Corsair CX750M PSU Review

Not open for further replies.


Bit early for a review,but just installed mine to replace my old one that was power surging.
Very easy to install and only use the extra cables you need.
Piece of mind knowing the 3 year warranty that comes with it.


Funny how it doesn't have OTP, when as a matter of fact, OTP was the very thing that saved the original CX750M's butt in Jonnyguru's testing; but the majority of people, rather than realizing that was a good safety measure, took it instead as "the CX750M can't do more than 650W" when in fact it was just Oklahomawolf's hot box. So maybe to avoid another mishap like this altogether they just removed OTP and cheered over the money savings at the same time?

I don't care for this PSU too much anyway. It seems to be a power supply that likes to focus on good ripple (as every modern PSU does these days) and decent voltage regulation but falls short in nearly every other aspect. I don't see it being much of any improvement over the original CX750M, the whole purpose of was probably just to cut costs. I'd happily take 60mv of ripple on my 12V rail in turn for some better holdup time, a higher quality fan and perhaps caps (if those Suscons aren't the best), and OTP.
I keep reading posts referencing the new CXM series, but the conclusions section which most readers skip to seem to follow Mom's advice ... "if ya don't have anything good to say, don't say anything" ... anything negative is left out. If the CX750M is going to be competitive, it has to address the elephant in the room that is the EVGA 750 B2 that sells for $50.



Aren't those usually either the 450w, 550w and 650 watts? Those are different units compared to the 750 watts I think.

Plenty of better options for power supplies around CX750M's price range, must be said. Isn't the EVGA B2 750w 65 dollars?
The CX450M, CX550M and CX650M were redesigned last year and manufactured by CWT using a custom Corsair design, while the CX750M (and I think the 850 model) was based upon CWTs PUQ B patform. This latest revision seems to be even newer .. and even beyond that, it appears to have changed yet again after the review samples went out as noted in the article the newer 650M's use a "470uF bulk cap but use instead a 330uF one".

Note that the CXM series is reported as the lowest quality Corsair PSU available in the US.... yet several VS models remain available thru US e-tailers. One thing I have always observed, specifically with regard to PSUs and coolers is that forum posters, even when referencing an article that says the reviewed item was a "good budget model" or "good for the money, tend to drop the words "budget" and "for the money" when recommending it. So while it may be a logical choice for a G3258 or GTX 1050 build w/ no overclocking, I don't quite understand why a Hyper 212 or a CX series PSU gets recommended for a 6700k / 1070 build especially where OP states an interest in overclocking the bejeezes out of everything.

Here's the only reviews I have seen besides this one:,4770.html

NCIX almost always has it for $50 in conjunction w/ Corsair MIR and instant savings... sometimes others like newegg

NCIX = $45 (Savings Code 97531-1714. SAVE $25.00 off our regular price of $89.99 Special price ends 11/30/2016 + Save $20.00 USD with manufacturer's mail-in rebate!)

newegg = $89.99 - $20 instant savings - $20.00 rebate card = $49.99


Sep 6, 2015
If i was short on budget i wouldn't hesitate to use Corsair's CX-M line of PSUs.
Solid units for that kind of price.
I don't expect to have Seasonic PRIME 750's performance with 80$ !!



If I needed 750W and it was between the EVGA 750 B2 and the CX750M, I'd probably take the former.

The Seasonic Prime comparison is however a false equivalency. The Antec HCG is built on Seasonic's AM platform (same as M12, XFX Core and XXX series) and is cheaper ... as is Corsair's "step above budget / step below premium" segment, the RMx platform. From EVGA, the B2 and GQ series would be ones I'd look at.

To my eyes, at 750 watts, the $45 EVGA B2 750 over the $80 Corsair CXM is the proverbial "no brainer". At this price range, we are talking "budget build" where the emphasis should be more on Build Quality than Performance and the B2 has the edge here for almost half the price.

The 650M was reviewed on Nov 3rd w/ a $65 price tag ... A Seasonic S12 620 is $50. Antec HCG 620 (same Seasonic Platform as S12) is $52... XFX TS is $62.. Seasonic M12 is $63



I just don't think not counting MIR and just leaving them out is not a far comparison personally, as you do pay the MIR right? Or you just don't pay them? I don't quite get that system. I do agree with the point, but using MIR prices and then leaving the actual MIR price out doesn't seem fair :p.

1. It is what it is ... especially when the MIRs are an "almost every every day" thing as w/ the B2. If we say"no discounts and MIRs, would we than have to list the CX650M at $106.98 in this case **in order to be fair** since the discount will be gone tmorrow ?

