Question about the Enermax modu 82+ 425W

[Hush]

Were does it stand in in terms of noise and efficiency by today's standards?
I have one since 2009, the fan has gotten louder - although the rpm is constantly at 530- and took on an annoying sound signature,
and I suspect that it's performance is starting to dwindle.
I recently upgraded the components to Z97 and Haswell,
oc'ed the CPU and found a voltage were it ran stable under aida64 for 4 hours.
2 days later the system started BSOD'ing left and right, today i increased the vcore a couple of notches,
so far so good.
Could it be an issue with the PSU, that it's starting to falter after 6 years?
If i do replace it, would the Seasonic 400FL platinum be a good choice?

[Cistron]

How much did you overclock and is there anything else drawing loads of power in the system?

I have the small Pro82+ 385W and that's still going strong after 7 years. The fan is equally annoying so I swapped that.

[CA_Steve]

It's more likely you were on the edge for a stable CPU than the PSU altered it's output. Maybe it was a warmer day when it BSOD'ed..a few degree C's can make the difference between stable timing and not stable timing when you are overclocking/overvolting.

[Hush]

It's a G3258 chip, OC is 4.3ghz, now @ 1.229 vcore.
a few days ago it ran stable at 1.119.
nothing else is drawing loads of power, external video card is disabled.

what made me think that it may be the PSU is, that it one BSOD occurred when i inserted a dvd,
and the DVD-rom began spinning.

Temps were actually stable, maxing out at 55C under load,while averaging 46C. (during stress test)
idle temp was around 30-32.

[Hush]

BTW the noise coming from the Enermax fan sounds exactly like that chirping sound at night in the trees, in suburban areas...

[xan_user]

the code on the BSOD might help. or is there anything for that event in the Windows logs?

could also be failing/aged ram. id run memtest overnight, or longer just to be sure. (after all you are overclocking, which can stress parts beyond there safe operating specs)

[xan_user]

the code on the BSOD might help. or is there anything for that event in the Windows logs?

could also be failing/aged ram. id run memtest overnight, or longer just to be sure. (after all you just strarted overclocking, which can easily stress parts beyond there safe operating specs)

id also run it stock for a while and see if BSOD happens again..

[Hush]

WHEA_UNCORRECTABLE_ERROR
what should i look for in the logs?
RAM is also new. but i will run memtest.

[CA_Steve]

https://msdn.microsoft.com/en-us/librar ... 85%29.aspx

[xan_user]

new ram + overclocking= BSOD? ill put $20 on that being the issue.

[Stevo_]

I had the 625W version of this PSU in an 3770K build, RMA'd one, replacement was only slightly better but at least I could boot. Turned out these PSUs couldn't handle the low power states on the newer processors, regulation would go to shit, ASUS BIOS would report unstable voltages after coming out of BSOD. Turning up the vcore probably gets you just enough load to keep it in regulation.

[Hush]

Ram passed 12 hours of memtest with flying colors, meanwhile the PC is still crashing.
so i'll go ahead and buy a new PSU.
I am contemplating either the seasonic 400 fanless v2, or the SS G 550W
which is the better choice?
is 400W enough, or could it bottleneck the PC in the future?

[edh]

400W is more than enough for virtually all computers, but it obviously depends on your system.

Why G550 and not G360 or G450? The G series are not as quiet and maybe a fan swap could help but they're a good PSU anyway. If money is tight, definitely stick to the G series as the X series is so much more expensive.

[quest_for_silence]

Turned out these PSUs couldn't handle the low power states on the newer processors, regulation would go to shit

That's spreading FUD (or bullshit, if you rather): it's a quality PSU, and C6 & C7 low power states doesn't imply the WHEA BSOD (moreover they are useless on a desktop, so just disable them, as the vast majority of users out there).
Besides, right now I've a i5 4690 on a Z97 Plus with a GTX 970, all feeded by a 425 and that rig works flawlessly.

I am contemplating either the seasonic 400 fanless v2, or the SS G 550W
which is the better choice?

After three Gold & Platinum fanless Seasonics my opinion is they don't worth the money: if you wanna go fanless, pick a Super Flower.
The G-series up to maybe 200W DC may be enough quiet (and it's a solid performing unit) but it strongly depend of ambient temp: moreover the fan isn't a quiet one, so probably some "clone" might work quieter. Said that, there are quieter options around, like BeQuiet.

[Stevo_]

Turned out these PSUs couldn't handle the low power states on the newer processors, regulation would go to shit

That's spreading FUD (or bullshit, if you rather): it's a quality PSU, and C6 & C7 low power states doesn't imply the WHEA BSOD (moreover they are useless on a desktop, so just disable them, as the vast majority of users out there).
Besides, right now I've a i5 4690 on a Z97 Plus with a GTX 970, all feeded by a 425 and that rig works flawlessly.

