• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Supermicro-PWS056 650W PSU Review

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

Super Nade

† SU(3) Moderator  †
Joined
Aug 30, 2004
Location
Santa Barbara, CA
WORK IN PROGRESS!

Updated with JonnyGURU's ATE-Results (Many thanks mate!).
See HERE and HERE



Hi all,

Upon Galvanized's (a member at Anandtech and Badcaps.net) recommendation and reading THIS thread, I decided to go for the Supermicro PWS-056 unit.

It was purchased at wiredzone.com FOR $136 SHIPPED, an authorized reseller listed on Supermicro's website. I had a lot of difficulty finding a vendor as this unit was out of stock at most resellers. I will recieve my unit (hopefully) by the end of the week. The unit will be shipped directly from Supermicro. What attracted me to this unit was the beefy 12V line. This PSU was made by Liteon and the model number is something like PS-5651-1A1 (not exact, but a brother). I have requested the official datasheet from Supermicro. Since this is usually sold as an accessory, you can't really find one in the general market.


Lets look at the manufacturer specs:


[/size] Electrical Characteristics :
  • Input Voltage 100V AC to 240V AC
  • Output Power 650 W
  • Frequency 47 Hz to 63 Hz
  • Current 11A @ Input Current
  • Current 30.0A @ 5V DC - Output Current
  • Current 4.0A @ 5V DC - StandbyOutput Current
  • Current 46.0A @ 12V DC - Output Current
  • Current 0.6A @ -12V DC - Output Current
  • Current 30.0A @ 3.3V DC - Output Current
Interfaces/Ports :
  • Connectors 24-pin Motherboard Adapter
  • Connectors 8-pin ATX Power 12V Plug
  • Connectors 4-pin ATX Power 12V Plug
*Regulation is unknown.


Some Terms to be aware of (Borrowed from here) :
  • OPERATING RANGE:
    The minimum and maximum input voltage limits within which a power supply will operate to specifications. A power supply with a wide input range is recommended when the line voltage is subject to brownouts and surges.
  • EFFICIENCY:
    The ratio of output power to input power expressed as a percentage.
  • EMI:
    Electromagnetic interference is the noise generated by the switching action of the power supply. Conducted EMI, that portion reflected back into the power line, is normally controlled with a line filter. Radiated EMI, that portion which is radiated into free space, is suppressed by enclosing the circuitry in a metal case. The FCC governs conducted and radiated emission levels.
  • PFC:
    Power factor is the ratio of true power (watts) divided by apparent power (volts x amps or VA). A standard power supply has a power factor of 0.70-0.75, while a power supply with active power factor correction (PFC) has a power factor of 0.95-0.99. A power supply with power factor correction is better able to convert the current into power. This results in lower peak current and lower harmonic current, putting less stress on wiring, circuit breakers, and transformers.
  • OUTPUT CURRENT:
    The maximum current which can be continuously drawn from the output of a power supply. PC motherboards and expansion cards draw 5 volt current. Drive motors draw 12 volt current.
  • LOAD REGULATION:
    The change in output voltage due to the output load varying from minimum to maximum with all other factors held constant. It is expressed as a percent of the nominal output voltage. A power supply with tight load regulation delivers optimum voltages regardless of system configuration.
  • LINE REGULATION:
    The change in output voltage due to variation of the input voltage with all other factors held constant. It is expressed as a percent of the nominal output voltage. A power supply with tight line regulation delivers optimum voltages throughout the operating range.
  • TRANSIENT RESPONSE:
    The time required for the output voltage to return within the regulation envelope following a 50% load change. A power supply with quick transient response will reduce the risk of read/write errors during access.
  • RIPPLE:
    The magnitude of AC voltage appearing superimposed on the DC output, specified in peak to peak volts or expressed as a percent of the nominal output voltage. A power supply with clean DC output is essential for computers with high-speed processors and memory chips.
  • HOLD-UP TIME:
    The time period, following a loss of input power, that a power supply's output will remain within specified limits. Adequate hold time keeps the computer running during the transfer time required by a UPS unit.
  • POWER GOOD SIGNAL:
    A delay circuit used to initialize the computer and provide a logic signal upon low line voltage.
  • OVERVOLTAGE PROTECTION:
    A circuit that shuts down the power supply if the output voltage exceeds a specified limit.
  • OVERCURRENT PROTECTION:
    A circuit that protects the power supply and computer from excessive current, including short-circuit current.
  • AGENCY APPROVAL:
    UL, CSA, and TUV are safety agencies that test specifications such as component spacing, HI-pot insolation, leakage currents, circuit board flammability, and temperature rating.
  • OPERATING TEMPERATURE:
    The range of ambient temperatures within which a power supply can be safely operated.
    FAN RATING:
    Airflow in cubic feet per minute. A 100% increase in airflow will reduce system operating temperatures by 50% relative to ambient. For each 10°C (18°F) reduction, the life of the system is doubled. (Arrhenius equation).
  • NOISE:
    Acoustical noise in dB(A) at 1 meter. Logarithmic scale. Each 3dB reduction represents 50% less noise. Issues include the pitch and speed of the fan blades, the hub size, the venturi depth, the bearing quality, and the layout of the power supply components.
  • MTBF/MTTF :
    Mean Time Between/To Failure. A measurement of the relative reliability of a power supply based upon actual operating data or calculated according to MIL-HDBK-217.(Ed.) These are statistical measures.
    MTBF is a basic measure of reliability for repairable items. It can be described as the number of hours that pass before a component, assembly, or system fails. It is a commonly-used variable in reliability and maintainability analyses.
MTBF can be calculated as the inverse of the failure rate for constant failure rate systems. For example: If a component has a failure rate of 2 failures per million hours, the MTBF would be the inverse of that failure rate.

