Wednesday, September 30, 2015 - link
Excellent review! Noted a couple of typo's in the board layout discussion. Regarding the capacitors, I think you meant "whopping". I'm sure you meant to say "coil whine" within the quotes.E.Fyll - Wednesday, September 30, 2015 - link
Thank you Bearmann. The spellchecker/auto-correct are not as smart as they should be sometimes. :)Cow86 - Wednesday, September 30, 2015 - link
There is another difference between the RMi and RMx series that the author is apparently unaware of: The RMi series uses a Fluid Dynamic Bearing fan, whilst the RMx series uses a Rifle Bearing fan, same as the RM series. The latter is supposed to not last as long, and also potentially be a little noisier. The difference is marginal though, so I think it's still very much justified to buy the RMx series if you don't need the Corsair Link. I'll definitely be looking at it for my next build :)E.Fyll - Wednesday, September 30, 2015 - link
Although the "reviewer's guide" that I received says so as well, both of the units that I received had an NR135P fan installed. Perhaps the fan of the lower wattage models is different or Corsair ran out of NR135L's while making our samples. :)Cow86 - Wednesday, September 30, 2015 - link
Seems strange, as the FDB fan would be superior normally, so if they're saying it has it, and it then doesn't, that seems like false advertising even to me. I mean, it's not huge, but getting a different (and potentially inferior) fan to what you're supposed to be getting is not cool. Maybe all the production models have the FDB fan (the NR135L I guess) though in the RMi series...I'd hope so.Oxford Guy - Wednesday, September 30, 2015 - link
Actually it is huge. FDB fans should never be advertised and then switched out for crappier fans. Ever.Cow86 - Wednesday, September 30, 2015 - link
You're right, and having thought about it a little more I feel more strongly about it now too. This is false advertising if these are making it to customers like this. You're paying a premium for Corsair Link AND the longer lasting FDB fan, so it should damn well have it. @E.Fyll: Any chance of a follow-up with Corsair on that?jonnyGURU - Wednesday, September 30, 2015 - link
The RMi shouldn't have had a rifle bearing fan. Let me see what happened there.jonnyGURU - Wednesday, September 30, 2015 - link
Actually... you have it backwards. Look at an old RM review and an old HXi review if you don't believe me. NR135L is the rifle bearing fan. NR135P is the FDB fan.E.Fyll - Wednesday, September 30, 2015 - link
True. That was a mistake on my part. The NR135P is the FDB fan.E.Fyll - Wednesday, September 30, 2015 - link
The NR135P is an FDB fan. The RM1000x is advertised as having a rifle bearing fan, but my sample also had a NR135P inside it (FDB).I would assume that what happened is the exact opposite of what you are suggesting - Corsair ran out of NR135L's at some point and, instead of switching to something inferior, they chose to install a superior fan inside some of their RM1000x's.
jonnyGURU - Wednesday, September 30, 2015 - link
Exactly.extide - Wednesday, September 30, 2015 - link
So we can all put this to bed!buxe2quec - Thursday, October 1, 2015 - link
Some clarification about fluid dynamic bearings, rifle bearings, ... http://www.hardwaresecrets.com/the-truth-about-flu... It's quite complete, and shows how the naming scheme is not always coherent.Cow86 - Thursday, October 1, 2015 - link
Hmm, interesting...So then that means that buying a RMx series PSU you normally get an inferior fan to this review. I guess it should still be possible to draw conclusions from reviews of the old RM series on the fan's performance then though. Shame anyway, it robs us of a good direct comparison here. Thanks for the follow-up jonnyGURU!ruthan - Wednesday, September 30, 2015 - link
Are sure that fan will not spinning until 500W load, even in longer load period? Because of i have Seasonic semi passivelly cooled PSU and fan not spinning worked only on paper, there was some temperature trigger set prettty low, so fan was spinning even in 150W,200W load after few minutes regardless of fan mode switch (PSU was Platinum 860W).jonnyGURU - Wednesday, September 30, 2015 - link
The fan uses an algorithm of load, temperature and duration. Similar to the Seasonic. The marketing represents the fan speed at different loads at 25°C. If your ambient temperatures are higher, the fan will start sooner.Cellar Door - Thursday, October 1, 2015 - link
Seasonic has been doing smaller footprint PSUs for a while with a 'fan only when needed' switch for a while now. For ex. a 750 Gold unit would top out at 975watts and do stay in platinum efficiency for most of the time.Why are we getting all excited about a bearing technology 3 years after superior units jonnyGURU?
