i think you're limiting the capabilities of the product by forcing it into eco mode under heavy loads. what's the benefit of eco mode if you're stressing the unit? the eco modes are designed to reduce fan speed under LOW LOAD situations. so why should it run exactly the same when under heavy loads like powering a powerful GPU+CPU under extended BOINC loads? there seems to be some dissonance in how the product is designed to work and how you're trying to use it. there's nothing defective.
your toaster analogy isn't accurate for what's happening. its like buying a Corvette, and limiting it yourself to 60mph, then complaining that it's not fast enough for such an expensive car. doesn't really make sense. except here, if you removed the limitations of eco mode, it would actually work as designed and you wouldn't notice anything anyway unless you looked to see the fan spinning.
I have to disagree with you on this. ECO mode as per the manufactures own specification should only limit the fan run times during low load (<40%) all other times the fan should run normally. (It is speed controlled but the fan curves are identical ECO mode vs Non-ECO mode, ECO simply cuts off below 40%) Again I say my toaster analogy is correct. If I try to use this FEATURE of this product it will randomly cut off the power rendering it useless. Not sure how you can defend your position on this, but knock yourself out.
The tidy graphs suggesting intended ECO mode behavior support your position, Gandolphi. However they only contemplate steady-state conditions. As we know the problem of actual concern involve rapid changes, it may be fruitful to ponder how that differs.
The obvious difference to me is that the supply which has been running for a good while high in the "no fan needed yet" portion of the ECO region has appreciably hotter interior components just before the offending transient comes along. It also has a fan which is running slower (actually stopped) at that moment, than the non-ECO case.
So along comes a transient at the speed of electrical things. Presumably the fan control logic notices and immediately sends 12V to the fan. But the fan, being a mechanical thing, takes many milliseconds to spin up, even to the speed that 5V gives at steady state. Meanwhile in the non-ECO case the fan is already spinning at the speed 5V gives, and goes zooming up the speed range, giving more cooling at any given instant after the transient begins than in the ECO case.
In summary, the temperature condition of the critical components will be materially inferior for several seconds after a transient begins for the ECO case than for the non-ECO case. Possibly they may behave worse, or possibly protection provisions may simply kick in to avoid damage risk.
I do find in intriguing that at least one forum participant is giving you exactly the opposite advice regarding your ECO mode as the same participant gave me for the "magic pushbutton" on my Seasonic Focus supply, which is documented to provide a very similar intended behavior difference.
As to whether this is a bug or a feature, I'd score it as "ECO can be expected to, and has been observed to, give less capability in the face of extreme transients than does non-ECO mode".
Peace
Peter A. Stoll
former microprocessor design engineer
former semiconductor reliability person
Nicely put archae86. I agree that MAY be correct in what is causing the random trips, but I still believe there is a design fault since my system shouldn't be drawing any more than 700 watts at 100% load. Regardless, the EVGA PS would not work, the Seasonic does. (Not a fair comparison because the EVGA was 1000 watt vs the 1300 watt Seasonic, but I'm not gonna buy a 1000 watt Seasonic to find out either...) I think this horse is dead now....
EVGA's eco mode is totally dependent on temperature. not power "load". load is only estimated by the temperature of the components. I'd agree with Archae86's assessment of the possibility that you could be operating just below the temperature threshold for fan spin, and the high temp component not being able to handle a transient that quickly overwhelms the unit to trip OCP/OTP when it wouldnt have otherwise if you just had the fan spin all the time.
it's still a bit of cognitive dissonance to try to run with ECO mode on for high load scenarios, even if it *shouldn't* make a difference based on projected loads. is it really providing any less noise or dust to the system if the system should always be over the threshold for fan spin anyway? I'd say no, and you likely can't hear the PSU fan over other system noise I have 6x 1200rpm Noctua fans in my case, and I can hear the GPU coil whine from Einstein tasks louder than anything else LOL. I just can't see the logic in operating this way at all. just making life harder for yourself and perceiving faults that are based on your operation and not the product itself.
also consider the products you're using. not all EVGA power supplies are from the same ODM, and so not all are equal. this goes for all PSU manufacturers. your particular P5 1000W EVGA is of inferior quality than their older P2 series units, which were larger with beefier components. both platinum rated, both 1000W, different ODMs, different quality, same producer (EVGA). the change in ODM was likely based on costs vs specs you'll find this kind of thing across the PSU industry. I'm not saying that you are intentionally conveying that, but it could come across that way.
i think you're limiting the
)
i think you're limiting the capabilities of the product by forcing it into eco mode under heavy loads. what's the benefit of eco mode if you're stressing the unit? the eco modes are designed to reduce fan speed under LOW LOAD situations. so why should it run exactly the same when under heavy loads like powering a powerful GPU+CPU under extended BOINC loads? there seems to be some dissonance in how the product is designed to work and how you're trying to use it. there's nothing defective.
your toaster analogy isn't accurate for what's happening. its like buying a Corvette, and limiting it yourself to 60mph, then complaining that it's not fast enough for such an expensive car. doesn't really make sense. except here, if you removed the limitations of eco mode, it would actually work as designed and you wouldn't notice anything anyway unless you looked to see the fan spinning.
