Corsair RM550x Power Supply Review

A Look Inside And Component Analysis

Parts Description

Before proceeding with this page, we strongly encourage you to a look at our PSUs 101 article, which provides valuable information about PSUs and their operation, allowing you to better understand the components we're about to discuss. Our main tools for disassembling PSUs are a Thermaltronics soldering and rework station, and a Hakko 808 desoldering gun.

Primary Side
Transient Filter4x Y caps, 2x X caps, 2x CM chokes, 1x MOV
Inrush ProtectionNTC Thermistor
Bridge Rectifier1x GBU1506 (600V, 15A @ 100 °C)
APFC MOSFETs1x Vishay SIHG30N60E (650V, 18A @ 100 °C, 0.125 ohm)
APFC Boost Diode1x Panjit SiC04A065T (600V, 4A @ 150 °C)
Hold-up Cap1x Nippon Chemi-Con (400V, 470uF each, 2000h @ 105 °C, KMQ)
Main Switchers2x Infineon IPA50R280CE (550V, 4.7A @ 100 °C, 0.28 ohm)
APFC ControllerInfineon ICE3PCS01G - CM03X
Switching ControllerInfineon ICE2HS01G
TopologyPrimary side: Half-Bridge & LLC Resonant Converter
Secondary side: Synchronous Rectification & DC-DC converters
Secondary Side
+12V MOSFETs4x Sinopower SM4021NAKP (40V, 100A @ 100 °C, 2.7 mohm @ VGS=6V)
5V & 3.3VDC-DC Converters: 2x M3006D & 4x M3004D FETs
PWM Controller: APW7159
Filtering CapacitorsElectrolytics: Chemi-Con (105 °C, KZE & KZH series)
Polymers: Nippon Chemi-Con
Supervisor ICWeltrend WT7502 (PG, *OVP, *UVP, *SCP)  * for 5/3.3V only
Fan ModelNR135L (12V, 0.22A, Rifle Bearing)
5VSB Circuit
RectifierPFR20V45CT (45V, 20A, VF: 0.42V max @ 125 °C)
Standby PWM ControllerOn-Bright OB5269

There are only three heat sinks: a very small one for the bridge rectifier and two thin ones without any fins. The first of those is for the APFC FET and diode, while the second is for the primary switchers. On the secondary side, the +12V FETs are cooled by the PCB that holds them and two metallic bars. The minor rails are generated by two DC-DC converters installed on a large vertical daughterboard.

Channel Well Technology (CWT) is the OEM that Corsair uses for its RM, RMi and RMx units. Inside, the RMx platform is similar to the RMi. Of course, there is no digital interface to facilitate monitoring and fan control. And although the RMi PSUs are equipped with a fluid dynamic bearing (FDB) fan, the RMx models use a lower-quality rifle bearing fan.

On the AC receptacle, two Y caps form the first part of the EMI filter. The latter continues on the main PCB with two more Y caps and two X ones, two CM chokes and an MOV. The EMI filter on the main PCB is protected by a copper shield to further suppress EMI noise. If this shield was grounded, it would definitely do a better job.

A small heat sink dissipates thermal energy from the single bridge rectifier, a GBU1506. This rectifier can handle up to 15A of current, so it covers the RM550x's needs.

The APFC converter uses a single Vishay SIHG30N60E FET and Panjit SiC04A065T boost diode. The bulk cap is by Chemi-Con (400V, 470uF, 2000h @ 105 °C, KMQ), and its capacity looks to be ideal for the RM550x. In front of the PFC, coil we find a small NTC thermistor that provides protection against large inrush currents. Although a bypass relay isn't used, we suspect that the diode right beside it somehow isolates it from the circuit once the PSU starts.

A small daughterboard houses the APFC controller, an Infineon ICE3PCS01G that's backed up by a CM03X Green PFC controller. On the same board, there's also an Infineon ICE2HS01G LLC resonant controller.

A couple of Infineon IPA50R280CE FETs are the primary switchers; they're arranged into a half-bridge topology.

Here are the inductive and capacitive parts of the LLC resonant converter.

The lack of any heat sinks on the secondary side makes quite an impression. This might be an efficient Gold-rated power supply, but we still expected at least a small sink for the +12V FETs. CWT obviously had a different opinion, and Corsair must be very confident about its design or it wouldn't offer such a long warranty.

The FETs that generate the +12V rail are four Sinopower SM4021NAKPs, all installed on a small PCB. The board and two metallic bars help cool these FETs. We also noticed that two short, thick cables connect the main transformer's output with the +12V PCB, minimizing energy losses.

Four Chemi-Con electrolytic caps (rated at 105 °C) filter the +12V rail. All of the PSU's other filtering caps come from the same manufacturer.

The PCB shown in the photos above houses both DC-DC converters responsible for generating the minor rails. The common PWM controller is an Anpec APW7159. In total, two M3006D and four M3004D FETs are used.

The supervisor IC is a Weltrend WT7502 IC, which is installed on the main PCB. This IC provides only the very basic protections including OVP, UVP and SCP for the minor rails. More than likely, CWT uses another circuit to implement +12V protection features.

