7-way 600 Watt PSU roundup in Venlo

Cases & PSU/Power Supplies by geoffrey @ 2010-03-17

Looking for a 600~700W PC power supply but haven´t decided yet what to get? In addition to our 2009 summer article we at Madshrimps have tested another four popular products from major brands like Nexus, Enermax, Scythe and Sharkoon. We compared features, performance and efficiency. Which one to get? Read on to find out!

  • prev
  • next

Scythe Kamariki 4 650W pics & results

Scythe Kamariki 4 650W box unpacked

Madshrimps (c)
Image courtesy of Scythe.com

Box unpacked, inside we found:
  • Scythe Kamariki 4 650W power supply
  • 3 x 3-pin FAN power cable
  • AC power cord
  • Manual
  • 4 x PSU Installation Screws

    A very basic box contents which aims directly at the lower price market. As extra feature we found 3 fan connectors at the PSU side, with this PSU you'll also get cables to hook up a maximum of 3 fans. As you can see from the picture below each cable can connect with 1 fan, one side of the cable has to be connected to the PSU to take its power and the third connector on the cable is used for making the connection with the PC mainboard for reading out the fan speed. I do wonder why one would actually use this feature as we can take power from the mainboard anyway. Thing about these PSU fan connectors is that it changes fan speed dynamically according to the load applied to the power supply, so as soon as you go full load the fan speeds up to maximum speed, quite noisy, I'd rather see fan speed control based on temperature levels, hence the reason why one would rather use the mainboard fan connectors instead of these, unless off course you've ran out of fan connectors but then again you're probable better of with a fan baybus.

    Madshrimps (c)

    Cables explored

    Madshrimps (c)

    In contrary to the previous PSU's we tested in 2009 this Kamariki 4 is not of the modular type. Though if you really want Scythe has released a second K4 power supply which is of the modular type and comes with nearly the same features as this model. Here is a small cable line-up:
  • ATX 20+4pin Cable
  • 2 x PCI-E 6+2pin Cable
  • 1 x 4+4pin 12V EPS Cable
  • 2 x ODD/HDD/FDD Cables
  • 2 x S-ATA Cables

    The Kamariki 4 power supply comes with 6 molex, 8 SATA and 2 floppy connectors plus 12V EPS and 2 times 6+2 pins PCIE connectors, what more could you want? The ATX plug is 20+4 pins which is most preferable too when taking compatibility in mind, cable length is also longer than average which makes the Kamariki 4 one of the better PSU's when looking at connectivity only.

    More pictures

    Madshrimps (c)
    Madshrimps (c)

    Scythe gave their K4 PSU a kind of reflecting metal surface, in my honest opinion not that beautiful plus it looks dirty once you've had it in your sweaty fingers. Size is normal, same as most other 600W PSU's, though the fact that this unit is not modular might make it less popular, it's always nice when you have the option to use as less cables as required. On the back you'll find large ventilation holes optimized for maximum airflow, inside we spotted the Scythe original Kaze Jyu fan which comes with a smaller fan 'core' compared to most other fans so to let more air travel through at the same fan speed. Gains are negligible but if you have it then why not just use it.

    Madshrimps (c)

    Scythe hasn't overdone in internal cooling heatsinks. In close up the Kamariki 4 looks exactly the same as the OCZ power supply which we tested earlier. To be correct, both power supply's are build by the same manufacturer, the printed circuit board matches except for a few details. The OCZ sample has revision 8 pcb where Scythe has the rev10 pcb, that is very probable just because of the difference in time we got both power supply samples. This also explains why both PSU's have matching specifications, though Scythe warrants up to 650W where OCZ thinks 600W is more then enough.

    Madshrimps (c)

    Scythe Kamariki 4 650W results

    How to read the table underneath? Well, we tried 3 different load settings: 100%, 50% and 20% of the total power rating. In the left green field on the right side of the table you can read the output power of the PSU, i.e. the load we set, it's calculated by measuring the live voltage per rail, and by measuring the flowing current per volt rail. You can also see what numbers we've measured inside the CM labs, and how many power is consumed by multiplying U (Volts) with I (Ampéres). In the red field on the left side represents the input power of the PSU, i.e. how many in total is being drawn from the wall plug. The second green value on the right side is the efficiency of the power supply, or output power divided by input power, in percent that is.

    Madshrimps (c)

    During our tests we had some problems with properly loading the power supply. The power supply has 2 12V rails, one for 12V EPS and one for all the rest. But that we did not know, it's nowhere explained in the manual/box/website so we divided current equally to not overload the power supply. Off course, with one rail build for 12V EPS only the other 12V rail was quickly being overloaded with roughly 2 Amps when we tried 100% load resulting in a power supply lockdown in only a matter of time, very probable due to overheating. Efficiency was still good but the PSU ran very hot. Looking back at the results we obtained with the OCZ unit you'll see that we noticed very little headroom left once we topped out the specs and considering the heatsinks really aren't that big maybe the 650W rating is a bit over optimistic?

