Jump to content
Corsair Community

is too big a PSU a problem?


kris228

Recommended Posts

Well, you get less efficiency the further you get away from ~50% load

 

So for example: if you know you won't be using more than about 400 watts, then you'll want to get an AX760 instead if an AX860

 

An AX760 will power that computer perfectly fine, but if you're thinking of getting another video card, then you'll want to get an AX860 or higher

Link to comment
Share on other sites

Thanks! Yes, that's the rub... It's not what I now have but what I might have over the next few years... a second R9? an i7? Best guess of current use doing 3DMark11 is around 350W. Furmark and Heaven all seem to be in that ballpark. I think I see an 850 in my future... That should settle the problem for long term... Thanks again!
Link to comment
Share on other sites

Even a AX760 would probably power 2 r9's, but at full load you would be running it near 100%, which wouldn't really be good

 

Unless you have my video card which doesn't throttle and uses close to 500 watts when running OCCT with the voltage turned all the way up (actually went even higher by editing the bios)

 

Set voltage to 1.35 and clock speed to 1170 and my video card was using about 500 watts

 

Power supply said it was sending 42 amps to my video card

 

Well, so... the hotter a video card is the higher the power consumption... it can be quite a bit, when I'm stress testing it's almost 3 more watts per degree hotter! Under normal circumstances it's usually only about a watt more per degree, though

 

So it doesn't start at 500 watts, but that's what it got up to before the VRM temp hit 110 and it started throttling

 

 

 

Anyway, I would just go with an AX860... unless you think you'll want more than 2 video cards in the future

 

But it's up to you

 

 

 

This is me stress testing my video card, lol

 

[ame=

]
[/ame]

 

 

I have a 7970... TDP says it's supposed to be 223, but for some reason it doesn't throttle no matter what I set the power limit percentage to... The power in for the core caps at 370 watts and won't display any higher... lol

Link to comment
Share on other sites

hot board...

 

500W?????

 

What sensor monitor are you using?

 

I finally got around to trying to stress the R9 - furmark + heaven would crash after a few minutes but i think that is a program prob in heaven as furmark ran 100% for 30min + with no prob - It's fans and on board circuitry don't allow it to go beyond 80.0C - got up to about 1700rpm and about 44%. Finally ran prime95 and it ran fine though heat got up to 87 or so, but i have the BIOS set to Normal for the H80i V2 pump/2 fans. If I was going to do that type of thing on a regular basis I would set it to Fast , and temp would be a bit better. But wasn't able to break it. I did finally get an rj45 to usb cable for the UPS and it is up to maybe 350 or 375, but it also has the 40" hdtv which is about 50 watts of that(plus a myriad of modem, router,etc). Max for it is 600W (1000VA), so at some point I will have to take a hard look at it, but for now it is fine.

 

EDIT: I do not OC the R9. and just OC the cpu 3.5 to 3.8 ("turbo boost")

Link to comment
Share on other sites

hot board...

 

500W?????

 

What sensor monitor are you using?

 

HWiNFO

 

But... yeah... I saved my bios and opened it up in a program and this is what it showed... not sure why it's not throttling lol

 

http://i.imgur.com/7AX70k0.png

 

I know folding@home can prevent underclocking thus the video card can't underclock, but I didn't have it running at the time... but that's how I got the VRM temp up to 135 multiple times lol, my video card shut off probably 5 times due to VRM temp

 

[ame=

]
[/ame]

 

The VRM temp 2 warms up eventually too, it just does it slower... actually, if I leave VRM temp 1 at the same temp for a while then eventually VRM temp 2 will be 1-2 degrees hotter than VRM temp 1

 

I currently have a case fan attached to my GPU for cooling... apparently the fans decided they kept wanting to turn down to a really low rpm and the temp kept hitting 104 and shutting the video card off... :/

 

In that video my stock fans were working, though

 

heaven would crash after a few minutes but i think that is a program prob in heaven as furmark ran 100% for 30min + with no prob

 

That's because furmark isn't good for testing stability lol, neither is Kombustor or OCCT

 

For one, they require way more voltage to run than any other game or program would, and two because they don't extensively test the GPU... they just heat it up a lot

 

Heaven is actually an AWESOME program for testing stability... it puts more of a real-world load on the video card, and tests it extensively unlike those other stress test programs which... they just aren't very good at all for testing stability, lol

Link to comment
Share on other sites

I have had to back off everything - realized I'm getting a 0005 error in d3d11.dll when the display stops and ultimately Heaven crashes. Trying to figure that out. Win7. dx11. arg!

 

I'm no expert but I think it might be a hardware error... maybe try turning overclock down?

 

"Error 0005

Exception occurs if parameter is less than a specific value.

You will receive this error if the parameter in the message is below or equal to a boundary value required for the module to process the data."

 

 

Uh... nevermind.. that wasn't even for DirectX lol... not sure what it means :/

 

 

Okay... I think it's a driver issue... first step would be to run DDU and uninstall the drivers completely and then reinstall them... and if that doesn't work then it's probably a hardware issue... maybe overclock is too high or the hardware is damaged

Link to comment
Share on other sites

I'm no expert but I think it might be a hardware error... maybe try turning overclock down?

 

"Error 0005

Exception occurs if parameter is less than a specific value.

You will receive this error if the parameter in the message is below or equal to a boundary value required for the module to process the data."

