Corsair RMi Series 1000 W

[crmaris]

Corsair introduced the RMi series at Computex 2015, and it offers some major upgrades compared to the previous RM line. Today, we will evaluate the strongest RMi unit with 1 kW capacity in order to figure out how it does against its predecessor and all other competing offers in its category.

Show full review

 

[LTUGamer]

HXi was the first reason to not buy AXi. And here is the second one

And pls don't say that 80+ Gold is not enough, everything is about DC quality

 

[R-T-B]

HXi was the first reason to not buy AXi. And here is the second one

And pls don't say that 80+ Gold is not enough, everything is about DC quality

The HXi line's DC quality was already excellent. I won't hesitate to admit I'm a little jealous seeing this "lower end" product having even better DC performance than mine, lol. Platinum is all well and good, but you are right, DC quality is king.

 

[damric]

I'm glad to see they pulled their heads out of their asses and put good caps in this thing. This has to be the best Channel Well product I have ever seen.

 

[Assimilator]

Don't like PSUs that use capacitors on the cables for ripple suppression, it's always felt a bit like cheating to me. Not to mention that doing so makes the cables more fragile - and in this case doesn't really seem to help the all-important 12V ripple, which somewhat negates their presence.

The "Modular Cables" table is wrong and/or confusing. It specifies 1x EPS connector/cable, but the unit includes 2x. It specifies 8x PCIe 6+2 cables, but there are actually 4x cables with 2 connectors each. And the SATA/Molex counts are completely indecipherable. I'd suggest splitting this table up by cable and connector counts to avoid confusion.

 

[crmaris]

6+2 pin PCIe (600mm+150mm) 8x describes the connectors, hence we have four cables.
The EPS are two of course. My bad there (when you have limited time to finish a review some things will go wrong).

SATA (400mm+100mm+100mm+100mm) 8x connectors. Basically the second column describes the connectors. I will revise the table in order to make it easier to understand.

 

[Joss]

Thanks for the test, thorough as usual.

I see some options on the Corsair Link screenshots (Profile, Configuration, Options). How granularly can you configure the fan speeds? Can you give it a profile where it is always on?
And if you tweak the options does the software need to run at startup? Or does it store the changes in the PSU?

 

[crmaris]

The fan speed can go as low as 40% of its max RPM and you can have the fan constantly running at least with a fixed speed.

About the changes, if they are stored in the PSU, I am not sure. I will ask jonny to comment on this one.

 

[Joss]

The fan speed can go as low as 40% of its max RPM and you can have the fan constantly running at least with a fixed speed.

About the changes, if they are stored in the PSU, I am not sure. I will ask jonny to comment on this one.

Thanks for the answer, I'll keep an eye on jonny's reviews.
The thing is: I dislike the fan curve options of most so-called silent PSUs. I'd prefer to have the fan spinning all the time albeit very slowly up to mid charge. After all I'm paying for the product so I should have the choice.
As to software everybody should do what Roccat (for example) does with its peripherals: tweak the settings and either uninstall or impede from startup.

 

[jonnyGURU]

Don't like PSUs that use capacitors on the cables for ripple suppression, it's always felt a bit like cheating to me.

Is it still cheating if it actually does what it's supposed to do? It's not like we're playing a game of Monopoly and someone drew two cards.

...and in this case doesn't really seem to help the all-important 12V ripple, which somewhat negates their presence.

It does? It seems those caps make a bit of a difference. Aris even stated that "Ripple suppression was outstanding!"

Not to mention that doing so makes the cables more fragile...

Legit concern. The ends of the 24-pin, EPS12V and PCIe are more bulky.

And if you tweak the options does the software need to run at startup? Or does it store the changes in the PSU?

The settings keep as long as the PSU is plugged into the wall and the rear switch is "on".

The thing is: I dislike the fan curve options of most so-called silent PSUs. I'd prefer to have the fan spinning all the time albeit very slowly up to mid charge. After all I'm paying for the product so I should have the choice.

Yeah. I have mine set to 40% and it rarely ramps up (two GTX670's on an AX760i). If I have it set to the Zero RPM mode, it stays quiet until I play a game and then it goes up and down as the temperatures change.

 

[Joss]

The settings keep as long as the PSU is plugged into the wall and the rear switch is "on".

Thanks.

 

[Assimilator]

Is it still cheating if it actually does what it's supposed to do? It's not like we're playing a game of Monopoly and someone drew two cards.

"Cheating" was the wrong word, it's more like CWT took the easy route and just stuck filter caps into the cables instead of spending the time refining their platform further. I also don't want to see a filter-caps-in-cables "arms rush" between PSU manufacturers in a quest to get that extra .1% stability.

It does? It seems those caps make a bit of a difference. Aris even stated that "Ripple suppression was outstanding!"

Yeah I'm dumb, I was looking at the 12V load regulation. D'oh!

 

[jonnyGURU]

"Cheating" was the wrong word, it's more like CWT took the easy route and just stuck filter caps into the cables instead of spending the time refining their platform further.

