Active PFC for PSU? Yes or no?

[bhcompy]

surprised noone has mentioned sparkle.. but then again, they're only a huge name in servers. i wonder why sparkle/fsp are run in so many high end servers? ah.. reliability, active pfc, wattages, etc. imho, do yourself a favor and get a sparkle/fsp 550. you'll be more than happy and it's cheaper than "designer" psu's

 

[Gothgar]

people always crap on TT psu's, but they are good psu's for a good price.

I have a TT 420 and a TT 480, both are amazingly stable, clean power, and they are way cheaper than a counterpart from another manufacturer

 

[mindless1]

TT are a good value in 340W and higher but their 300W & lower are decidedly worse, built completely low-end for cheap case bundling instead of proportional compared to their higher wattage models.

 

[lrmat]

mindless1, i can't help but feel that some of your comments in your previous reply are partly directed to me (which is fine by the way). my Tt crapped out but that was an unfortunate event, it happens once in a while and i don't hold it against Tt; they took my rma and are sending me a replacment. the problem was i couldn't wait, i built a (what i think is) a nice rig and i wanted to use it. i wanted to finish transfering my files and give my dell to my wife, i could not wait any longer and i didn't. the ultra appealed to me for purely cosmetic reasons, the sleeved cables, the modular design, i only connect the cables i need (my case is a Tt tsunami with a window) giving a clean look. i decided f*&k it! i'll get another psu and keep the Tt as back up so if god forbid this happens again, i will only be slowed down and not stopped dead in my tracks.

i could instead put the Tt back in instead of the ultra and i truly don't doubt that this time it won't crap out (even chintzy psu's last, my first build was with cheap parts and i think that machine is probably still running all these many years later) but the ultra asthetically compliments the tsunami case MUCH better then Tt's own psu. what can i say? IT'S PRETTY!!!

 

[mindless1]

lrmat, it wasn't meant as a rant against you, I feel more like anyone who paid top dollar (or even $70) for an Ultra has been cheated. Customers should be able to buy product and count on specs, that the specs adhere to same standards set forth by the major name-brand PSU manufacturers. Eye candy DOES cost money, it's only fair to charge more for a PSU having it, but hundreds of percent more? Plus that sidesteps the whole capacity issue.

I do think the Ultra looks nice. If it were a Sparkle, Delta, PCPower & Cooling 50W+ all chromed and such I'd call it a bargain to get it for less than $120, but Ultra isn't even close to any of the aforementioned brands' 500W PSU and that's not fair to the customer.

 

[Antoneo]

Originally posted by: bhcompy
surprised noone has mentioned sparkle.. but then again, they're only a huge name in servers. i wonder why sparkle/fsp are run in so many high end servers? ah.. reliability, active pfc, wattages, etc. imho, do yourself a favor and get a sparkle/fsp 550. you'll be more than happy and it's cheaper than "designer" psu's

Is it sparkle/fsp/fortron?

 

[imported_jediknight]

Originally posted by: Elcs

Originally posted by: amdguy
don't get thermaltake PSU, they are CRAP VERY CRAP!!!

get enermax or antec !!!!

Ignore this person.

Thermaltake in general are crap. Their PSU's are not. Quality units. Like all manufactured items, they can fail just like any Enermax or Antec.

Ask either one of my 2 TT PSU's. They have no problems handling mine or my dads machine.

Their manuals are crap, though.. reads like someone in Taiwan (or wherever TT is based) just ran their manual through babelfish or something.. not that you need much of a manual for a PSU, but anyway..

 

[sharkeeper]

This is exactly the kind of illogical thinking that causes false impressions.
OCZ has won editors awards because the editors thought they looked good. Frankly I'm sick of arguing about it, over and over clueless people review them and don't bother to do accurate testing of full load capacity. When a PSU labeled as 500W can't sustain that, it's fraud. No CPU or video card overclock requires remotely near 500W, and yet THEY chose to label them as 500W, which is not some kind of "guideline", it is a precise scientific measurement of power that is not variable based on personal interpretations or marketing gimmicks. Nor is it basked on chrome, cable sheaths, LEDS or clear fans, etc, etc, etc.

This is no different than the Pyramid EQ boosters that claim 600watts yet connect to your car's fusebox with a 16 gauge wire!

The same could be said about computer speakers...

If these people bypassed the laws of physics, we would no longer depend on oil and have 400 hp cars running on drycell batteries!

Cheers!