There's a reason they discontinued them at that time, that was after conversing with their customer service after I received the RMA unit, so straight from the horses mouth. Enermax makes pretty good PSU's but by no means the best.

Edit: I'd add they weren't the only ones having issues with their older lines of PSUs at the time.

[quest_for_silence]

There's a reason they discontinued them at that time

Please don't take offense, but here it's not THG or any kids' playground, and that sort of "tech gossip" is not really acceptable.

I strongly disagree with your thoughts: that platform was discontinued just because it's an old design, no more competitive (and no more usefully tweakable: the newer and more modern resonant controllers with magnetic DC-DC converters are more performing and more economical to build).
Set aside those aspects, those units were built like tanks with specs up to about any contemporary job, and an expected lifespan of more than a decade (with no issue running an Haswell rig as in my own experience).

The referred OP issue is not plainly related to the PSU, when you're dealing with overclocking and/or undervolting, anything can happen: as a matter of fact one of my rigs with a Seasonic PSU, a Phenom II X6, run flawless for years undervolted to 1.13 Vcore, but since a while it get issues with less than 1.18 Vcore despite no changes in hardware.

Is it possible that the problem is related to the PSU? Yes, it is: but it's far from certain, as there's no evidence up to now it's a PSU issue, and your own issue with your 3770-rig also doesn't mean anything in this respect.

[Hush]

I do strongly suspect that it was the PSU.
my old rig started to break down and was crashing daily.
i assumed it was the e8400 that ran it's course,
but i now recall that some of the crashes then, were not even BSOD,
sometimes it just reset in an instant at random.

The G 550 is in the same price range as the platinum x400FL
so i was wondering if the G is silent enough, and whether the FL has enough juice.

Super flower is not available were i live. neither is Be quiet which i really wanted.
so i went with the SS 400FL.

Now running the 400 fanless i noticed immediately,
that temps all across went up, CPU by 5C, and MB by 6C.
also fan speeds turned up.
this is normal i assume, right?

[quest_for_silence]

so i went with the SS 400FL.

So now do you revert the vcore to 1.119V?

Now running the 400 fanless i noticed immediately,
that temps all across went up, CPU by 5C, and MB by 6C.
also fan speeds turned up.
this is normal i assume, right?

Well, it couldn't be so: actually an ATX PSU shouldn't have an exhaust role in the airflow scheme (and given the lower wasted heat, those noticeable temps risings are a bit suspicious to me).

[Hush]

I changed it to 1.212, so far it's held up.

you are sure about the PSU fan not serving as an exhaust?
i think that in my case it did.
with the old PSU when i placed the hand near the case fan,
i felt a strong stream of hot air flowing.
now i feel almost no flow at all, i opened the case to see if the fan was working.
apparently the old psu was pushing loads of air to the case fan.

[Hush]

Now it has already failed at the reduced vcore..
but at 1.229 it does seem entirely stable,
which wasn't the case before

i think what may have been the problem with the enermax,
was the compatibility of the fan system with z97 mobo.
with the p45 MB, when running intel burn test, the fan would ramp up to 900rpm,
otherwise it would stay at 470-500.
in this asus z97-a it would always show it at 530rpm.
unlike the p45 there isn't a designated pwr connector,
so i plugged the enermax into the chassis 4 connector.
so maybe it was excessive heat in the PSU that caused something?

another thing, what's the deal with these new sensors?
it has auxtin0,1,2,3,
auxtin0 and 1 show temps at either -128C or 128C
when i run cpu tests it shows 128C otherwise it's at -128C
this was the case with both PSU's
auxtin2 and 3 does seem normal,
and shows lower temps in the new seasonic.

[edh]

i think what may have been the problem with the enermax,
was the compatibility of the fan system with z97 mobo.
...
so i plugged the enermax into the chassis 4 connector.
so maybe it was excessive heat in the PSU that caused something?

Fans on PSUs are not run from external power. The Enermax has a cable for fan rpm sensing which you can connect to the motherboard but the motherboard has no control at all over the PSU fan. Changing the motherboard would make no difference.

The additional sensors will be nothing to do with the PSU.

I don't think your understanding of airflow in your system is quite right either.

[Stevo_]

There's a reason they discontinued them at that time

Please don't take offense, but here it's not THG or any kids' playground, and that sort of "tech gossip" is not really acceptable.