Well.... that's not 100% true. The RMi already had good ripple suppression without the caps. Yes, better engineering could yield better ripple suppression, but when the Corsair engineers (not CWT's) noticed that the ripple suppression was helped further by putting caps on the ends the same way Seasonic and SuperFlower/EVGA had been, they said "why not?"

 

[bogami]

Very nice results Well that's not all. I got chills when I see on the 12 V MOSFETs.
Such MOSFET does not belong to the power supply and there is no cooling element on them. My money and advice for $ 200 PSU will definitely go elsewhere.

 

[crmaris]

Thanks for the answer, I'll keep an eye on jonny's reviews.
The thing is: I dislike the fan curve options of most so-called silent PSUs. I'd prefer to have the fan spinning all the time albeit very slowly up to mid charge. After all I'm paying for the product so I should have the choice.
As to software everybody should do what Roccat (for example) does with its peripherals: tweak the settings and either uninstall or impede from startup.

Actually I meant that Jonny will reply since he deals with this stuff at Corsair. He doesn't write reviews any more

 

[Frick]

Very nice results Well that's not all. I got chills when I see on the 12 V MOSFETs.
Such MOSFET does not belong to the power supply and there is no cooling element on them. My money and advice for $ 200 PSU will definitely go elsewhere.

It comes with a seven year warranty, it'll work just fine.

 

[jonnyGURU]

Actually I meant that Jonny will reply since he deals with this stuff at Corsair. He doesn't write reviews any more

Right. Jeremy and Tony write the reviews now.

Very nice results Well that's not all. I got chills when I see on the 12 V MOSFETs.
Such MOSFET does not belong to the power supply and there is no cooling element on them. My money and advice for $ 200 PSU will definitely go elsewhere.

That MOSFET doesn't even need to be sinked, TBH. This is 2015, not 1995. That FET is VERY efficient and puts out very little heat, has a very high operating temperature and that BUSBAR is very effective for dissipating any heat there happens to be (it's solid copper). I actually get very tired of reviewers saying that the FET doesn't have "proper cooling". That tells me that they're not even reading the data sheets that they're linking to (sorry Aris).

 

[Joss]

Actually I meant that Jonny will reply since he deals with this stuff at Corsair. He doesn't write reviews any more

Thanks for the clarification.
I did a search and found this about him, although the information is a bit outdated.

 

[Delta6326]

If you go with completely new cables for custom sleeving and don't use the capacitors on cables will it cause harm or just not quite as good performance?

 

[jonnyGURU]

If you go with completely new cables for custom sleeving and don't use the capacitors on cables will it cause harm or just not quite as good performance?

Harmful? No. But ripple will be around 16 ~ 20mV instead of the results you see here in the review.

 

[crmaris]

That tells me that they're not even reading the data sheets that they're linking to (sorry Aris).

I actually read them since among others I always look at Rds (on) values and max current at high ambient. Actually inside the review I note down that CWT/Corsair is very confident about this design else you would't provide a 7 year warranty. Now according to my opinion with a proper heatsink you could lower noise even more since you would need even less airflow, but this is just my personal view in the subject.

 

[jonnyGURU]

I actually read them since among others I always look at Rds (on) values and max current at high ambient. Actually inside the review I note down that CWT/Corsair is very confident about this design else you would't provide a 7 year warranty. Now according to my opinion with a proper heatsink you could lower noise even more since you would need even less airflow, but this is just my personal view in the subject.

I'm not convinced you could reduce noise because bigger heatsinks would only dissipate heat, but not remove it. The fan still has to remove the heat from the housing.

 

[crmaris]

a heatsink would offer a larger area from which is easier to remove heat. Also the bus-bars that cooldown the +12V fets are from copper and this material might absorb heat faster but on the other hand aluminium dissipates it faster. Also in the specific place where this bus-bars are located the airflow isn't optimal. A heatsink with proper fins would definitely do a much better job.

 

[jonnyGURU]

Well... you're the one with the thermal camera. I'd love to see you use it on something like this. ;-)

But I digress.. it's been tested in a lab with multiple thermistors placed throughout and that busbar hardly gets hot. Agreed, a larger, aluminum heatsink would do a better job. But we're talking about something as insignificant as putting active cooling on heatsinked DDR3.

 

[GreiverBlade]

But we're talking about something as insignificant as putting active cooling on heatsinked DDR3.

awww i knew that my Corsair Vengeance Airflow standing above my G.Skill Trident X 2400 was only a cosmetic (wait ... can it still help in case of a water cooling to create a airflow for the VRM's, not that i don't think the 140mm i use on the back of my case is not enough (or that the airflow will be enough to come from the RAM area to the VRM area )

as for the review: good as usual, happy to see the RM series updated, this is the only line from corsair i like (tho seeing it's the same CWT as a HX make me wonder ... well at 200$ a 1k psu ... no worries in the end )

even if i don't like corsair for anything else than SP/AF120L, cases(when not overpriced)and cosmetic RAM cooler (personal opinion and experience ofc.)