The 5VSB rail is regulated by a PFR20V45CT Schottky Barrier Diode (SBR) installed on the PCB's component side. The standby PWM controller (an On-Bright OB5269 IC) is installed on the solder side of the main PCB. Right next to it is a M03N65D FET; we don't know what purpose it serves.

The RMi family's MCU board is populated with a PIC32MX, which is the brains of the digital interface. That IC is missing in the RMx line-up. And because the cooling fan isn't PWM-controlled, its fan header only has two pins instead of four.

On the front side of the modular PCB, many Chemi-Con capacitors provide an extra layer of ripple suppression.

A number of cables on the back of the modular board transfer power. We also spotted a single M3004D FET back there.

Overall, the soldering quality is pretty good.

Here are some closer shots of the main PCB's solder side. We noticed two current shunt resistors under the +12V islands, suggesting that this platform can support OCP for at least two +12V rails with the proper supervisor IC. There is space on the modular PCB for two extra supervisor ICs, which you find installed on members of the RMi family.

Again, Corsair uses a rifle bearing fan on its RMx units in an effort to lower their prices. The NR135L (135mm, 12V, 0.22A) is a low-speed fan featuring quiet operation, even when it's spinning quickly.

This thread is closed for comments
36 comments
    Your comment
  • Dark Lord of Tech
    I just wish they would get the price down , these are a great lineup \ RX.
  • basroil
    From the performance it seems like CWT is finally something to consider... It's showing Leadex Gold/Seasonic levels of performance.
  • Dark Lord of Tech
    I love the RMX and RMI series , price just keeps me away from the purchase , very very solid.

    Knock a little off the price and these would fly out of warehouses.
  • William Henrickson
    They were on sale when new. I snagged an RM750i for $105 -w- shipping
  • Dark Lord of Tech
    Yeah the 550w should be about 79.99 to 89.99 , no rebates.

    Then I would grab a few.
  • JQB45
    Quote:
    Yeah the 550w should be about 79.99 to 89.99 , no rebates. Then I would grab a few.


    Corsair RMx 550W 80+ Gold Certified Fully-Modular ATX Power Supply
    http://pcpartpicker.com/part/corsair-power-supply-cp9020090na
    $79.99

    $89.99 for the 650W version.

    UPDATE:

    Sorry thats with mail in rebates...
  • Dark Lord of Tech
    I don't do rebates , takes to long , and maybe you get it maybe you don't , I'll wait for a newegg drop.
  • Nintendork
    We really need more platinum/titanium PSU's at 300-500w. Most PC's stays near idle and with the efficiency focused gpu's/cpu's they rarely exceed 100w unless you tax them.
  • turkey3_scratch
    Plus companies often don't even send you the rebates, sometimes they just say it was too late or some other bull crap like that. I agree with Blackbird. I've been waiting for a review of the 550 RMx, and what I get out of this review is that it trades blows with the 550 G2 that saying one or the other is better is just silly and extremely nit-picky. They are both incredible. Both offer a 7 year warranty, as only higher-wattage G2s offer the 10 year warranty. They are just so close, that when it comes to picking the better one, the cheaper one is better, and the G2 is cheaper.

    I've actually quit including rebates in my pcpartpicker lists. They are a pain and I don't think they reflect the true cost of an item.
  • turkey3_scratch
    205977 said:
    We really need more platinum/titanium PSU's at 300-500w. Most PC's stays near idle and with the efficiency focused gpu's/cpu's they rarely exceed 100w unless you tax them.


    I wish so, but unfortunately if this were to happen they would end up priced the same as any Platinnum/Titanium 650W unit. It's just the way it works. Quality low-wattage models are priced almost the same as the higher-wattage models. I would like to see something like a Titanium 250W model come out from Seasonic. Something like $40, fully modular. Will never happen, though.
  • logainofhades
    Yea the price for these is a big turn off. It is selling at a price that the Evga G2 750w is often seen at. The Evga P2 650w even sells for less. There are quite a few Seasonic and Superflower based units, cheaper than this, actually.
  • turkey3_scratch
    The Seasonic Platinum Series is quite expensive, and much of the X-Series is not even sold anymore. Only the S12ii comes to my mind as an option under $60.
  • logainofhades
    The cheapest RM550x is $99, before rebate, at newegg. You can get a G2 and P2 Evga for same or less. Seasonic 550g is selling for $69. Even if I needed a fully modular unit, there are less expensive models from Evga, XFX, and Antec. For a few $$ more, I can get a GS 850w or a G2 750w. The pricing on this unit is just not good enough to compete. Even their own 650w RMX is only about $10 more.