_________________________________________________________________________
I have to disagree with you
)
I have to disagree with you on this. ECO mode as per the manufactures own specification should only limit the fan run times during low load (<40%) all other times the fan should run normally. (It is speed controlled but the fan curves are identical ECO mode vs Non-ECO mode, ECO simply cuts off below 40%) Again I say my toaster analogy is correct. If I try to use this FEATURE of this product it will randomly cut off the power rendering it useless. Not sure how you can defend your position on this, but knock yourself out.
The tidy graphs suggesting
)
The tidy graphs suggesting intended ECO mode behavior support your position, Gandolphi. However they only contemplate steady-state conditions. As we know the problem of actual concern involve rapid changes, it may be fruitful to ponder how that differs.
The obvious difference to me is that the supply which has been running for a good while high in the "no fan needed yet" portion of the ECO region has appreciably hotter interior components just before the offending transient comes along. It also has a fan which is running slower (actually stopped) at that moment, than the non-ECO case.
So along comes a transient at the speed of electrical things. Presumably the fan control logic notices and immediately sends 12V to the fan. But the fan, being a mechanical thing, takes many milliseconds to spin up, even to the speed that 5V gives at steady state. Meanwhile in the non-ECO case the fan is already spinning at the speed 5V gives, and goes zooming up the speed range, giving more cooling at any given instant after the transient begins than in the ECO case.
In summary, the temperature condition of the critical components will be materially inferior for several seconds after a transient begins for the ECO case than for the non-ECO case. Possibly they may behave worse, or possibly protection provisions may simply kick in to avoid damage risk.
I do find in intriguing that at least one forum participant is giving you exactly the opposite advice regarding your ECO mode as the same participant gave me for the "magic pushbutton" on my Seasonic Focus supply, which is documented to provide a very similar intended behavior difference.
As to whether this is a bug or a feature, I'd score it as "ECO can be expected to, and has been observed to, give less capability in the face of extreme transients than does non-ECO mode".
Peace
Peter A. Stoll
former microprocessor design engineer
former semiconductor reliability person
Nicely put archae86. I agree
)
Nicely put archae86. I agree that MAY be correct in what is causing the random trips, but I still believe there is a design fault since my system shouldn't be drawing any more than 700 watts at 100% load. Regardless, the EVGA PS would not work, the Seasonic does. (Not a fair comparison because the EVGA was 1000 watt vs the 1300 watt Seasonic, but I'm not gonna buy a 1000 watt Seasonic to find out either...) I think this horse is dead now....
EVGA's eco mode is totally
)
EVGA's eco mode is totally dependent on temperature. not power "load". load is only estimated by the temperature of the components. I'd agree with Archae86's assessment of the possibility that you could be operating just below the temperature threshold for fan spin, and the high temp component not being able to handle a transient that quickly overwhelms the unit to trip OCP/OTP when it wouldnt have otherwise if you just had the fan spin all the time.
it's still a bit of cognitive dissonance to try to run with ECO mode on for high load scenarios, even if it *shouldn't* make a difference based on projected loads. is it really providing any less noise or dust to the system if the system should always be over the threshold for fan spin anyway? I'd say no, and you likely can't hear the PSU fan over other system noise I have 6x 1200rpm Noctua fans in my case, and I can hear the GPU coil whine from Einstein tasks louder than anything else LOL. I just can't see the logic in operating this way at all. just making life harder for yourself and perceiving faults that are based on your operation and not the product itself.
also consider the products you're using. not all EVGA power supplies are from the same ODM, and so not all are equal. this goes for all PSU manufacturers. your particular P5 1000W EVGA is of inferior quality than their older P2 series units, which were larger with beefier components. both platinum rated, both 1000W, different ODMs, different quality, same producer (EVGA). the change in ODM was likely based on costs vs specs you'll find this kind of thing across the PSU industry. I'm not saying that you are intentionally conveying that, but it could come across that way.
_________________________________________________________________________