Part 1: Four Cheap 80 PLUS Bronze Power Supplies, Reviewed

[ddpruitt]

There is, however, a term called 'total harmonic distortion', which (IIRC) refers to the amount of noise produced by a device..

THD is already accounted for in the PF measurement as the "distortion factor" which accounts for all power transferred at frequencies other than the fundamental.

So on DC which has no phase error, only the distortion factor remains. What you said was implied in what I said.

This only applies to a purely resistive load, go outside of resistive and the numbers may be different.

 

[InvalidError]

So on DC which has no phase error, only the distortion factor remains.

This only applies to a purely resistive load, go outside of resistive and the numbers may be different.

Huh? It applies to all loads.

The power factor is the percentage of apparent power provided by the source that actually ends up consumed by the load. If the grid has to provide 200VA to power your 100W load due to phase shift and distortion factor (harmonics, crest factor, etc.), the overall power factor is 0.50 or 50%, meaning your 100W load puts twice as much strain on the grid than it ideally should for the power it requires. That's why companies are often billed based on VA rather than Watts - to incite them to keep their power factor in check.

Net power transfer as far as a monotonous source is concerned only occurs at the source's fundamental frequency and for DC sources (PSUs) feeding DC-DC converters, that would be 0Hz. If you know integrals, you can try doing the Fourrier integral or transform for v(t)i(t) yourself for a source of a given frequency and a given current (load) waveform. You will find out that only 0Hz has any net power on a DC source and only 60Hz has any net power on a 60Hz AC source. All other frequencies will average zero (no net transfer) over a common factor number of cycles.

But capacitor and inductor impedance only matters for AC. On DC, capacitors have almost infinite impedance while inductors are almost short-circuits apart for their leakage and wiring losses so unless output voltage or currents change, they are practically invisible to their source.

Since DC has no phase at its fundamental frequency to worry about, power factor simplifies down to DC current / RMS current and due to how little power noise and ripple contribute to RMS unless you have a really awful DC-DC converter, the result will be very close to 1: if you have a 5A load that produces 1A RMS ripple+noise on its input, your power factor would still be 0.98.

With sufficient decoupling to minimize operating noises and ripple at their operating frequency. DC-DC converters operating at constant output power will look almost perfectly resistive between load changes as far as their upstream DC source is concerned.

 

[smeezekitty]

I love the redneck load tester!

 

[acreoaeneas]

I read:

We even asked Enermax to be quiet! to help us troubleshoot the issue with their equipment.

 

[acreoaeneas]

Double post, please delete.

 

[Patrick Tobin]

Lol, that Antec HCG-400 is a discontinued product!

 

[vertexx]

Anyone up for taking a collection for a Corsair cx430 to be included in the next round?

 

[computer_nugget2]

always judge a power supply by its weight.

While weight is an important factor, overall weight is 100% useless

for example check out the SIGMA SHARK SP-635

http://www.ocia.net/reviews/sigma635/page2.shtml

it is probably one of the heaviest power supplies you will find (with the exception of some of the 1200 watt ones)

I purchased it a while back and was disappointed at the internals. they used an insanely thick case that made the power supply weigh a lot but the insides were were the result of the owner of the company asking while walking around china, what are the cheapest items we can find on the shenzhen market today.

(at least the power supply can double as a flail or a boat anchor)

thats why, always judge it.

 

[smeezekitty]

always judge a power supply by its weight.

While weight is an important factor, overall weight is 100% useless

for example check out the SIGMA SHARK SP-635

http://www.ocia.net/reviews/sigma635/page2.shtml

This is somewhat off topic but, the reviewer had no clue what they were talking about.
They were so concerned about 11.91v on the +12 but that is within 1% WELL within normal limits.

I would be more concerned about the 4.8v (pushing 4% tolerance) on +5v.

 

[computer_nugget2]

always judge a power supply by its weight.

While weight is an important factor, overall weight is 100% useless

for example check out the SIGMA SHARK SP-635

http://www.ocia.net/reviews/sigma635/page2.shtml

This is somewhat off topic but, the reviewer had no clue what they were talking about.
They were so concerned about 11.91v on the +12 but that is within 1% WELL within normal limits.

I would be more concerned about the 4.8v (pushing 4% tolerance) on +5v.

yeah, but you know how most computer people are, they demand exact perfection or nothing at all.

 

[computer_nugget2]

crap double post, stupid error 500.

 

[mpotran]

Please make some reviews about LC-Power german PSUs. They are reasonably priced and should be good quality. A handful of their PSUs are marked bronze, silver or gold - i wonder if they stand up to their marketing.

 

[flong777]

Wow I am so unimpressed. At least they won't burn up your PC, but that is not saying much.

 

[Someone Somewhere]

Even then, some of those caps are from downright shoddy brands. Not so bad on these ones, but part two has several with Samxon. *shudders*

 

[5ocrates]

Probes with higher bandwidth have lower capacitance. If you do not calibrate them with your oscilloscope, you end up with mysterious spikes. The higher the frequency of the measured signal, the bigger the spikes.

 

[InvalidError]

Even then, some of those caps are from downright shoddy brands. Not so bad on these ones, but part two has several with Samxon. *shudders*

Samxon is usually fine for input caps since input caps are subjected to relatively low current ripple.

