Rosewill RD600N-2SB-SL-BK review @ hardware secrets

[370forlife]

http://www.hardwaresecrets.com/article/558/1

Ok, so it's a review of another rosewill unit, quite a old review actually. Whats so special?

Solytech

We were surprised with this power supply. With so many people saying bad things about the many previous incarnations of Solytech we were expecting a power supply with low efficiency and that couldn't deliver its labeled power. This unit could deliver its labeled 600 W at 47.3º C, which is wonderful, plus we could pull up to 675 W from it, and it survived! But this isn't a “perfect” power supply. Efficiency drops below 80% when pulling its full 600 W power and also noise level is far higher than we wanted to see.

[i4004]

Re: Rosewill RD600N-2SB-SL-BK review @ hardware secrets

both hardwaresecrets and jonnyguru don't seem to pay any attention whatosoever to most important circuit in the psu: the +5vstby.

you don't know if psu is good or bad untill you know its +5bstby circuit, because it has the potential to destroy both psu and mobo if it's done wrong.

how much power can ti deliver is less important than that, because most of the time it'll be pulling what, 200-300w....
i would much prefer 400w branded that can deliver only 350 with decent 5vstby, than this resewill that can produce more than it says, if it has poor 5vstby design.

and also, short review is not same as putting the psu in machines and then seeing the number of times psu fails. i don't believe in accelerated tests, or in short-sharp-shock techniques....
these tests tell us nothing aboout long term reliability ie durability.

rosewill was famous for being crappy(and newegg feedback proved that) but now, dunno?
buy one(aren't they cheap?), dismantle, photograph it, and then we'll know.

[370forlife]

Re: Rosewill RD600N-2SB-SL-BK review @ hardware secrets

This unit runs about $65, so no it's not "cheap" but it is cheap for a unit that delivers more than 600W reliably. I wouldn't imagine such a new platform from solytech would have the two transistor 5vsb circuit. The point of these reviews is to see if the unit delivers it's rated wattage and if it does it in spec. It would be very expensive and time consuming to do a long-term reliability test. The fact that this unit is very overspec'd on both it's primary and secondary is a good indicator of how long it could last as well as the capacitors used, which looked like either the puke green Anodia or Yang-Chung. You can really only infer on how long units in these tests will last, not test how long they will last.

Also if you look at the table he tests the 5vsb up to it's labeled wattage too, and tests the ripple, dosen't show the scope shot but just puts a pass or fail there as well as tests the short circuit protection on it. JG also does this, they only write about it if it is something bad, such as on the bestec atx-250-12E review at JG.

[i4004]

Re: Rosewill RD600N-2SB-SL-BK review @ hardware secrets

long term would be in the sense "i'll take this apart and see what's inside, and then(if i like it) will be putting it to my customers' machines" etc.
but given the price tag, you're right, it won't happen.

as-is, i don't really know where to put this psu, but i would rather get corsair for that money, even if it had somewhat less power.

what was it that you said about the caps? good or bad?
those names seem unfamiliar...
and pix don't really show them.

>The point of these reviews is to see if the unit delivers it's rated wattage and if it does it in spec.

true.
that's why i have more appreciation for good pix and descriptions posted by badcaps memebers than for such tests.
(and knowing that decent brands usually pass such tests without issues)

as for the spec, did anybody actually read the "specs", ie what's expected from psu and in what conditions, etc.
i think i recently saw thread at jonnyguru forum where it's suggested that tests(ie "standard" to use on psu testing) don't even specify what sort of load should be used, which brings me back to the premise i used last time, about mobo load being mighty different than the loads they use on tests, which would(if so) question the legitimacy of such tests overall.
which brings us back to the point about "field tests" being most important.
ie "i had such and such luck with such and such brand", etc.
either way, specs (ie methodology they all use to say what's ripple and on what load) should be read.

and indeed

page22 doesn't go into detail about the load at all(and if the 2 capacitors used on scoping can simulate the mobo, i'm an airplane).

i think i would prefer scoping the psu when attached to particular mobo, or mobos, like keri did.
(but one should know that another mobo could give different load, ie different results)

[370forlife]

Re: Rosewill RD600N-2SB-SL-BK review @ hardware secrets

what was it that you said about the caps? good or bad?
those names seem unfamiliar...
and pix don't really show them.

as for the spec, did anybody actually read the "specs", ie what's expected from psu and in what conditions, etc.