2. Not sure where the logic comes from on the MIR. Isn't the "net cost" the thing ? I only concerned with out of pocket ... which is what I take out minus what i put back in. I have oft gotten over $100 in MIRs which is not something I think should be ignored.

Put it this way, which one is the better deal ?

Option 1 - MSI 1070 GaminX w/ Free $60 game when it comes out (one ya want assumed) = $435
Option 2 - MSI 1070 Gaming X w/ nuttin = $420

Isn't it "what you get when all is said and done" that counts ?.... not getting the game is cheaper on day 1 but more expensive 6 weeks later when the game comes out. Isn't this the exact same thing with the MIR ?

Or what about ?

Option 1 - EVGA B2 = $100 w/ $25 MIR
Option 2 - EVGA B2 = $100 w/ no MIR

I'm gonna take Option 1 every time ...

3. All prices come from pcpartpicker. The prices from pcpartpciker include current prices **as delivered after all discounts**. If that wasn't "standard practice", then why is the entire PCPP web site based upon this structure ? Why aren't web site users up in arms and clamoring for change ?

It is the current practice here on this forum, is it not ? If you have ever posted a PCPP build on the forum, the total price includes all discounts, MIRs and shipping. Isn't this the real cost ? or, when you posted such builds, are you going back and editing ... removing all discounts and MIRs ?

It is also "standard practice" according general accounting rules and the US Government. If I pay $10,000 for a new plotter and get $2000 "cash back" and $2000 on a "trade-in", I only get to use the $6,000 "net cost" as a tax deduction. If I reported $10k and based my deductions on that, I risk a jail term. Of course ya still get screwed if ya pay sales tax, but if ya paying sales tax on on-line purchases, ya doing something wrong :)

3. However ... one more point. I did in fact provide the all applicable dollar values for the PSU I recommended (EVGA B2)....see 4 posts up from yours.

NCIX = $45 (Savings Code 97531-1714. SAVE $25.00 off our regular price of $89.99 Special price ends 11/30/2016 + Save $20.00 USD with manufacturer's mail-in rebate!)

newegg = $89.99 - $20 instant savings - $20.00 rebate card = $49.99
4. There's still many choices which are better / cheaper even w/o rebates.

Antec HCG - No MIRs

S12 - No MIRs

And yet jonnyguru gave the S12 a 9.7 rating, 9.4 on the M12 and the CXM (Version 2) falls a full point below at an 8.7 rating.

The S12 and Antec HCG were from the GB Bronze platform, the M12 520 and 620 same. The bigger M12s were from Seasonic's AM platform which also included Anyec HCG, CCC Core and TS series

I normally wouldn't compare PSU reviews from different size ranges but the S12 520 and 620 are build on same platform.. And we are talking about the GB Bronze platform here, the S12G was a different (G) platform.

While we don't use this level PSUs in any of our own bulds, we have used the S12 and M12 in user builds on numerous occasions have done numerous S12 / M12 builds. Can Not recall any fan noise issues. We had a EVGA 1000 G2 in which the fan was the noisiest thing in the PC ... other users reported here and in other forums that they experienced no such issue. So yes, there is some variability "luck of the draw" so to speak between samples. Except the unit failed after 18 months and the RMA replacement was just as noisy.

But if these are complaints about the S12... what about what even THGsays about the CXM .... among the cons are Efficiency, Sleeve-bearing fan, Hold-up time, 3.3V rail performance in Advanced Transient Response tests, Inrush current with 115V input. Distance between peripheral connectors and Fan speed at light loads could be lower (which would reduce noise at low loads). I am not saying that they are the best options (currently partial to Seasonic Snow and X Series myself) ... I am saying there's few options out there between $50 and $65 that are better and the CXM is not one of them.




Jonnyguru is not the only site that reviewed these units, though. HardOCP's review of the M12ii-620 shows extremely poor transient response performance - something just as important if not more than standard voltage regulation tests.

A 9.4 score 6 years ago is not the same as a 9.4 score today. Everyone agreed with this, scoring has gotten more strict and that PSU would not score that much today. In addition, score only tells you so much, so much more is there in the review.

The first S12ii units were so noisy that shortly after Seasonic had to change the fan entirely to a different one, which still isn't the quietest.
Again, we have had these units ... we have not seen issues... we always test every build for voltage instability is too high, it doesn't go out the door. We have used dozens of these units w/o a single noise issue. We have had CX and CXM series PSUs come in the door and they were definitely louder. Been here in this forum for over a dozen years and you just don't see complaints on the S12 and M12 units ... now compare that with the CX series where forums (i.e. that have entire threads dedicated to poor experiences with the CX / CXM series PSUs. More on fan noise later.

Look at what THG wrote about the CX750M .... hardly flattering and certainly not of the magnitude of the S12 comments. Your claim about reviews being more picky is arguable ... but look at it this way.