MTBF = (1,000,000 hours) / (2 failures) = 500,000 hours
MTTF is a basic measure of reliability for non-repairable systems. It is the mean time expected until the first failure of a piece of equipment. MTTF is a statistical value and is meant to be the mean over a long period of time and large number of units. For constant failure rate systems, MTTF is the inverse of the failure rate. If failure rate is in failures/million hours,
MTTF = 1,000,000 / Failure Rate for components with exponential distributions.
Technically MTBF should be used only in reference to repairable items, while MTTF should be used for non-repairable items. However, MTBF is commonly used for both repairable and non-repairable items. (source)



Testbed and Methodology :
This PSU will be tested in two ways. First, a professional PSU tester would be used to load the unit and perform efficieny, power factor and cross-loading tests. These tests will be performed by JonnyGURU, who's help I sought on badcaps.net. He's been testing PSU's for a while, both as an enthusiast and as a professional, so these tests should be fun. More information regarding his methods can be found HERE.

Now onto more real-world tests, I'll be using the following system, loaned out to me by corpsjockey, one of our members here.
  • DFI LP NF4 SLI-DR
  • x1800xl-->XT 512 Mb
  • As many HDD's and Optical drives I can lay my hands on.
  • Opteron 144
  • Mushkin lvl II Black PC3500 BH5
For a more severe test, I would need somebody with a dual-core CPU/Prescott and a CF setup to help me out.

I will run the following standard benchmarks. Please let me know if I can add anything to this.
  • 3DM 2001
  • 3DM 2005
  • 3DM 2006
  • Stress CPU
  • Prime95
  • Super Pi
  • CPU Burn
  • 3DS Max from Autodesk
Well, what am I actually doing here?
  • Stick multimeter probes into a molex and get average DC voltage numbers on all the lines simultaneously.
  • Stick probes on selected spots on the mainboard and get another set of readings
  • Time permitting, I have at my disposal a GPIB interface setup to acquire data from an Oscilloscope. So, I'll stick in a 10x probe into a molex (and use a divider network) and acquire data for about an hour.
  • I also have a very precise Temperature controller/readout unit, so I'll stick in a few thermistors and maybe we can discern how hot this unit gets under load.
As always, I'm open to new ideas. Please let me know if you have any specific tests you would want me to run.

:)

Update: July 26 2006

The JonnyGURU Test Results :

It's been a while since this was updated. JonnyGURU has finished testing the unit with his Automated Test Equipment setup. For more information on his methods and setup, please consult HIS website..