 

[Zepper]

Actually Enlight doesn't even make their own PSUs any more, so how could they be making Tt's PSUs? It is Sirtec (HighPower is its direct brand) that actually makes most of both brands. Some Sirtec models are very good (the Enlight, et al., 360W unit for example) and others aren't so good (e.g. the 300 and 400W units).
. Ultra is Powmax inside - crap is crap. One should know it is crap just from knowing its principals (Infotel/TigerDirect). Those clowns are always looking for junk they can sell at an extreme margin.
The regular Powmax 480W unit is $12.00!!! on Newegg and the Ultra-X is around $100... Man oh man, what a margin!!!
. Fortron makes at least the top Sparkle line (the ones with FSP in the model no.). And that line (with any brand on it as you can find FSP PSUs under lots of rebadges (AOpen, InWin, etc.)) is excellent. As far as I know even their budget lines hold up well.
. Silverstone PSUs made by Enhance are also very good.
. Active or passive PFC is important if you have lots of PCs as you can save some money over the long term. It is also good if you care about the environment as they present the most friendly load to the power grid and can help reduce the rate at which generation capacity may need to be added.
. As another said, just slapping an active PFC circuit on top of a pre-existing design isn't the best way to go, but it does the job. Another thing is that most PSU makers only put active PFC (which makes it look like the Current draw curve of the PSU is lined up with the voltage curve - IOW makes the PSU look like an ordinary resitor to the power grid rather than an active circuit) on their top line models. PFC is not a protection circuit.
. Re. PSU reviews on techie sites. None of them test PSUs to failure - which to my mind is the only way to really test them. How do they react at a near-failure state and what is their failure mode? IOW, is it likely to take your mobo or drives along with it if it fails? Most of the tests don't even bring all rails to half power simultaneously. When is the last time you saw an O-Scope trace of the outputs from a PSU at a near-failure state in an online review? Take those that don't test as above w/ a grain of salt!
.bh.

 

[nealh]

Ok....mindless1

if you are saying OCZ Powerstream is fraudulent in rating their PSU at 520w ..shows us the data..I have read countless reviews on PSU recently...the OCZ seems built better as compared to Antec NeoPower, 550w truepower, enermax noisetaker or at least as good

These review sites open and looked at the components..

so please show me why the PSU I bought is crap and cheap....

Just note I never said it was the best but I would like proof it is a POS ... please show me a review or website..because I do not want to pay a premium for crap that is so highly regarded

My was just delivered to day by fedex

 

[whattaguy]

I have a TT Purepower 480 with PFC in my system, and it's been holding its own just fine.

DFI Lan Party UT NF3
AMD64 3000+
1 gig Patriot (2-3-2-5)
4 HDD's (2 on the SATA chain)
2 Optical drives
2 Recording sound cards
1 4-port USB hub
MSI 9800Pro
1 floppy drive

haven't had a problem with this PSU for over a year.

 

[nealh]

BTW which PSUs are actually true to there rating ..is there a place with a list...

I am pissed if the OCZ PSU is not very good

 

[mindless1]

Originally posted by: Zepper
Actually Enlight doesn't even make their own PSUs any more, so how could they be making Tt's PSUs? It is HighPower that actually makes most of both brands. Some HighPower models are very good (the Enlight, et al., 360W unit for example) and others aren't so good (e.g. the 300 and 400W units).
. Ultra is Powmax inside - crap is crap. One should know it is crap just from knowing its principals (Infotel/TigerDirect). Those clowns are always looking for junk they can sell at an extreme margin.
The regular Powmax 480W unit is $12.00!!! on Newegg and the Ultra-X is around $100... Man oh man, what a margin!!!

"Highpower" is just another (re)label, like Enlight and Thermaltake- All are Sirtecs. You have a pretty good grasp of what's going on with the Ultras, they just make them pretty as a marketing tool.

 

[Zepper]

Yeah, I knew that Sirtec was the actual mfr. when I wrote the earlier post, it just slipped my mind at the time. Thanks for the missing piece.
.bh.

:moon: Man it gets dark early now...

 

[irrationalJohn]

Originally posted by: mindless1
Active PFC uses more power than "same" passive PFC unit, many people mistakenly assume the opposite.

I am one of those who believe(d) the opposite. One of the reasons I thought this way is the claim below taken from this TT webpage.

Non-PFC offers around 0.5~0.6 PF (Power Frequency), 40%~50% power lost.
Active PFC provides more efficient PF(Power Frequency), 0.95~0.99, it means only 1%~5% power has gone.

I don't claim to be anything other than an ignorant idiot. So I'd appreciate it if someone would reconcile "uses more power" with the "more efficient" TT marketing-speak.

-irrational (and often ignorant) john

 

[Dethfrumbelo]

Originally posted by: fanerman91
I LOVE my Seasonic Super Silencer 460 Watt. It has active PFC plus lots of other useful bells and whistles. It's also very close to silent, with a nice automatic fan control.

I think I'll end up going this way as well. I like the fact that Seasonic is actually honest about their power ratings, where most manufacturers give you peak numbers which aren't practical in a real world situation.

 

[mindless1]

Originally posted by: zjohnr

Originally posted by: mindless1
Active PFC uses more power than "same" passive PFC unit, many people mistakenly assume the opposite.

I am one of those who believe(d) the opposite. One of the reasons I thought this way is the claim below taken from this TT webpage.

Non-PFC offers around 0.5~0.6 PF (Power Frequency), 40%~50% power lost.
Active PFC provides more efficient PF(Power Frequency), 0.95~0.99, it means only 1%~5% power has gone.