It's pretty much universally agreed that the original CX series was to be avoided. The CX-M was better but still not so hot. But today (... well Sep-12-2016) we read:

Last time, I remember recommending passing on this unit and going with the CS-M series instead. And that range is still superior to this one, I think.
And even more importantly ...

No way is anyone buying one of these when a Seasonic G-750 is the same price. Fry's is much better, but which version are we getting?
Is the noted similarity in packaging and naming convention intended to help clear the shelves of the old units that no one bought ?

The S12 520 and 620 were on the old GB Bronze platform, the 750 S12 is on the Seasonic G. So when JG says no one is going to be buying the CX750M when the Seasonic G is the same price, I am gonna be among the bunch that isn't buying them. I have had dozens of S12's in here and having not experienced any noise issues, your claim of noisy fans... if you pardon the pun "is gong to fall on deaf ears". But let's take a look at this. The Yate Loon sleeve bearing fans do not have a Fluid Dynamic Bearing (S12 does) and THG writes in this review that they wonder if it will last thru the 5 year warranty

We do know that the CX750M v2 uses the exact same fan as the previous CX750M (Yate Loon D14SH-12)....noise therefore can be expected therefore to be no different than the older model ... and we can dig up a long history about that. In addition to the above noted personal experience ...

I have a corsair cx750m and its fan is noticeably the loudest component in my system. Its a low-pitched humming sound with a little bit of a whining noise to it.
Now I have seen S12s and M12s with problems w/ fan noise ... but it was never endemic, some peeps had it some didn't. When i stayed with the thread, an RMA solved the problem

So i gotta ask ... Does the impact of noise change when a different label is on the side ? Surprised that you find the S12's FDB fan a problem but don't consider the CX750Ms history of noise issue of significance.

You used HardOCP as a **your source** to argue the poor of the s12, and yet HardOCP writes (your link)

to date the single quietest power supply we have reviewed (the X-750) was produced under Seasonic's own brand so we had great hopes that the M12II-620Bronze would follow more in its footsteps. These hopes were largely realized as the M12II-620Bronze stuck to Seasonic's roots of being almost dead silent. In general, the M12II-620Bronze easily matches all other Seasonic units, save for the X-750, that we have seen to date in its perceived noise output. There was no time during our load testing or working with the unit that it became audible in our test environment. As such, with the M12II-620Bronze Seasonic continues to produce products that are certainly appropriate for quiet cooling environments.
On to the transient load tests.... which you indicate is "something just as important if not more than standard voltage regulation tests". I agree that it is certainly an important test ... but it is important in a different respect than voltage stability. Certainly variances in voltage stability beyond the ATX spec are worrisome in that you can damage equipment if you stray outside that realm. The same is true with regard to transient tests as, again, if you exceed the ATX spec, you can damage equipment. But as long as you stay inside the ATX spec, the threat is just not something I am too concerned about in a $65 PSU.

I don't have that equipment so I can't address the issue of what *a* reviewer got when others didn't. I can only rely on my own experience, and what I have read in the forums over the last 6 years ... and that doesn't include a lot of complaints from S12 owners.

And while you see this comment about transient response being significant for the S12, somehow the same comment by THG on both reviews with regard to the CX650M and CX750M has no significance. Why is it that's a big negative for the S12 and of no consequence for the CXMs ? Yes, the transient response is not all that great, but none of them are $150 PSU ... your own reference goes on to say...

On the downside, the older group regulated design shows some weaker 12v regulation and transient load response. However, with a current price of $89.99 it is hard to find very many other units on the market currently that can continuously outperform the M12II-620Bronze at the same price point.
So yes, it doesn't compete with high end units, but at this price range it a) isn't expected to, b) you aren't going to find many units at this price range that can compete with it from a quality standpoint and c) your own reference that you used to imply poor quality of the unit gave the danged thing an Editor's Choice award !

Let's look at a more common thing we see in every day life. We all want to own a waterproof watch and a waterproof cell phone for when we fall in a pool or wade / swim in a lake. And while it's kinda cool to have devices that are waterproof to 100 foot depth ... is one that is waterproof to 100 feet as opposed to 50 feet of any real value if you're not "Diver Dan" ? So when your only risk is falling in a pool or swimming / boating in a lake where ya head barely ever dips below the surface ... is being waterproof to 100 feet really an issue ? If we are are talking about a $50 - $65 PSU, we are certainly not talking about a 4.8 GHz CPU OC and 28% overclocks on SLI'd GFX cards.