XFX units(based on them), were a steal..
jonnyGURU - Thursday, October 1, 2015 - link
Uhhh.... "Excited"? I was simply answering his question.elforeign - Wednesday, September 30, 2015 - link
I can vouch for the quality of this product, the RMi1000W in particular. I have had one since late July when I got it for my 24/7 BOINC Crunching rig. It powers my overclocked i5 and my 2 GTX 970's along with the other components in the system, SSD, HDD, 110i GT AIO etc. The ambient temps in my home are around 22c and thus, even when my system is pulling ~600watts the fan does not turn on because the PSU itself is not heating up that much even under load. Also, the efficiency meter from the Link interface marks around 91.7%The only knock one can give this PSU is the fact that the EPS, PCIE and ATX power cables are rather rigid and thick. It took some work to get the ATX cable into the motherboard. I brought up this complaint to Corsair. As of yesterday, Cablemod is now offering a set of EPS/PCIE/ATX individually sleeved cables for this PSU that I plan to purchase.
LaRock0wns - Wednesday, September 30, 2015 - link
For anybody going to get or make custom cables for this PSU, Corsair changed the connector for the ATX 24-pin cable that plugs in to the PSU. The PSU is a 18 + 10 pin connector instead of the standard 14 + 10 pin.HOOfan 1 - Wednesday, September 30, 2015 - link
They also have capacitors on the cables to aid in ripple suppression...so getting custom cables would give you different results from what is shown in the reviewlozikosaz - Wednesday, September 30, 2015 - link
Aren't capacitors useless in interior cables, which use direct current?AnnihilatorX - Thursday, October 1, 2015 - link
It will help suppress high frequency noise.DanNeely - Wednesday, September 30, 2015 - link
I've seen the 18+10 connector on a number of high end modular PSUs over the last year or so; I suspect it's the ODM's design not corsairs. Elsewhere I was told that the extra wires are used for voltage monitoring at the mobo end to actively adjust for any voltage drop.jonnyGURU - Wednesday, September 30, 2015 - link
It's not the ODM design. The extra 4-pins (actually, 3. But you can't have an odd number of pins on a connector) are for +12V sense, +5V sense and extra ground. Seasonic and Super Flower have been using this method in the past to improve voltage regulation. Corsair started doing this recently as well for the same reason.To address the concern of using cables without sense and caps: You can do it, but you'll end up with the same level of voltage regulation and ripple suppression performance as an HXi.
extide - Wednesday, September 30, 2015 - link
Cant believe it's taken this long for high end PSU's to implement separate sense lines. Any decent lab grade PSU has had that for decades, lol.jonnyGURU - Thursday, October 1, 2015 - link
Correct!But it's a cost adder and since it's only something you'll see in reviews and pretty much any decent PSU is well within Intel's (sloppy) ATX spec, it's not very common.
elforeign - Wednesday, September 30, 2015 - link
I do remember seeing that this was accounted for in the Cablemod custom cables and they do indeed have the proper cabling design to uphold the ripple suppression seen in the stock cables.FriendlyUser - Wednesday, September 30, 2015 - link
Great review as usual. I was wondering why nobody makes PSUs with EPS 8pin outputs... Such high-wattage units would be perfect for a dual-CPU WS.jonnyGURU - Wednesday, September 30, 2015 - link
What do you mean? MOST PSUs have EPS 8-pin connectors. Typically two for dual CPU WS's.FriendlyUser - Thursday, October 1, 2015 - link
I am sorry, I read the table quickly and just saw the line EPS-8pin empty. However, I still don't get what sort of PSU I need to power, for example, an ASUS X99-E WS, which needs (I quote from the ASUS website):2 x 8-pin EATX 12 V Power connector
1 x 6-pin EATX 12 V_1 Power connector(s)
1 x 24-pin EATX Power connector(s)
1 x 8-pin ATX 12V Power connector(s)
DanNeely - Thursday, October 1, 2015 - link
Looking at the manual (as opposed to the speclist) or pictures of the board it only has 1x24, 2x8, and 1x6 pin power connectors. The two 8 pin ones are at the top of the board and presumably are 8 pin EPS connectors. The 6pin is between the CPU and 1st PCIe slot, I assume it's a PCIe power connector. The two 8 pin and 1 6 pin connectors have opposite polarities (the 8pins have the 12V on the side with the clip, the 6pin has the ground on the clip side); which fits my recollection of differences between the two types of 12V connectors but I don't recall which is which. No clue where the supposed 3rd 8pin connector came from...DanNeely - Wednesday, September 30, 2015 - link