The ATX specification as outlined by intel calls for 5% regulation on all lines except for the -12v, which calls for 10% regulation. Ripple should not exceed 50mv on 3.3v and 5v lines, and 120mv on 12v lines, allowing for 150mv on both the 5vsb and -12v. The temperature rating is outlined by the manufacturer, solytech did not say what temp this is rated at, but it delivered more than it's labeled wattage in spec at over 40C.

The caps anodia and YC are known bad, but not like Fujiyuu's one year wonder caps, as long as they are well cooled they seem to do ok.

I would agree I would rather buy corsair or any other reputable seller over this, but the fact that a solytech could actually run it's labeled ratings and beyond at a high temperature with pretty good efficiency is a wonder in itself. Still, if I was in a pinch and needed a 600w unit capable of it's ratings for cheap, this would be it.

[Krankshaft]

Re: Rosewill RD600N-2SB-SL-BK review @ hardware secrets

Bah just because the 2 transistor 5 VSB circuit caused problems in one bad power supply because of a badly placed resistor near a bad cap doesn't mean it's not reliable with good caps.

Been running 4 totally recapped (all Fukyuus and Koshins purged) Antec SP and TP PSUs and no problems here. I believe they use the 2 transistor 5 VSB design.

I don't believe in reverse engineering unless the model of power supply in question has a history of that circuit failing.

I do agree though with ICs being so cheap a dedicated PWM with integral protection for the 5 VSB wouldn't have cost that much.

[i4004]

Re: Rosewill RD600N-2SB-SL-BK review @ hardware secrets

The ATX specification as outlined by intel calls for 5% regulation on all lines except for the -12v, which calls for 10% regulation. Ripple should not exceed 50mv on 3.3v and 5v lines, and 120mv on 12v lines, allowing for 150mv on both the 5vsb and -12v. The temperature rating is outlined by the manufacturer, solytech did not say what temp this is rated at, but it delivered more than it's labeled wattage in spec at over 40C.

what i'm aiming is this: i think that if you were to connect same psu to 3 intel mobos of different vintages(say this year's, 2007 and 2005) you would probably see 3 different ripple oscillograms for 3 different mobos.
and ripple as measured by hooking it to sunmoon would probably be different than either of those 3.
ie the importance of load you test psu on.
(in a same way psu has smps switchers that are making noise on primary and is transferred to secondary, mobo cpu vrm also is a switcher in itself...and it adds to noise on the secondary of psu
http://en.wikipedia.org/wiki/Voltage_regulator_module )

the aim should be to be close to "real life" conditions upon measuring, and i dunno how close sunmoon resembles mobo...my guess is "not at all", as i doubt it employs cpu vrms mobos use.

it's good thing that rosewill is getting better, sure, because they would probably go bust if they were to continue the old path.
overall things are improving in mid and higher range units, as for low end, i dunno, but perhaps even there mobo killers are now eliminated.

[i4004]

Re: Rosewill RD600N-2SB-SL-BK review @ hardware secrets

>Bah just because the 2 transistor 5 VSB circuit caused problems in one bad power supply because of a badly placed resistor near a bad cap doesn't mean it's not reliable with good caps.

it wasn't realyl just one, there were many cheap designs with it, ie many examples of mobo killers.
if you recently know only of bestec fiasco, i know of at least few more in my part of the world.

as for "just caps", ok, but how did you solve the issue of that resistor heating too much? putting good cap won't solve the resistor overheating, will it?
and even best caps don't like the heat....and otoh, even badcaps will work for a long time if cooled properly...examples of badcaps working for 6-7 years is something i see very often...i tell people "you should change this" as soon as i see it..but they usually opt to change it only when it fails...luckilly, in most cases they get away with it...ie mobo killers are not that prevalent.
but once spotted, i would certainly apply everell's mod than to leave it as is.
offcourse if i must repair....and i don't, i put fortron and that's it.
i really need that peace of mind because crappy psu destroyed my first pc.