The outside of the unit:

What can I say here? Looks very military, no embellishments whatsoever. Who would expect bling from a server unit anyway?


img0656ud8.jpg



The Power ratings:
Just as the info sheet said. The ratings look pretty solid on paper.

img0660yo8.jpg



Cables:
No sleeving or any accesories. Typical server unit.
img0662ln4.jpg



Internals:
Now we come to the fun part! The primary side caps are made by United-Chemicon, a reputed manufacturer of low ESR devices. The goo in there is to damp physical vibrations. Physical vibrations (manifesting as a hum) occurs when harmonics of the line frequency are produced. Mathematically, any waveform can be decomposed into sine and cosine waves (the basis of Fourier analysis) and since the audible range extends from 20Hz-20kHz, harmonics, if produced can be heard. You can see that the inductor is not wound properly. Usually, one avoids crossing windings. Looks to be a solid build. The green caps are Taicon.


img0664ra7.jpg



img0665nh2.jpg



img0666lo3.jpg



img0671yh0.jpg



img0673ag3.jpg



img0674cp7.jpg






















 
Last edited:
That's a hell of a lot of power. Only one 12v rail though? Hard to wrap my head around a single rail PSU. Haven't had one in a while.
 
Good choice on that PSU - OEM Lite-On. I look forward to seeing how it does.
 
Generous of corpsejockey to loan you that system, but IMHO it won't come close to putting the PSU to the real test. A single core with a single card won't prove much and hard drives only really draw a significant power at spinup.

If you can, see if you can load the +12v rail with power resistors like proth and Ross did here.

Found this site which seems to have a pretty nice selection of them and they look pretty cheap too.
 
The artificial loading tests would be taken care of by JonnyGURU. He basically has a black box which does the same thing as the resistor pack. After I'm done playing with it I'll find someone with a CF setup to help me finish this up. I was in the process of adding this to the first post, but the Mexico-Portugal game had already started. :D
 
Last edited:
Pictures (please excuse the poor quality)

img01364dm.jp
Initial impressions:

Well, this is a server PSU after all, so I'm not surprised by the lack of bling! However, the heatsink looks really flimsy, just a couple of pieces of Aluminum each. The redundant fan feature is nice and there is an alarm which goes off if a fan fails. Caps are Kemicon (Although the pic isn't clear, I can see a KMM on the big storage Caps). Other smaller caps are Taicon. It isn't really clear because there is some kind of a goo covering the names. It's hard to believe that this is a 650W unit because it is so light. Let us see how this holds up.
.
.
.
.
img01374qt.jpg

.
.
.
.
img01524ky.jpg

.
.
.
.
img01510zz.jpg



More images to follow when I get a better digital camera. G0dm@n yhpm :D
 
Last edited:
It looks like jonnyGURU upgraded his test bench to be able to test 10 rails at a time :D
Have you sent him the PSU yet?
(He hasn't mentioned it on his site yet)
 
I sent it over last week and he recieved it on Friday. He hasn't gotten back to me yet, but the USPS tracking number say's "Delivered". I've sent him a pm on badcaps.net, waiting for him to reply.

Maybe he, like me, is watching the World Cup Final. :)
 
I think he's swamped with stuff to review - there's the OCZ 700W and Silverstone 750W for sure he's working on, plus the Fortron FSP600-80GLC I just sent him (though he wouldn't have gotten that one yet). And then Silverstone is sending him the new 850W too.

Plus he has that new toy to figure out yet ;)
 
ATE Results:


COLD TEST -->3.3V, 5V,12V, WATTS, EFF %, PF, INTAKE, EXHAUST
Test One Load 2A 5A 8A
Test One Result 3.35V 5.04V 11.97V 144W 73% 0.99 28°C 31°C

Test Two Load 5A 7A 14A
Test Two Result 3.32V 5.01V 11.95V 275.7W 77% 0.98 28°C 32°C

Test Three Load 7A 12A 22A
Test Three Result 3.29V 4.98V 11.91V 398.8W 77% 0.99 29°C 34°C

Test Four Load 10A 17A 30A [A buzz started to come from the PSU. Also, the outside fan started to spin at full RPM. VERY LOUD! ]
Test Four Result 3.26V 4.95V 11.85V 520W 75% 0.99 29°C 37°C

Test Five Load 12A 20A 40A
Test Five Result 3.23V 4.93V 11.79V 638.4W 74% 0.99 29°C 39°C

Crossload 3A 4A 40A
Crossload Result 3.29V 5.06V 11.57V
HOT TEST--> 3.3V, 5V, 12V, WATTS, EFF, % PF, INTAKE, EXHAUST
Test One Load 2A 5A 8A
Test One Result 3.35V 5.04V 11.97V 143.1W 73% 0.99 30°C 32°C

Test Two Load 5A 7A 14A
Test Two Result 3.29V 4.98V 11.95V 273.1W 76% 0.98 32°C 35°C