I strongly disagree with your thoughts: that platform was discontinued just because it's an old design, no more competitive (and no more usefully tweakable: the newer and more modern resonant controllers with magnetic DC-DC converters are more performing and more economical to build).
Set aside those aspects, those units were built like tanks with specs up to about any contemporary job, and an expected lifespan of more than a decade (with no issue running an Haswell rig as in my own experience).

The referred OP issue is not plainly related to the PSU, when you're dealing with overclocking and/or undervolting, anything can happen: as a matter of fact one of my rigs with a Seasonic PSU, a Phenom II X6, run flawless for years undervolted to 1.13 Vcore, but since a while it get issues with less than 1.18 Vcore despite no changes in hardware.

Is it possible that the problem is related to the PSU? Yes, it is: but it's far from certain, as there's no evidence up to now it's a PSU issue, and your own issue with your 3770-rig also doesn't mean anything in this respect.

As I said there was a reason they discontinued them, outdated. CS guy admitted they were not rated for new low power designs, even JohnnyGuru saw big transients on the 87s at low loads(barely stayed in tolerance) from tests around that time. After getting the RMA I also purchased a PSU tester, which would load it enough, but when shorting the pwr_on pin to ground no load, pretty much nothing or way out of spec voltages. The first PSU output absolutely nothing at no load condition, and why I RMA'd it. When I bought it they had already been discontinued as Ivy Bridge was coming online. If the OP wants to try some troubleshooting try shorting the PWR_ON pin to ground and measure the no load output on the PSU for the different outputs, that would help tell what's going on.

As one raises core voltage, leakage increases drastically as you go down in nm and thus system power which gets the system away from low load condition. In the 28nm HPM TSMC library I design in right now, leakage power can equal dynamic power if the silicon is at the fast corner at 1.05V/125C(top temperature the library is characterized to). For that library 1.05 is also an overvoltage corner, 1.0v being nominal max, but performance gains are notable also.

[Stevo_]

There's a reason they discontinued them at that time

The referred OP issue is not plainly related to the PSU, when you're dealing with overclocking and/or undervolting, anything can happen: as a matter of fact one of my rigs with a Seasonic PSU, a Phenom II X6, run flawless for years undervolted to 1.13 Vcore, but since a while it get issues with less than 1.18 Vcore despite no changes in hardware.

Is it possible that the problem is related to the PSU? Yes, it is: but it's far from certain, as there's no evidence up to now it's a PSU issue, and your own issue with your 3770-rig also doesn't mean anything in this respect.

Now you're being ridiculous comparing a different brand PSU which means nothing in this case as is claiming that there's plainly nothing PSU related due to OC/OV conditions when that has everything to do with PSU loading. But the OP had upgraded to a newer system while keeping an older PSU so it has everything to do with my 3770 last gen with an already outdated PSU.

[Stevo_]

It's a G3258 chip, OC is 4.3ghz, now @ 1.229 vcore.
a few days ago it ran stable at 1.119.
nothing else is drawing loads of power, external video card is disabled.

what made me think that it may be the PSU is, that it one BSOD occurred when i inserted a dvd,
and the DVD-rom began spinning.

Temps were actually stable, maxing out at 55C under load,while averaging 46C. (during stress test)
idle temp was around 30-32.

This is where my RMA'd unit failed as soon as I tried loading Win7 from an external DVD. The original never got past booting.

Can you perform a manual no-load check of the PSU?

https://www.youtube.com/watch?v=hMP_zqlgzDc

[quest_for_silence]

Now you're being ridiculous comparing a different brand PSU which means nothing in this case as is claiming that there's plainly nothing PSU related due to OC/OV conditions when that has everything to do with PSU loading. But the OP had upgraded to a newer system while keeping an older PSU so it has everything to do with my 3770 last gen with an already outdated PSU.

It's crystal clear you didn't understand what I wrote: despite that fact, I won't call you either an idiot or a ridicolous man, but I'll just rephrase the concept, given that it might be due to my not so good english.

When you're operating a rig outside its normal range of operation, anything can happen: for instance (as a matter of fact) I myself experienced a weird issue with an undervolted system which, after running for years at 1.13Vcore flawlessly, since a while it's simply refusing to do the same and actually needed a small but noticeable bump in Vcore. A side note is that system don't use an Enermax 82+ PSU, but the key argument is that nothing is granted when you're operating a rig outside its standard range of operation, so that you can't easily assume there's a PSU issue, even because the possible reasons are many, and often more likely than a real PSU issue (not to mention that PSU issues due to ageing are almost exclusively related to mantaining the full rated power at the rated temperature).