    We were more successful when we lowered combined 12V1 current to 3 times 8 Amps which resulted in 90% load, now the PSU kept running and efficiency was good:

    Madshrimps (c)

    Some more test results we would like to share:

  • Inrush current: 59,5 Ampére
  • Power Factor: 0,91~0,99
  • Standby power consumption: 1,16 Watt
  • 12V shortcut protection: OK
  • 5V shortcut protection: NOK

    After the problems that occurred with the power supply locking down due to overheating we also discovered that the 5V overload protection was not working properly. When pressing the short circuit button on the Chroma test unit the power supply did not power off itself instead it kept keep on running for few more seconds until we powered it off ourselves. The Chroma unit showed us 60Amps at the 5V rail which is the maximum this unit can provide. We repeated the test with the same outcome following up and since we did not encounter these problems with the OCZ unit I guess this might have been a problem with the reviewing sample.
    • prev
    • next
    Comment from EsaT @ 2010/03/19
    Overall good job but as you "crack" them open in any case you should also either check and write about capacitor selections or take such photos which allow recognizing capacitors.
    For most global brand PSUs used capacitors can be found from numerous other reviews but for European PSU brands thorough reviews aren't so common.

    While looking all good now PSU with crap capacitors can easily go bad after only couple years (or faster) so use of known high quality capacitors isn't just PR stunt. Instead of el Cheapo products even PSU with more average performance, say in voltage regulation, but high quality capacitors, is lot better investment and can keep going lot longer.
    Comment from geoffrey @ 2010/03/23
    You're correct and I must confess that I don't have enough knowledge about quality capacitors and such stuff. I'd love to tell you about what components to look out for and which you would rather want to avoid. In order to do that one must have lots of experience in electronics, for myself I would pay a visit to some of these capacitors producers and see how they do it. I would ask repair departments in order to know their thoughts about certain weak components instead of reading and quoting other man's words. But frankly, knowing that it took a very long time to get this review done, and knowing that writing reviews is just one of the many hobby's I have, I can tell you that getting more in detail would take only more time which I can't make available at the moment.

    "Then why bother the review?" Well why would you? I'm telling you from the start that internals is not what we're going to focus on, instead we'll be testing the product 'as it comes' and judge by that those facts what it is worth for us. On the longer scale there're lots of influences which can degrade the product life, and yes quality components will most certainly extend the joy you'll have from it, but than again you're judging by name and not by real world facts like you would get when testing hundreds of units for multiple years. I say it's a nice extra thing to mention but sometimes I get the thought that some reviewers are just quoting stuff and that's why I rather not spend time researching and reading other man's work. But I hope that what you got in this article was good enough to get you going
    Comment from EsaT @ 2010/03/24
    Sometimes some smaller caps have only minimal markings on them but most have manufacturer's name or logo so simply listing identifiable capacitors and leaving arguing/deductions to readers wouldn't increase workload more than minute or two per PSU.
    Cap makers often sticking to certain colour theme can give fast first hint of manufacturer but still gettings markings to show in photos would actually take more effort.

    judging by name and not by real world facts like you would get when testing hundreds of units for multiple years
    Even longer testing of big sample isn't reliable if quality is inconsistent. Like those Fuhjyyus making Antec SmartPowers and old TruePower I/IIs random timebombs. Some of them have obviously worked well while lot of them have been ready to pop at any time.
    And even if lower quality caps work long in one product lowering cooling (as silence is now major advertising point in PSUs) below some level can kill those caps fast in other product.

    So there simply aren't any valid reasons to keep cheap dozenware equal to known long time manufacturers.
    It's rather sure bet that quality PSU manufacturers do some tough testing to capacitors (also they have to stay sharp because of huge counterfeit production in China) they keep using in their high end models so that should be good hint as to what is reliable because in capitalistic world use of more expensive parts than necessary is rare.

    Personally I have one 13 years old high end Nokia CRT still working but despite of good looking design and good sized heatsinks for critical components Samsung 959NF started going haywire after 6½ years simply because of lot of capacitors failing. (don't remember their brands except that they weren't known high quality brands)
    And now in TFT monitors cheap capacitors of their power supplies are common cause of their premature deaths before reasonable operating life.
    So with 1½ year of use behind I have to start really considering when I open my LG IPS monitor for checking is it better to replace caps for making sure it lasts because real step forward OLED and FED/SED monitors are still years away and current monitor markets are falling towards vertically cripled cheap crap. (Millions of flies can't be wrong: Crap is good.)