EVGA SuperNOVA 1600 P2 Power Supply Review

A Look Inside And Component Analysis

Our main tools for disassembling PSUs are a Thermaltronics soldering and rework station and a Hakko 808 desoldering gun.

EVGA 1600 P2 Parts Description
Primary Side
Transient Filter
6x Y caps, 2x X caps, 2x MC chokes, 1x MOV, NTC Thermistor
Inrush Protection
NTC Thermistor & Electromagnetic Relay
Bridge Rectifier(s)
Bridgeless Design - 1x US30K80R & 8x Infineon MOSFETs
APFC MOSFETs
8x Infineon MOSFETs
APFC Boost Diode
4x CREE C3D08060G
Hold-up Cap(s)
4x Nippon Chemi-Con
(400V, 390uF each, 1560uF combined, 105 degrees C, KMW)
Main Switchers
4x 50R140CP
APFC Controller
SF29603
Switching Controller
SFAA9013
Topology
Bridgeless PFC & Full-Bridge LLC & Resonant Converter
Secondary Side
+12V MOSFETs
16x Infineon BSC035N04LS
5V & 3.3V
DC-DC Converters: 8x Infineon IPD060N03 FETs
Filtering Capacitors
Electrolytics: Chemi-Con, 105 °C, KY & KRG
Polymers: Chemi-con
Supervisor IC
AA9013 (probably) & LM324ADG
Fan Model
Globe Fan RL4Z-B1402512EH
(140mm, 12V, 0.6A, 2000 RPM, 153.47 CFM, 39.5 dB(A), 70,000-hour MTBF)
Topology
Syncronous Rectification & DC-DC converters
5VSB Circuit
Rectifying Diode
Mosped S10C60C
Standby PWM Controller
29604


The original manufacturer of this PSU is Super Flower, which is behind most of EVGA's power supply offerings. The platform is the same as the one used in the Gold-rated EVGA 1600 G2, but sporting the modifications needed to pass the 80 PLUS Platinum requirements. This is a cutting-edge design with a bridgeless APFC converter in the primary side, along with a full bridge topology and a resonant converter for lossless switching. In the secondary side, synchronous rectification is used, meaning that the regulation of the +12V rail is handled exclusively by MOSFETs and a pair of DC-DC converters generate the minor rails.

The small PCB behind the AC receptacle, strangely enough, doesn't host any part of the transient filtering stage. That's installed on the main PCB, and includes three pairs of Y caps, a pair of X caps, an MOV and two CM chokes. The transient filter is complete.

There is only a single bridge rectifier, a US30K80R, which most likely is used only by the 5VSB circuit, since the main rectification takes place in the APFC converter.

In the APFC converter, eight MOSFETs, which are much more energy efficient compared to diodes, fully rectify the incoming AC voltage. The aforementioned MOSFETs, which we were unable to identify, are hosted on two vertical daughter boards, along with four C3D08060G boost diodes. As bulk caps, four parallel Nippon Chemi-Cons (400V, 390uF each, or 1560uF combined, 105 °C, KMW) are used, and our tests demonstrate whether their combined capacity is enough to allow for at least 16ms of hold-up time.

A proprietary IC code-named SF29603 is used as a PFC controller. It is installed on the shielded PCB shown in the photo above.

There is a big NTC thermistor for protection against large inrush currents during the supply's start-up phase. A relay cuts it from the circuit once it finishes its job, allowing it to cool down and at the same time to restrict energy losses. This works because, in essence, an NTC thermistor is a temperature-controlled resistor.

Two boards hold the main switchers, four Infineon IPP50R140CP, which are arranged into a full-bridge topology. An LLC converter is also used to boost efficiency through lossless (or at least close to lossless) switching.

The main transformer should be huge to meet this unit's needs. Apparently there wasn't enough space for it because Super Flower had to use two smaller, parallel transformers. Also in the secondary side, there are many high-quality Nippon Chemi-Con electrolytic (105 degrees C, KY series) and polymer caps, which are used for filtering purposes.

Four vertical PCBs house all MOSFETs that regulate the single +12V rail. In total, 16 Infineon BSC035N04LS FETs are used, as in the 1600 G2 unit. Two DC-DC converters handle the minor rails, and each converter uses four Infineon IPD060N03 FETs.

Next to one of the DC-DC converters is a PCB with the fan-controlling circuit and a Mospec S10C60C SBR that generates the 5VSB rail. The above PCB also houses an LM324ADG. Since this PCB is held in place only by the solder joints on its base, you can break it easily while trying to remove the fan header. So, we applied lots of glue around the base to secure it after learning our lesson the hard way on another high-end Super Flower PSU.

The standby PWM controller only has the number 29604 written on it, which doesn't prove useful in identifying this component.

On a vertical daughter board is the resonant controller, an AA9013 IC, for which there is no available information. There is also a well-hidden LM324ADG on the same board.

The many polymer and electrolytic caps on the modular PCB's primary side provide further ripple filtering. These caps ensure that the output DC voltages are clear of AC fluctuations (ripple).

Soldering quality is good. It may lag behind Flextronics and Delta, but it is close enough. Super Flower has made huge steps forward in this area over the last few years.

The cooling fan is provided by Globe Fan, and is the same one found in the 1600 G2 and 1200 P2. Its model number is RL4Z-B1402512EH (140mm, 12V, 0.6A, 2000RPM, 153.47 CFM, 39.5 dBA, 70,000-hour MTBF), and it is equipped with dual ball-bearings for increased longevity. This unit, in contrast to the 1600 G2, features a semi-passive mode that makes it much quieter under light loads.