I don't claim to be anything other than an ignorant idiot. So I'd appreciate it if someone would reconcile "uses more power" with the "more efficient" TT marketing-speak.

-irrational (and often ignorant) john

"PF" is Power Factor. Their website is either wrong or translated incorrectly to english (same diff, still wrong).

Google groups thread
Active PFC's reduction of peak current decreases loss slightly in wiring and transformer, but it, itself, causes loss. The difference isnt enough to be relevant for relatively low-powered equipment like only a few PCs, the real issue is whether your home AC circuits can handle the current (most have no problem) and whether you're an industry billed disproportionately more for higher current, since it takes a larger percentage of their (the electric co.) capacity to supply current to equipment with lower power factor, even if the equipment used same amount of power either way.

 

[irrationalJohn]

Mindless1,

Thank you for your post, especially for the link to the sci.electronics.equipment newsgroup thread. It helps. At the continued risk of being wrong, what I think I know now is that the Power Factor is not related to how much power is consumed by the power supply. Rather the PF is an indication of how "efficiently" (??) the power used is delivered from the source to the load.

Hmmmmmmm, so in my case focusing on PF alone in the market-speak is actually very misleading. The measurement I think myself and others are actually interested in is the EFFICIENCY of the power supply, no? The higher the efficiency of the power supply the smaller the portion of the total power consumed by the power supply which is wasted as heat, correct? Would I also be correct in assuming that if one power supply has a higher efficiency than another that ... all other things being equal ... the more efficient power supply would (1) cost less to operate and (2) generate less heat in the PC case?

-irrational john

 

[jterrell]

lrmat,
a spare psu is indeed a good thing.
its generally a good first checl when your box wont boot or depending on who this happens to. lol.

 

[mindless1]

Originally posted by: zjohnr
Mindless1,

Thank you for your post, especially for the link to the sci.electronics.equipment newsgroup thread. It helps. At the continued risk of being wrong, what I think I know now is that the Power Factor is not related to how much power is consumed by the power supply. Rather the PF is an indication of how "efficiently" (??) the power used is delivered from the source to the load.

Hmmmmmmm, so in my case focusing on PF alone in the market-speak is actually very misleading. The measurement I think myself and others are actually interested in is the EFFICIENCY of the power supply, no? The higher the efficiency of the power supply the smaller the portion of the total power consumed by the power supply which is wasted as heat, correct? Would I also be correct in assuming that if one power supply has a higher efficiency than another that ... all other things being equal ... the more efficient power supply would (1) cost less to operate and (2) generate less heat in the PC case?

-irrational john

Power Factoring does slightly increase the efficiency of delivery, but it's not really the efficiency at issue, it's the ability to deliver it. For example, if your system had a Via C3 CPU, integrated video, and 12 hard drives, it would be a relatively steady load, the current might not change so much, it "might" consume 270W. On the other hand, if you had a P4 3.4GHz, Geforce 5900, and 4 hard drives, you might also consume 270W, BUT the current demand would greatly vary, above and below the level needed by the Via C3 box in this example. So, a typical 300W PSU might power the C3 box but you'd need more for the P4 box because the peak amperage is higher. Similar applies to the power company, they need a larger power "supply" to deliver higher amps to you, the more your equipment has current swings, and "same" equipment with lower power factor has higher current swings.

Efficiency is good, it will produce less heat, but keep something in mind. In the same way that input PF causes loss due to filtering, so does other filtering in a power supply. Take same PSU with very good output filtering, a clean power source (as much as possible considering it's a SMPS), and reduce it's filtering by component substitution, and suddenly the power supply is more efficient... but not for a good reason. Another way to increase efficiency is by careful matching of power supply to load such that power supply is very near it's max capacity, but it must be matched on multiple voltage rails AND running such a power supply near this max load will tend to wear it out faster, due to the higher wattage supplies having better heatsinks, (often fans) and capacitors... or even more protection and regulation circuitry allowed by the higher price-points set for the higher wattage models. So, again seeking a higher efficiency "may" have drawbacks.

Then there's yet another issue, ratings standards. We can clearly see that all PSU are not rated the same per capacity even when this is a pretty standardized way to rate PSU, or at least it should be pretty standard, I'm one of those who considers it fraud when a PSU's label states a wattage it can't sustain and doesn't explicitly mention that the wattage is only a Peak rating. Point being, there's even more room to fudge numbers when spec'ing an efficiency rating, I would tend to believe the efficiency ratings from the main manufacturers much moreso than relabelers or generics. Another alternative is to consult reviews that test efficiency (though they are rarer), providing they also account for the matching of system load to PSU amperage specs as i mentioned earlier, else they're only comparing efficiency of PSU when mated with that particular load distribution, system it's powering for the test. Generally speaking the easiest way to get a good PSU is to just buy a name-brand that bears the manufacturers name on it, though there are some fairly decent relabels too like Zalman, Antec, Thermaltake.