They do, generally in 4+4 type (eg the two EPS 4+4 Pin from the feature table here). If you have both halves of the 4+4 together it's an 8pin EPS connector. If you pop the plug apart you get a 4 pin "P4" connector suitable for use with older/low power systems or as supplemental power on some enthusiast boards. (My LGA 1151 board has both an 8 pin EPS and 4 pin P4 power connector, with the latter to provide more power for overclocking - not sure why they didn't just do 2x 8pin connectors honestly, even if they didn't need it for the CPU might be enough to remove the need for an optional PCIe6 connector around the expansion card area to power 3 way GPU setups.)extide - Wednesday, September 30, 2015 - link
Probably because they couldn't ship that much current from above the cpu socket where that connector most likely was down to the pcie slots without significant voltage drop, making it ~pointless. With really high current on a PCB you want that connector as close to the point of load as possible. I mean even with 2 oz copper you are looking at a pretty fat trace, and boards these days are pretty packed ... So you are talking about possibly adding in 2 entire extra layers...FriendlyUser - Thursday, October 1, 2015 - link
You're right, I misread the EPS-8 line which was empty and did not notice the 4+4 part. Thanks for your comment.lozikosaz - Wednesday, September 30, 2015 - link
I wish that Anandtech made more reviews for PSU in the range of 500/750W.Even entusiast PCs are maxed at that power. Even with multiple GPU.
It does not make sense to use more PSUs of higher wattage than needed, because efficiency drops, and efficiency on idle or light use (like browsing Internet or using office tools), is much worse with a high wattage PSU.
Arbie - Wednesday, September 30, 2015 - link
@lozikosaz - What do you mean? By my count, at least six people are interested in this, though of course not all will buy it.HOOfan 1 - Thursday, October 1, 2015 - link
There is about a 2% difference in efficiency between these 1000W PSUs, and a 650W 80 plus Platinum PSU at 100W of DC usage. So, we are talking about using an extra 2 Watts...big whoop.lozikosaz - Thursday, October 1, 2015 - link
1- Those numbers are made up by you.2- You are ignoring different vendors and models.
3- A modern, powerful rig may consume like 50W when browsing or doing office work.
4- There is nothing wrong in saving 1% of power by using a more appropriate PSU, and surely a cheaper one.
HOOfan 1 - Thursday, October 1, 2015 - link
1- No they are made by actually reading reviews and looking at the numbers.Go find Kitguru's review of the RM1000x. They get 87.32% efficiency from the RM1000x at 100W. Go look at the Jonnyguru.com review of the EVGA P2 650 he gets 87.6% efficiency at 65W and 90.7% at 142W. So you are looking at around 88%-89% efficiency at 100W. Guess what, that is a 2%-3% difference.
You haven't provided ANYTHING
2- This article was about the RM000i/RM000x...so why should I talk about different vendors?
3- At 50W, a 5% difference in efficiencies is 2.5W....so the lower the power draw, even with a difference in efficiency, you are still wasting less power than you would be at higher loads.
4- There is also nothing wrong with using an over rated PSU which can provide a lot more power before the fan kicks on.
5- These sites publish what brings in readers. Reviewer after reviewer has said that the high power PSU reviews bring in more viewers.
HOOfan 1 - Thursday, October 1, 2015 - link
Or you could do a more apples to apples comparison from the Kitguru, since they use the same testing equipment and AC sourceCorsair RM1000x 87.3% efficient at 100W DC draw
Be Quiet Dark Power Pro 11 89.5% efficient at 110W DC draw
Coolermaster V 550 88.6% efficient at 110W DC draw
So the 80 plus platinum 550W unit is only 2.2% more efficient than the 1000W 80 plus gold and the 80 plus gold 550W unit is only 1.3% more efficient than the 1000W 80 plus gold
lozikosaz - Wednesday, September 30, 2015 - link
@ArbieIn case you didn't read, I mean :"I wish that Anandtech made more reviews for PSU in the range of 500/750W", which overkills any powerful single GPU rig.
Arbie - Friday, October 2, 2015 - link
@lozikosaz - I was being facetious... I agree with you.paul878 - Sunday, October 4, 2015 - link
Too many models Corsair, your product line is getting confusing.