This thread is closed for comments
25 comments
    Your comment
  • damric
    Good review, but missing the hot box testing to see if this thing regulates and suppresses ripple at 50C as advertised.
  • Aris_Mp
    All tests were conducted at high ambient temperatures which during full load were above 47C. Only the Cross-Load tests were conducted at 28-30C.
  • SinxarKnights
    I appreciate the detailed review.
  • Giannis Karagiannis
    Very detailed review indeed. There isn't really anything that could be covered and it is not. I don't think that there are many PSU manufacturers out there that can test their products so extensively.
  • Dark Lord of Tech
    Too light for me I have the 2000w coming from Dabs when it comes to retail.
  • Aris_Mp
    I had the opportunity to test the 2 kW model (from Super Flower) and it is indeed superb. But it will provide 2 kW only with 230 VAC input since a normal socket can deliver only up to 15 A of current.
  • damric
    276663 said:
    Too light for me I have the 2000w coming from Dabs when it comes to retail.


    Where are you from that you need all that power? Cybertron?
  • damric
    1903369 said:
    All tests were conducted at high ambient temperatures which during full load were above 47C. Only the Cross-Load tests were conducted at 28-30C.


    47C ambients? Must have been sweating your language, please off, or you are language, please me.
  • Dark Lord of Tech
    I'm going to power my Skynet build with it.
  • damric
    276663 said:
    I'm going to power my Skynet build with it.


    One day you need to show us a picture of everything. I've seen little snapshots here and there, but I'd like to see it all in one thread.
  • Dark Lord of Tech
    I'm starting a white and black build , a Snow Beast , next week.
  • damric
    With the Krait motherboard? Or is there another white/black?
  • Dark Lord of Tech
    No way this one.

    http://www.newegg.com/Product/Product.aspx?Item=N82E16813132414
  • damric
    Oh wow I haven't seen that yet.
  • Aris_Mp
    410076 said:
    1903369 said:
    All tests were conducted at high ambient temperatures which during full load were above 47C. Only the Cross-Load tests were conducted at 28-30C.
    47C ambients? Must have been sweating your language, please off, or you are language, pleaseing me.


    The PSU is inside a special-made box (hot-box) for the high temp tests.
  • Dark Lord of Tech
    Sweet board going to order it on monday.
  • damric
    1903369 said:
    410076 said:
    1903369 said:
    All tests were conducted at high ambient temperatures which during full load were above 47C. Only the Cross-Load tests were conducted at 28-30C.
    47C ambients? Must have been sweating your language, pleaseoff, or you are language, pleaseing me.
    The PSU is inside a special-made box (hot-box) for the high temp tests.


    Wouldn't your FLIR show that the box is hot?
  • Aris_Mp
    no point since the box is insulated so from the outside the temperature will be lower. Also if i open the lid to take a snapshot with FLIR the temperature will drop immediately by 5 C at least (ambient) messing with the test.

    I already know what happens inside the box thanks to two temperature probes I have installed in it, so no need to use my FLIR on it. However on next review I will try it.
  • damric
    1903369 said:
    no point since the box is insulated so from the outside the temperature will be lower. Also if i open the lid to take a snapshot with FLIR the temperature will drop immediately by 5 C at least (ambient) messing with the test. I already know what happens inside the box thanks to two temperature probes I have installed in it, so no need to use my FLIR on it. However on next review I will try it.


    Fair enough. It will be excellent when Tom's Hardware adopts a consistent review standard. Yours was one of the best PSU reviews yet. There's other nitpicks I have but they are nothing major like chart/graph formats could be easier to read.

    In your opinion everything held up in close to 50C conditions then?
  • loki1944
    Nice, if I ever go 3-4 way SLI I'll upgrade to this from my 1300W G2.
  • ykki
    Quote:
    I'm going to power my Skynet build with it.

    Dont forget the OFF switch. Arnold is a busy man.
  • Aris_Mp
    "In your opinion everything help up in close to 50C conditions then?"

    around 45C is much more realistic. 50C are too much for modern chassis. However when needed I crank up the heat inside my hot box to see how the PSU performs. Like in this case that I deliberately exceeded my usual 45C and went to almost 48.
  • g00ey
    The "Power Specifications" table on the second page is wrong. The third line must be "Maximum Watts" and not "Maximum Volts". Moreover, the Wattage calculation for the 3.3V rail is missing (should be 3.3*24 given that the maximum amperage for the 3.3V rail is 24A, this must be verified). The value 120 (W or VA) must be for the 5V rail since 5V*24A=120 VA/W...

    Also note that I'm making a distinction between Watts (W) and Volt-Amperes (VA) although they have the same dimension. The reason is this:

    http://electronicdesign.com/energy/what-s-difference-between-watts-and-volt-amperes

    I hope you do the same.
  • Aris_Mp
    275838 said:
    The "Power Specifications" table on the second page is wrong. The third line must be "Maximum Watts" and not "Maximum Volts". Moreover, the Wattage calculation for the 3.3V rail is missing (should be 3.3*24 given that the maximum amperage for the 3.3V rail is 24A, this must be verified). The value 120 (W or VA) must be for the 5V rail since 5V*24A=120 VA/W... Also note that I'm making a distinction between Watts (W) and Volt-Amperes (VA) although they have the same dimension. The reason is this: http://electronicdesign.com/energy/what-s-difference-between-watts-and-volt-amperes I hope you do the same.


    The max power is the combined max power that both rails can deliver. Hence while each rail can go up to 24 A both of them can deliver only up to 120 W (combined). This means that either the 5V rail can go up to 120 W alone (so zero W for the 3.3V one) or the 3.3V rail can go up to 24 A and the 5V at 8.16 A.

    In DC we use Watts. VA is for AC. In DC Watts = VA

    As for the max voltage yeap this will be fixed.