What usually kills output caps is being grossly under-rated for current ripple: many cheap PSUs use capacitors rated for 700-1200mA on rails rated for over 20A. If you are drawing 10A from one of those rails, the RMS current ripple on the capacitor will likely be over 3A and the PSU is destined for guaranteed short/medium-term failure.

Using knock-off brands might not help but using grossly under-rated parts from one of the top-brands wouldn't save poorly designed PSUs either. Under-rated parts remain under-rated parts regardless of vendor.

 

[JackNaylorPE]

Power supply manufacturers condition us to think that efficiency is the one variable you need to pay the most attention to.

No, Id have to disagree ..... reviewers seem to think that efficiency is the one variable we need to pay the most attention to. This statistic is emphasized above all else in every review and the one I pay least attentions to. Voltage stability is the 1st thing I look at...... a review which fails to include such information is one I consider not worth reading.

There's been several reports here on the THG forums where peeps experiencing system crashed under heavy loads had a CX series PSU that fails to maintain the 5% max voltage variation on the 12v rail resulting. This kind of information I think is more important than whether I am going to spend an extra 70 cents a month on electricity.

 

[smeezekitty]

Power supply manufacturers condition us to think that efficiency is the one variable you need to pay the most attention to.

No, Id have to disagree ..... reviewers seem to think that efficiency is the one variable we need to pay the most attention to. This statistic is emphasized above all else in every review and the one I pay least attentions to. Voltage stability is the 1st thing I look at...... a review which fails to include such information is one I consider not worth reading.

There's been several reports here on the THG forums where peeps experiencing system crashed under heavy loads had a CX series PSU that fails to maintain the 5% max voltage variation on the 12v rail resulting. This kind of information I think is more important than whether I am going to spend an extra 70 cents a month on electricity.

80+ Certification means quite a few things that suggest a better supply.

A manufacturer is likely not going to pay to have their supply go through testing if they didn't feel it was decent enough to be worthwhile.

Higher efficiency also suggests higher quality components were used and design we developed to keep efficiency up.

80+ testing also verifies the supply can actually output its labeled power output.

Finally, a more efficient PSU generates less heat which means it should potentially last longer. Electrolytic capacitors hate heat

 

[InvalidError]

80+ Certification means quite a few things that suggest a better supply.

A manufacturer is likely not going to pay to have their supply go through testing if they didn't feel it was decent enough to be worthwhile.

Higher efficiency also suggests higher quality components were used and design we developed to keep efficiency up.

80+ testing also verifies the supply can actually output its labeled power output.

Finally, a more efficient PSU generates less heat which means it should potentially last longer. Electrolytic capacitors hate heat

80+ / 80+ Bronze targets are too easy to meet and while 80+ certification may prove that a PSU is able to meet 80+ efficiency targets, it does not mean much in terms of how well its output regulation and noise will hold through the process nor how long the unit will manage to do so before failing.

You have to step things up to 80+ Silver if you want PSU manufacturers to seriously have to try. This does not guarantee the PSU will retain its 80+ Silver qualification through its lifespan but it does mean that most of what went in it is at least a cut above.

While heat might not be electrolytics' best friend, the main reason electrolytics get hot in the first place is because poorly designed PSUs often use grossly under-rated caps so those caps are destined for imminent failure even if you manage to keep them at room temperature.

 

[Someone Somewhere]

Yup. So long as it falls within the (very old, very easy to meet) ATX standard (and I'm not sure if they even test that), they don't care about ripple or regulation. And they get worse as it ages.

Cap life also depends on the fan speed - in most cases a PSU will work fine at high temps for a few weeks/months, which is enough to get it through testing. A slow/quiet fan and higher temps makes it more appealing to consumers.

Cap temp also depends on the layout - how much heatsinking there is and how many hot FETs are nearby. If you reduce the heatsinking, the caps might get warmer.

Besides, there's no reason you can't do a Gold PSU with crappy caps. I can't find anything, but I doubt this uses anybody reliable: http://pcpartpicker.com/part/raidmax-power-supply-rx850ae

 

[smeezekitty]

Of course.
80+ is is no guarantee of quality, but in general an 80+ PSU probably would be a bit better in quality then a compilable non-certified supply.

 

[Someone Somewhere]

Yes. That's why I think that the only thing 80Plus is good for is the test of being able to output rated power without immediately blowing up or shutting off.

Gold does not necessarily mean more reliable than Bronze, however.

 

[InvalidError]

Gold does not necessarily mean more reliable than Bronze, however.

I think it is fairly reasonable to expect a 80+ Gold to take longer to degrade to 80+ level. Only way to know for sure though would be to give them a good burn-in at high-power (say 80% of max continuous power for a month at 40-45C) and see how much that beats them up.

 

[JackNaylorPE]

Well look where it's got the buying public..... thinking they got a great PSU because it says Corsair on it, because it says it's 80+ and THG says it's great and thru all of that there are forum posters complaining it's delivering 11.3 volts on the 12 volt rail.

It's like saying a car is reliable because it has high mpg.

 

[Someone Somewhere]

Gold does not necessarily mean more reliable than Bronze, however.

I think it is fairly reasonable to expect a 80+ Gold to take longer to degrade to 80+ level. Only way to know for sure though would be to give them a good burn-in at high-power (say 80% of max continuous power for a month at 40-45C) and see how much that beats them up.

No, because you're still assuming that efficiency = reliability.

My point is that a 80Plus Gold PSU with bad caps and no protection can still blow up your system if it fails, while even an uncertified one ccould be fine if it's built properly.