    http://pcpartpicker.com/parts/power-supply/#e=5,4&W=550,2000&sort=a9&page=1&m=52,11,14,71,39&p=1,2
  • turkey3_scratch
    Well I know that's what I said, I can't find a good argument as to why one should purchase the RMx for $20 more than the EVGA G2, so I just don't recommend it.
  • powernod
    I've got one very good reason for you: The fantastic performance of Corsair's /CWT's platform at the new "Power-OK signal" tests that Aris added at his reviews. On the contrary, the performance of Leadex's platform (*in which EVGA's G2 units are based) in these tests is very discouraging ( http://www.tomshardware.com/reviews/super-flower-leadex-gold-550w-power-supply,4416-4.html )
  • logainofhades
    Still cannot see spending $99 on a 550w Corsair, when I can get a 650w P2, for the same price. Heck, the Seasonic built GS 850w is selling for a mere $3 more.
  • turkey3_scratch
    Yes G2 hold-up time is 12ms whereas ATX spec is 16ms, but that doesn't mean the PSU is going to cause damage to the computer, it just means that if there is AC loss for 12ms the computer will shut down, as the PSU won't be able to maintain the DC. Better hold-up time is not worth the additional $20, especially if you're not running something like a server, in which you would have a UPS anyway that quickly kicks on the battery power in 12ms or less (you'd do research to buy one that does so, as some may be higher than 12ms).

    I'd say the hold-up time is a small gripe and not enough for a $20 more PSU. If hold-up time as super important, every PSU reviewer site would test it like Jonnyguru.
  • powernod
    Quote:
    Better hold-up time is not worth the additional $20, especially if you're not running something like a server, in which you would have a UPS anyway that quickly kicks on the battery power in 12ms or less (you'd do research to buy one that does so, as some may be higher than 12ms).


    Indeed, those who have a good UPS to protect their system, they shouldn't worry about that, but if you don't have a UPS, this failure at Hold-UP time /Power-OK signal, could be very dangerous.
  • turkey3_scratch
    2077519 said:
    Quote:
    Better hold-up time is not worth the additional $20, especially if you're not running something like a server, in which you would have a UPS anyway that quickly kicks on the battery power in 12ms or less (you'd do research to buy one that does so, as some may be higher than 12ms).
    Indeed, those who have a good UPS to protect their system, they shouldn't worry about that, but if you don't have a UPS, this failure at Hold-UP time /Power-OK signal, could be very dangerous.


    How would hold-up time be dangerous?
  • powernod
    I didn't say only Hold-up Time, I was referring at Aris's new test for Power-OK signal. Take a look of one of his analysis : " The hold-up time tests don't go well. Not only is the measured hold-up less than 16ms (the ATX spec's minimum), but the Power_OK signal drops after, and not before, the PSU's rails go out of spec. This means that your motherboard gets a false power-good signal from the PSU. Indeed, we measured the +12V rail floating at around 10.8V when Power_OK dropped to zero. This is a very low voltage level that applies lots of stress to the voltage regulators of components fed by +12V " ( http://www.tomshardware.com/reviews/super-flower-leadex-gold-550w-power-supply,4416-4.html )
  • turkey3_scratch
    That is the Superflower Leadex 550W where the Pwr_OK signal drops after the voltage is below 10.8V, and while the EVGA 550 G2 is based off that unit, Aris does not mention that same behavior for the G2 on the 550 G2 review. So it is possible the 550 G2 does not have this same mistake, but I am unsure myself. They could have made modifications.

    Edit: Ahh, he didn't check AC loss to PWR_OK in the G2 review, so the behavior is most likely still present.
  • turkey3_scratch
    This behavior is also present on Sirfa-made units, which have very high AC loss to PWR_OK off hold up times because their PWR_OK signal does not turn off until the voltage is somewhere below like 10.2V! http://www.jonnyguru.com/forums/showthread.php?t=12886

    The real question though is how fatal, in terms of the 550 G2, this really is. Is the 10.8V+ applied to the unit's VRM components for 12ms really going to harm the unit fatally? If anything, those Sirfa units are more dangerous. A higher AC loss to PWR_OK loss time is not necessarily good if the voltage is too low. They should be setting the threshold to 11.4V, not these lower values, so shame on Superflower and Sirfa for this. But it's actually a good thing the 550 G2 is only 12ms, you don't want those low voltages to be handled by the 12V VRM components for any longer, shorter is better in this case. But yes you are correct, they should improve these, but again does this merit the $20 extra for the 550 RMx? I'd have to ask a PSU expert how much of a harm this 10.8V+ for 12ms will make, as I'm not an expert on that.
  • Aris_Mp
    hold-up time tests are not the easiest thing to do since besides extra equipment you also need GOOD knowledge of what you are doing. On top of that in order to measure hold-up time along with power good time and other stuff as well you need an accurate (thus expensive) four channel, at least, scope. Till I raised enough money to buy one I didn't have the capability to measure both hold-up times so I provided only the power good signal's time.

    If your PSU has a really short hold-up time in some cases your UPS won't be able to handle the job, especially if its transfer time is higher than 15-20ms. Moreover the longer the PSU goes without power, the larger the in-rush current it will draw when it has power again. This means that its bulk caps will be almost empty so the In-rush current could exceed the current handling capacity of the UPS and cause its shut down. This means that a PSU's hold-up time IS important!
  • turkey3_scratch
    Aris, you say the in-rush current could exceed the current handling of the UPS, don't you mean the PSU? And are you the Aris who wrote the article? Also, if the in-rush current exceeds the current handling capacitor of the bulk caps and the unit shuts down, will that cause hardware damage at all?