[370forlife]

Re: Rosewill RD600N-2SB-SL-BK review @ hardware secrets

The thing about the sunmoon and techred's and other ATE equipment is that they provide a worst case scenario testing environment. With a normal computer, the load is low until the computer needs more power, which it might fluctuate from low to high to medium to high to low, etc.....

Say you had this power supply hooked up to a pretty standard core 2 duo 2.4ghz with a 9800gt, 1 hard drive, 1 cd/dvd drive, 3 fans. At full load you are probably pulling no more than 300W while playing a high graphics game and really loading it down. The load may fluctuate while playing this game anywhere from 200W to 300W, then 300W down to 200W. The sunmoon will load it to 300W and hold it there, constantly pulling 300W. The thing about this is you know exactly what wattage you are pulling from the power supply and you can measure exactly what ripple and what voltage it is running at.

It's like test driving a car. In real life, most people stop and go, never really going faster than 55mph, and keeping it under about 3000rpm's. Then you go and test it while finding the top speed of the car and bounce it off the rev limiter for a while. Ya, it's probalby not real world but if it dosen't blow up at 7-8 grand going 150mph than it probably won't blow up putting around town or driving down the highway.

as for "just caps", ok, but how did you solve the issue of that resistor heating too much? putting good cap won't solve the resistor overheating, will it?

The resistor typically dosen't overheat, it just gets hot. I believe a couple of people here have made circuits that remove the resistor and that anti-frost diode and use it instead to keep the fan running during standby.

[i4004]

Re: Rosewill RD600N-2SB-SL-BK review @ hardware secrets

which just strains the fan, and it may fail when you really need it(when psu is on).
screw that.

>It's like test driving a car. In real life, most people stop and go, never really going faster than 55mph, and keeping it under about 3000rpm's. Then you go and test it while finding the top speed of the car and bounce it off the rev limiter for a while. Ya, it's probalby not real world but if it dosen't blow up at 7-8 grand going 150mph than it probably won't blow up putting around town or driving down the highway.

try this: sunmoon test more resembles trying the car using the fuel it'll never use in normal day-to-day usage.

are you familiar with different forms of load when it comes to electricity, ie resistive, capacitive and inductive load.
for example
http://www.markallen.com/teaching/uc...lecture3/3.php

now, what i'm saying is that in this aspect, sunmoon and mobo are not same type of load.
think about truck carrying a solid concrete block(sunmoon load) vs. truck carrying a big tank of water(same total weight as concrete block) that's half empty(mobo load): which load will destroy the truck faster?
well, you know water moving around will strain the truck more in all aspects(engine, suspension, gearbox).
and you drove the block for 5miles, and you drive the water tank for years, day-in, day-out.

it's not just the total watts, but what those watts are made of, so to say.

offcourse, if you can prove mobo(which psu will be powering, psus are not bought to power sunmoon testers) is a pure resistive load, i'll concur...but i don't hink you can.
mobo is a complex load, and sunmoon is not.

like gianni said here
https://www.badcaps.net/forum/showth...robe#post60210
"but I think you will have a lot of spike due to the switching regulator and CPU operation because the working frequencies are high."

[370forlife]

Re: Rosewill RD600N-2SB-SL-BK review @ hardware secrets

The sunmoon wolf uses in the JG review has the ability to cycle through the tests rapidly, if you watch one of the videos in the bestec review you can see it goes through the 5 different loading stages very quickly.

The main problem about this is you cannot measure the watts of the power supply if it is running in a computer accurately. This is what so many psu reviews out there do and it is wrong. They will pass a leadman psu because it can power a computer without 'sploding. But how many watts was actually used? Don't know. Hook it up to a sunmoon or a techred and you can precisely see how many watts it took to blow it up.

Fact of the matter is a psu load tester is as close as you can get to a real world scenario without having to build 4 or 5 different computers with different power consumptions to test psu's ($$$$) or until a company comes out with one that accurately simulates a real computer ($$$$)

[i4004]

Re: Rosewill RD600N-2SB-SL-BK review @ hardware secrets

>Fact of the matter is a psu load tester is as close as you can get to a real world scenario without having to build 4 or 5 different computers with different power consumptions to test psu's ($$$$) or until a company comes out with one that accurately simulates a real computer ($$$$)

true