Test Three Load 7A 12A 22A
Test Three Result 3.24V 4.93V 11.91V 393.6W 76% 0.99 33°C 37°C

Test Four Load 10A 17A 30A A buzz started to come from the PSU. Also, the outside fan started to spin at full RPM. VERY LOUD!
Test Four Result 3.18V 4.86V 11.85V 509.7W 74% 0.99 36°C 39°C

Test Five Load 12A 20A 40A
Test Five Result 3.13V 4.81V 11.79V 621.7W 71% 0.99 37°C 40°C
Oscilloscope waveshapes:

Test One 3.3V


test133vfm1.jpg



Test One 5.0V

test15vhg5.jpg
]


Test One 12.0V

test112vqr2.jpg



Test Two 3.3V

test233vxz6.jpg



Test Two 5.0V

test25vql9.jpg



Test Two 12.0 V

test212voe0.jpg




Test Three 3.3V


test333vzx9.jpg



Test Three 5.0V

test35vrq7.jpg



Test Three 12.0V

test312vjf0.jpg
 
Oscilloscope Waveforms Contd...

Test Four 3.3V:

test433vxo8.jpg


Test Four 5.0V

test45vfp0.jpg



Test Four 12.0V

test412vie9.jpg



Test Five 3.3V

test533vuh3.jpg



Test Five 5.0V

test55vri2.jpg



Test Five 12.0V

test512vhn0.jpg


 
Analysis of Test Results:

Please visit www.jonnyguru.com, to understand his testing procedure!
  • First off, it is easy to see that this unit has poor power efficiency. I expected something on the lines of 85%.
  • In both the Cold and Hot tests (please refer to JonnyGURU's site for a detailed description), the 12V line has held very well. The maximum deviation was just about 1.5% (under the hot test ), which is well below the 5% limit. This is remarkable considering the fact that supplies rarely maintain their rated efficiency or regulation at high temperatures.
  • JG said the fan made a loud noise when the temperature increased. Note that this unit has a redundant backup fan if the primary exhaust failed. I don't see any advantage in this. They could have used a 120mm or a dual fan configuration instead.
  • In the hot/full load tests, JG reported a buzzing sound from the inside of the unit. I would guess that this is due to a small valued secondary cap which is not filtering out higher harmonics. Also, we discussed the possibility of the toroidal inductors acting up. If you look carefully, the windings are horrible, with plenty of criss-crossing. Usually, in an SMPS device, there is always going to be a magnetic hysterisis effect, which can result in an audible hum, as energy is dissipated. This can be exarcerbated by a poor quality Inductor.
  • The cross-load test result is really bad. I would chalk this up as a near failure (11.45 V is the failure mark). What does this mean? Since all rails interact ith each other for some measure, unbalancing load on one rail would affect the others. In our case, lightly loaded 5.0 V and 3.3V rails and a heavily loaded 12V rail yeilds a near failing result. An easy way to circumvent this issue would be to have about 4 HDD's and an Optical drive, plus lots of fans.
  • JG mentioned that the 12V line drops by about 0.1V and immediately recovers "2ms sharp drop". This could again be chalked up to undersized caps. I was also thinking in terms of Ground Bounce due to small ground loops or if it's step like, it may have something to do with the fly-back of the voltage regulator. I mean, during a switching cycle, the Volt x time product is not being maintained momentarily. If its a rapid exponential, maybe the secondary caps are too small in value?
Still to come...
-Scope data analysis.
-Comparing this with other units JG has reviewed.
-Real World tests with fully overclocked P4 and Opteron rigs in Part II of this review. (Need your help in obtaining CF x1800xt 512Mb or 7800GT for SLI).
 
Last edited:
Based on what I'm seeing so far, I have no problem keeping Lite-On in my top 5 list. The crossload results aren't too awful for something lacking independant voltage regulation (I don't see the circuitry on the secondary to support it), and the ripple stays pretty much under control.

So, my top 5 in order of build quality is:
1-Zippy/Emacs
2-Etasis (Silverstone Zeus 560, 750, 850)
3-Win-Tact (PC P&C Turbocool)
4-LiteOn (Supermicro/Ablecom)
5-Delta

Edit - for my fellow Canadians, this unit is dirt cheap at the moment:

http://search.ncix.com/displayprodu...2&vpn=PWS-0056&manufacture=SUPERMICRO - CASES
 
Last edited:
Back