Group regulated designs are not known for great performance in crossload or transient response testing. Using group regulation is a means to cut costs and this is not considered a bad thing. Jonnyguru writes with regard to the S12

Those new to the site may need some explanation as to how these overshoot transient tests go. Well, these tests are intended to go looking for power on spikes that exceed ATX specifications. What you do is, you load the unit being tested to 100%. Then you apply power, all while watching the rails on an oscilloscope to see if you see a spike that exceeds 10% over mean value.

What do the above shots tell me? First, the 5VSB rail is very well controlled. There is a brief spike there before that rail settles in for the long haul, but it's nowhere near the spec. Over on the 12V side, even with me running the thing to full power from a "no power applied" condition as seen in the third shot up there, the only spike we get is a measly little 3V dimple in the scope shot. This is one of the better controlled units I've tested, I would have to say.

This unit does 80 Plus Bronze down to load levels you'd probably never see in real life, crossloads exceptionally well for a group regulated design, and has almost no ripple and noise to speak of. While I don't know that this unit deserves the scoring of a much more expensive independently regulated design, I also cannot punish a group design just for being a group design. Group designs aren't supposed to do well in crossload testing. And yet this one did. A half point off for voltage regulation that slipped outside the 1% number on the progressive tests, and that is all. 9.5.
When ya look at the CXM, I don't see you you can so easily ignore the noted poor efficiency, sleeve-bearing fan (if ya gonna use a sleeve bearing, at least provide FDB), poor hold-up time, poor performance in Advanced Transient Response tests, less than satisfactory inrush current with 115V input, inadequate distance between peripheral connectors, high fan speeds and resultant noise at light loads...and then choose it over another less expensive unit which reviews note just one of those things. And, on top of that, your chosen reference to support this conclusion, comes froma source that gave the unit you consider inferior an Editor's Choice award.
I noticed the S12ii mentioned a lot in the comments. Noise was mentioned quite a bit as being a problem with the S12ii. I can say with certainty that my Seasonic S12ii 520W with a max load applied of around 400W has never been loud to me. Its quieter then both the EVGA GTX 950 FTW and the Cooler Master Hyper 212 EVO. I've never sat next to the computer and thought to myself - oh that must be the PSU fan making that noise.

Perhaps I got a good one?



The S12-II was reviewed six and a half years ago.

The M12-II was reviewed more that three years ago.

Context, people.

Product are expected to, and for the most part, DO BETTER year over year.

You can't look at a six year old review's conclusion score and say, "yep. It's awesome."

Think about it. Using that logic, an AMD Phenom II X6 1090T is a perfect CPU for today.



I'm not saying it's a bad CPU. That wasn't my point. My point was that it's not apples to apples to compare something that was great six years ago against something that's great today.
It's different for power supplies than for CPUs or other components. Our CPUs need to handle heavier and heavier loads (at least in theory) whereas a 550W PSU only needs to handle up to a 550W load reasonably well be it now or ten years ago. If a PSU design is five years old, but it performs well, then it is still fine to use today without causing any problems and even to compare to modern PSUs.

If I happened to have a 550W unit from ten years ago that was never opened, but was a good unit for its time, I could use it just fine today. It'll be less efficient, sure, but it will work fine. Grabbing a ten years old CPU is a whole other matter. It's fine for web browsing or even light gaming with say Core 2, but an intense workload will show its age.

While it might not be an apples to apples comparison, it isn't far from it. PSU designs at their heart aren't changing much other than rail design. The biggest change in recent years was making the 3.3V and 5V lines use DC-DC converters from the 12V line. Seasonic's latest Prime platform even goes so far as to show that the new digital PSUs aren't beating good'ol analogue yet.


Agree.... but still not the point. And it's actually bothering me that you guys aren't getting the point. Makes me realize that you're just going to the last page and reading the summaries.

PSU's were certainly "adequate" 10 years ago. But ripple suppression has gotten better. Voltage regulation has gotten better. Transient response, etc. has all gotten better over the years. While this doesn't mean that a PSU that scored a 9+ 10 years ago isn't still a good PSU, when reviewing them, you can't fairly judge a PSU that got a 10 way back when against a PSU that has even better regulation, etc. today. Unless you change your scale to include 11, 12, etc. So, the scale has to move. A PSU with better voltage regulation, ripple suppression, etc. and scores a 9 today is far superior than a PSU that scored a 9 in the past. Once you grade on that curve, everything else has to be graded on the same curve. So if that same PSU scored a 9 in the past, it certainly would score a 9 today.

So yes... while a PSU that scored a 9 a long time ago is still "adequate", it's not as good as a PSU that scores a 9 today. Think about it. PSUs didn't have DC to DC so they wouldn't pass the C6/C7 sleep states. They had more ripple. The voltage dropped more going from zero to full load....
Not open for further replies.