EVGA 450 B3 PSU Review

[Yuri_BR]

You should buy one of the Aerocool P7 power supplies. Here is a review of the 650W https://www.techpowerup.com/reviews/Aerocool/ACP-650FP7/ - it doesn't get much quieter than that. It's an incredible PSU.

That PSU would indeed fit all my needs. Amazing review, dind't know about this PSU brand!

The problem is that they don't sell this PSU brand here in Brazil, and if I would buy it on ebay, the taxes would be too high (like 100%).

The options I have here at a resonable price are (1 USD= 3,3 R$)

Corsair RM750I.......R$781 (or 236 USD)
EVGA 850 P2..........R$809 (or 245 USD)
EVGA 750 P2..........R$780 (or 236 USD)
EVGA 750 G3.........R$730 (or 212 USD)
EVGA 850 G3.........R$800 (or 242 USD)

Corsair RM650I......R$714 (or 216 USD)
EVGA 550G3.........R$524 (or 158 USD)

Among those, what would the best pick?

 

[the nerd 389]

IMHO, if a PSU represents a potential fire hazard, it seems fitting that it should not receive a better score than a PSU that doesn't.

EVERY PSU or device handling high power is a potential fire hazard and not immune to the occasional material or manufacturing defect that could cause an otherwise good design to fail catastrophically.

With a sample size of one, it is impossible to tell whether this is an isolated defect, a rampant QA problem from a parts supplier or an outright design flaw. The only thing that can be said is that one sample failed. While this may not inspire confidence, it isn't sufficient to definitively condemn the product.

I agree that a sample size of one means that anomalies will occasionally occur. I will defend my initial comment by simply stating that you can solve that issue by requesting a second sample if the first fails catastrophically. That doesn't seem unreasonable to me. Regarding all PSUs being potential fire hazards, you are technically correct. In practice, I think it's reasonable to assert that if a PSU destroys itself catastrophically on a reviewer's workbench, and the manufacturer fails to provide a replacement or explanation, then that PSU should probably not receive a performance score that ignores those facts.

 

[turkey3_scratch]

-snip-

That PSU would indeed fit all my needs. Amazing review, dind't know about this PSU brand!

The problem is that they don't sell this PSU brand here in Brazil, and if I would buy it on ebay, the taxes would be too high (like 100%).

The options I have here at a resonable price are (1 USD= 3,3 R$)

Corsair RM750I.......R$781 (or 236 USD)
EVGA 850 P2..........R$809 (or 245 USD)
EVGA 750 P2..........R$780 (or 236 USD)
EVGA 750 G3.........R$730 (or 212 USD)
EVGA 850 G3.........R$800 (or 242 USD)

Corsair RM650I......R$714 (or 216 USD)
EVGA 550G3.........R$524 (or 158 USD)

Among those, what would the best pick?

The RMi will be the most quiet ones, probably those. I really don't know your components, though, so it's hard to advise. Maybe start a new thread.

I agree that a sample size of one means that anomalies will occasionally occur. I will defend my initial comment by simply stating that you can solve that issue by requesting a second sample if the first fails catastrophically. That doesn't seem unreasonable to me. Regarding all PSUs being potential fire hazards, you are technically correct. In practice, I think it's reasonable to assert that if a PSU destroys itself catastrophically on a reviewer's workbench, and the manufacturer fails to provide a replacement or explanation, then that PSU should probably not receive a performance score that ignores those facts.

The protection circuits all working is something of extreme importance. Protection for any electronics, that is. If theoretically this unit is just a fluke and OPP is working on most other units, then I'd still be unsatisfied. A good manufacturer should have good quality control and ensure that all the PSUs on the line are in working order. They usually hook them up to a tester of sorts, I would expect them to test all the protections before sending the unit out. Then again, I could be all wrong here since I don't work in the industry (but hope to in 4 years). Of course duds happen, but I am honestly doubtful that is the case, this to me seems like the engineers didn't properly implement a protection.

 

[InvalidError]

Of course duds happen, but I am honestly doubtful that is the case, this to me seems like the engineers didn't properly implement a protection.

Vibrations can break components. All you need s a broken SMD diode, capacitor, resistor, etc. during shipping in a semi-critical circuit such as the APFC's snubber (a simple resistor-capacitor-diode circuit meant to absorb energy from the transformer core's stray inductance to relieve switching stress from the FET) to cause a catastrophic failure.

I'd expect at least a statement about what the OCP/OPP limit is supposed to be. The other option would be reverse-engineering the board to find how the OCP/OPP is implemented, locate the range-setting resistors and determine the limits from those if they haven't been charred beyond recognition.

 

[turkey3_scratch]

Yeah packing foam would have been a better idea on their part.

 

[Aris_Mp]

Doing a reverse engineering in this sample and find OCP/OPP limits will be time consuming without having the circuit's design and I have around 30-40x more PSU reviews to work on along with other things.

This isn't the first B3 that died on my hands. More on the next review(s).

 

[InvalidError]

This isn't the first B3 that died on my hands. More on the next review(s).

Same failure mode or something different? If it is the same or similar, then that would already strongly suggest a design flaw.

 

[Onus]

If more of them croaked similarly, then perhaps EVGA laid an egg with this one.

 

[turkey3_scratch]

Aside from the catastrophic failure alone, I'd personally rather see better build quality sacrificing performance than vica versa.

 

[tsnor]

Yeah it could have better holdup time, though that's not very important to most people.

... until they see their PC crash when the lights flicker ... so they buy a cheap UPS that doesn't help when the lights go out because the cheap UPS can't switch-over fast enough when their PSU's hold up time is so short.

Agree you are right that most people buy crap PSUs based on $$ per rated watt delivered and do not know or care about holdup time.. Sigh. But they SHOULD care about holdup time.

 

[turkey3_scratch]

Yeah it could have better holdup time, though that's not very important to most people.

... until they see their PC crash when the lights flicker ... so they buy a cheap UPS that doesn't help when the lights go out because the cheap UPS can't switch-over fast enough when their PSU's hold up time is so short.

Agree you are right that most people buy crap PSUs based on $$ per rated watt delivered and do not know or care about holdup time.. Sigh. But they SHOULD care about holdup time.

You can't tell people what to and not to care about. I don't care much about holdup time. To be more explicit, I care that AC_LOSS to DC_LOSS is higher than AC_LOSS to PWR_OK. I don't like when the PWR_OK signal cuts after the 12V voltage is out of spec.

Normally, the PSU wouldn't crash, either, it would shut down because that's what cutting the PWR_OK signal does (after which the motherboard ignites PS_ON signal to the IC on the PSU).

If somebody is buying a UPS, it's their job to do their research to make sure the UPS's initiation time is lower than AC_LOSS to PWR_OK. Can't blame the PSU if people don't find a compatible UPS.

Plus, holdup time is far higher when not under 100% load, and most people don't have their PSUs loaded so high.

Probably holdup time, though, is down near the bottom for me in terms of importance. Fairly reliable power here in America.

 

[Darkbreeze]

So, since EVGA says that they are switching OEMs on this unit, can we expect to see all of these (Hopefully) B3 models revisited? Is it likely that these design flaws can be easily corrected or is it, as it seems it would be, a case where they would actually have to fully go back to the drawing board and completely redesign the platform which is something I don't really see them doing?