The other thing you apparently didn't understand is that even if we grant for the sake of argument that you experienced a PSU issue with an Enermax 82+ series and a 3770-equipped rig, that fact would not imply anything definitive in the OP case, because we're not dealing with a widespread issue, or a cheap PSU, or a faulty series: in the worst case scenario, your PSU can even be defective but that's all (as I already wrote some posts ago, I myself upgraded a system using an Enermax 82+ 425W with an Haswell CPU, as well as the OP, but I didn't get any issue in my case).
As I wrote with other words, that possible, unfortunate occurrence doesn't mean anything with reference to the referred OP issue, and, as a matter of fact, he's still unable to run the system with 1.119Vcore even with a very modern and high performing Seasonic Platinum.

I hope you will be a bit more patient next time, before bringing up ridicolous words into this thread.

[quest_for_silence]

As I said there was a reason they discontinued them, outdated. CS guy admitted they were not rated for new low power designs, even JohnnyGuru saw big transients on the 87s at low loads(barely stayed in tolerance) from tests around that time.

Set aside the 87+ were different PSUs (and which 87+ series? The Revolution? The Gold?), that's grossly misunderstanding and misrepresenting the facts (spreading FUD): in the old Enermax units there's nothing out of specs related to those "new low power designs" (while about those tests you're talking about, linking would be welcome: as a matter of fact, with reference to the more similar unit to the 82+ PSUs, Jeremy said: "the Modu87+ does what it does pretty well. It holds reasonably stable voltages, it is extremely efficient, it keeps ripple and noise to about half the ATX spec or lower, and it is very resistant to heat. That said, it is not the best performing unit I've ever seen. Except for the efficiency, there is room for improvement on just about everything here from voltage regulation to turn on spike suppression. With this in mind, I'll give the unit an 8.5 for performance.". Not too shabby, I dare to say).

As one raises core voltage, leakage increases drastically as you go down in nm and thus system power which gets the system away from low load condition. In the 28nm HPM TSMC library I design in right now, leakage power can equal dynamic power if the silicon is at the fast corner at 1.05V/125C(top temperature the library is characterized to). For that library 1.05 is also an overvoltage corner, 1.0v being nominal max, but performance gains are notable also.

So what?

[Hush]

i think what may have been the problem with the enermax,
was the compatibility of the fan system with z97 mobo.
...
so i plugged the enermax into the chassis 4 connector.
so maybe it was excessive heat in the PSU that caused something?

Fans on PSUs are not run from external power. The Enermax has a cable for fan rpm sensing which you can connect to the motherboard but the motherboard has no control at all over the PSU fan. Changing the motherboard would make no difference.

The additional sensors will be nothing to do with the PSU.

I don't think your understanding of airflow in your system is quite right either.

That was my understanding too, but the fact is the rpm was a static 530 with the new MB.

I am no expert in this, mere speculation, the psu fan sat from the top right near the rear exhaust fan,
blowing out hot air, so maybe that stream converged with and bolstered the flow from the CPU 120 glidedtream fan - which is aligned straight against the rear exhaust fan - thus helping carry the load of that too.
and if i am wrong on this, should i be worried that the seasonic is defective due to the rise in temps across the system?

[quest_for_silence]

That was my understanding too, but the fact is the rpm was a static 530 with the new MB.

That might be normal: power draw was different, heat dumped into the case too.

should i be worried that the seasonic is defective due to the rise in temps across the system?

Currently I don't think it's likely defective (though a handy PSU tester is a useful tool), but might we start from some more comprehensive, detailled specs (which is the case, how you mounted the PSU, which are the fans, how you run them, and so on)?

[Stevo_]

So what?

Obviously you have no clue about the sources of power draw in small geometry semiconductors. You talk about OC/OV conditions as if they are completely independent from power.

Regarding the Johnyyguru article it was his remarks on the overshoot that while in spec they were below average, but this was a better spec PSU also. This is the kind of thing that will drive BIOS nuts. Neither modu 82+ I had would even output correct no load voltages as pretty much all PSUs made in the last few years do.

[quest_for_silence]

Obviously you have no clue about the sources of power draw in small geometry semiconductors

Said differently: does that matter?

Regarding the Johnyyguru article it was his remarks on the overshoot that while in spec they were below average

I beg your pardon but (as maybe as above) you're misunderstanding and misrepresenting those data (and again you miss to link the actual page), and moreover those are on-off conditions, not working ones.

This is the kind of thing that will drive BIOS nuts. Neither modu 82+ I had would even output correct no load voltages as pretty much all PSUs made in the last few years do.

What? I'm sorry but what you say, the way you said, it seems just purposedly spreading FUD, nothing else, nothing more. But I might have you misunderstood, obviously.

By the way, we're going deeply off topic (the Seasonic should be almost surely not defective, as well as the Enermax would look like not defective), so (if you don't mind) let's give a rest.