Evga G2 series concerns

[turkey3_scratch]

You know I own the EVGA 850 G2 and have never had an issue with brownouts. IN fact, I've had situations in my house where the lights gradually darken and lighten as the AC voltage amplitude fluctuates, and my computer has remained on perfectly fine with the 850 G2. And when the power does go fully out while using my PC, I simply turn it on again.

And sit'hari is right, we don't care about the holdup time, we're more concerned about the PWR_OK signal cutting off at a safe voltage. I honestly could care less if my holdup time was 3ms.

[Jon Gerow]

And if they can, then for how long and for how many times ??
Can someone take this responsibility? If noone can, then i believe that i have the right as a customer to be worried whenever i notice a Power_OK signal failure in a PSU, right?

You're talking about a 3 to 4ms discrepancy. That will have little to no effect on hardware in real world application.

Corsair, for example, will continue to make sure hold up time and power good signal remain in spec and not play these "games", but I've tested this with a scope and there are a lot worse things that a lot better PSUs can do to damage your hardware.

The only reason they did this in the first place is to not cheat the customers.... but to "fool" the reviewers. Problem is, they got caught.

[sith'ari]

You're talking about a 3 to 4ms discrepancy. That will have little to no effect on hardware in real world application......................

I see, but then the next logical question is after how many ms should we be worried about the tolerances of mobos components? (*regulators and mosfets IF i remember correctly Aris's words)
Because it's not always a given that the failure will be near 3-4 ms.
There has been a case when "the Power_OK signal" failure was 17ms !! , such as the case of Zalman ZM750-EBT . ( http://www.tomshardware.com/reviews/...su,4373-4.html )

[Jon Gerow]

I'll need to ask my board partners for that info.

[TELVM]

It's the 650 G2 that likely has the issue, not the 550 G2 (or any of the higher wattage ones). The 550 G2 review is here http://www.tomshardware.com/reviews/...ly,4244-4.html

12ms holdup time, they ain't cheating.

650 G2 review https://www.techpowerup.com/reviews/..._G2_650/5.html

Hold-up time easily surpassed the 16 ms mark the ATX specification sets. Please note that although the unit uses the exact same bulk cap as the 550 G2 that failed this test, it somehow managed to produce a very long hold-up time. Given the 550 G2 has 100 W less capacity, it should register an even lower hold-up time than its small brother. The assumption is that significant improvements that somehow affect the hold-up time were made, or that the bulk cap of our 550 G2 sample wasn't at its best.

From the pics in the reviews it looks like the 650 G2 uses 2700 uF caps on the secondary, while the 550 G2 uses smaller 2200 uF ones:

My guess is that the larger secondary capacitance in the 650 G2 has a positive influence on hold-up time.

[turkey3_scratch]

The 3-4ms is for 100% load, if the PC is idle it becomes way higher, probably 100ms+ (I never measured it).

[turkey3_scratch]

From the pics in the reviews it looks like the 650 G2 uses 2700 uF caps on the secondary, while the 550 G2 uses smaller 2200 uF ones:

-snip-

My guess is that the larger secondary capacitance in the 650 G2 has a positive influence on hold-up time.

I'm unsure how much of an effect this would have on holdup time. For one thing, the expression that equates the charge on each capacitor plate over time for a discharging capacitor is:



Where Qmax is max charge on a plate, t is time, R is resistance, C is capacitance. So, it's true that a higher capacitance will decrease the rate at which the charge of each plate drops. But this is also dependent on the resistance of the circuit, and I have no idea how that compares for the bulk capacitors to those on the secondary. A higher resistance will decrease the current so the plates will discharge more slowly.

But what we really need to be looking at is energy here, and the energy stored in a capacitor = 0.5*Q*V, so at the end of the day it's more important that we look to charge and voltage in addition to capacitance for energy. Capacitance is relevant in that it relates charge to voltage since Q=CV, but the important thing is that capacitance is not the only factor here.

Two capacitors can have identical capacitance but have great differences in energy storage. This is generally seen with bulk capacitors having a lower capacitance than smaller ones on the secondary, but both the charge and voltage are higher for the bulk capacitors.

In fact, looking at the EVGA 650 G2 review by Aris, the 650 G2 has a single 470uF 400V bulk capacitor, but the secondary capacitors are 16V 2700uF capacitors. Using math here, let's say we see there are about 10 of those capacitors on the secondary.

Energy = 0.5 * Q * V = CV^2

We see voltage (at max charge on each plate) dominates capacitance when it comes to energy. Combining the energy of all the secondary capacitors we get 6.92J of energy stored in those when the charge on their plates is maximized. With the single bulk capacitor, we get 75.2J of energy. Total energy would be about 82J.

This means that the single bulk capacitor stores almost 11X more energy than all those secondary capacitors combined. In fact, if we're looking at the fact that the 650G2 uses 2700uF instead of 2200uF, we find that there is about a 1.3J difference in energy, which is totally negligible and would never account for nearly double the holdup time, at most maybe about 10uS.

[sith'ari]

From the pics in the reviews it looks like the 650 G2 uses 2700 uF caps on the secondary, while the 550 G2 uses smaller 2200 uF ones:
My guess is that the larger secondary capacitance in the 650 G2 has a positive influence on hold-up time.

Wait a minute, i'm a little confused here:
Among the 3 units ( SF LeadexGold 550, EVGA G2550, EVGA G2650 ) we have the following facts:
1) They all use the same bulk-cap
2) Their "Hold-UP time" is :
2a)SF 550: 11ms ( http://www.tomshardware.com/reviews/...ly,4416-4.html )
2b)G2 550: 12,4ms ( http://www.tomshardware.com/reviews/...ly,4244-4.html )
2c)G2 650: 21ms ( https://www.techpowerup.com/reviews/..._G2_650/5.html )
Between the 2 550watt units and the 650watt one, -although they all use the same bulk-cap- , we have about 10ms difference at Hold-UP time, and even Aris couldn't explain the reason for such a deviation !!
I remember a comment that @turkey3_scratch has made in the past which says that

The old testing equipment did not test hold-up time, it instead tested AC_LOSS to PWR_OK, even though they were falsely called hold-up time tests ( http://www.jonnyguru.com/forums/show...2&postcount=80 )

IF this is accurate, then that means that the G2 650 is far more likely to have a "Power_OK signal" issue than the G2 550 because its "Power_OK" time is way too long, 21ms (*i say again, IF @turkey3_scratch's comment is correct) and, since the bulk-cap was considered as a low capacity one by Aris during his SF 550review, this means it would be highly unlikely to achieve a "Hold-UP" time that could overpass the 21ms of "Power_OK signal".

[turkey3_scratch]

My comment is correct, it's AC_LOSS to PWR_OK because look at what the graph for the 650 G2 is measuring:



The blue is the AC voltage, the yellow line is the voltage of the PWR_OK signal. If it was measuring AC_LOSS to DC_LOSS instead it's droop down like in the Superflower review.



In which now yellow is 12V voltage out instead of PWR_OK, and green is now PWR_OK.