 

 

Uh... nevermind.. that wasn't even for DirectX lol... not sure what it means :/

 

 

Okay... I think it's a driver issue... first step would be to run DDU and uninstall the drivers completely and then reinstall them... and if that doesn't work then it's probably a hardware issue... maybe overclock is too high or the hardware is damaged

 

I got the first part of that message on my phone and was wondering at that point - - yes, I thought about SSD, memory, mercury retrograde, etc - I was thinking AMD driver but I couldn't uninstall AMD, and windows update wouldn't work, and I tried to fix that and that wouldn't work, (I did scan for viruses at this point but none.) so I tried restoring an easeus image and that stalled, so by then, linux was gone and everything else was gone and so I offline booted a thumbdrive and restored C from 10 days ago but had to kick the MBR to get it to boot, then restored images for other 3 partitions, sans Linux (I'll get that later if need be)(clonezilla!) So i'm 10 days previous and heaven runs like a top, but turbo is off and, yes, AMD driver is older (yay). And oh yeah I put the H80i on FAST. So here is HWInfo from 2 runs of Heaven - not taxing it much. I will put turbo boost back on and run heaven for a while and go to bed and look at the rest tomorrow.

:D: :hihi:

501077333_hwinfoheaven2x.thumb.jpg.7786e2188892edae106230e8f463ec3d.jpg

Link to comment
Share on other sites

Great! Btw, I recommend you disable the VRM temp sensors for the VRAM... all it does is flip the order from the GPU core

 

It also appears to be reporting the memory voltage wrong... if it was reporting correctly it should be about 1.5 volts, also ram shouldn't downvolt at idle... so it's definitely reporting wrong

 

Also I believe the wattage for the RAM is off... should be more like 20-50... idk, maybe the ram on the 380 is more power efficient

 

But I guess those things aren't that important

 

For me at 1500mhz with folding@home it was about 30, OCCT was 33 watts, Saints Row was 35, Furmark was going from 30-45, Heaven was 30-50, Kombustor was 50-52

 

This is running folding@home

 

http://i.imgur.com/CetiLi7.png

 

I currently have one case fan mounted to the video card and one just blowing on it, both at a low rpm... VRM temps are high but it keeps the core temps decent at least

 

Plus, my video card will start throttling if the VRM temp hits 110... so I shouldn't really have to worry about it too much

 

I was thinking AMD driver but I couldn't uninstall AMD

 

Did you try DDU? I tried using the AMD uninstaller and it didn't do !@#$, said it uninstalled but left the current driver installed

Link to comment
Share on other sites

My vote is for display driver but will prove that later...

 

I did a quick furmark and got this for UPS - I don't get anything for PSU... the tv is also running on that.

 

I do not get the extreme power use that you get, and, yes the R9 may be more efficient. I chose it because all the R7s I was looking at were purported to be noisy and old architecture. This is silent Just wanted to try an AMD as I had been on Nvidia forever and just blew a 650gtxTi.

 

I did a brief prime95 run last night with BIOS set to FULL for the H80i and it seemed 10 deg better than what I got before with it on Normal, - but all these tests would asymptote way high if run over time as the case internal would rise and esp in summer the ambient would go up.

 

But I tend to get something new and hammer it briefly to see if it breaks(during 30 day return for MC), and then use it normally at way lower loads.

 

As I am new again to AMD I will need to remember to hammer it with heaven and furmark esp after any driver update, but that's minor. Will also remember to do a full image backup of spindle (5 partitions)(SSD) before upgrading, which I will do tonight. Also must remember that a display driver from ASUS is different from a driver from AMD! LOL I knew about DDU long ago but had forgotten its name as I was scrambling last night. Thanks. Now have it sitting in wait.

 

If HWINFO is reporting R9 mem voltage or power wrong, I would have no idea. Sometime I'll check it against "tweak II". Heaven reports the gpu temp as some crazy 5 or 6digit number.

 

I currently have a 2 fan radiator sandwich on the back blowing in for the H80i, 2 blowing in on the front, and one blowing out on the top. There is slight neg pressure in the case. the R9 has 2 fans controlled by the R9 circuitry and either 0, 1, or 2 are running. mobo BIOS has nothing to do with it. Seems to work well though I did see one of the reports saying something hit 103.

 

Caveat - I know/assume that if I run prime95 and/or heaven/furmark for , say, 24 hours, I will break something and prove nothing.

 

Therefore I don't! :D:

 

You way OC your memory (great job by the by!) so I would expect your voltages and current draw to be higher. I have never figured the gain from that would be more than 1% so after trying it in years past I abandoned it . For me an increase would probably mean an i7 or a second R9. Time will tell.

 

For the immediate time and future, I LOVE that H80i and any increase would probably involve a larger case (750 or whatever it is) and a larger cooler like a 110 or something - but that is toy stuff for futures.

364793519_hwinfobrieffurmark.jpg.79250bf4a868b5b06d381606d27ed368.jpg

Link to comment
Share on other sites

I do not get the extreme power use that you get, and, yes the R9 may be more efficient. I chose it because all the R7s I was looking at were purported to be noisy and old architecture. This is silent Just wanted to try an AMD as I had been on Nvidia forever and just blew a 650gtxTi.

 

Yeah, Idk why mine doesn't throttle... although one video card causing PCIe #1 (the power draw that the PSU was reporting for the video card) to go over the OCP limit is not good, lol

 

You can go a little over the OCP limit without the PSU shutting off as long as it's not for more then about 30 seconds or so... but doesn't mean it's good

 

Well for the power supply itself it's probably okay... but for the video card? yeah, no

 

I currently have it set to 40 amps which would be about 480 watts

 

I currently have a 2 fan radiator sandwich on the back blowing in for the H80i, 2 blowing in on the front, and one blowing out on the top. There is slight neg pressure in the case. the R9 has 2 fans controlled by the R9 circuitry and either 0, 1, or 2 are running. mobo BIOS has nothing to do with it. Seems to work well though I did see one of the reports saying something hit 103.

 

Probably the processor... prime 95 will heat it up a lot, 104 is when some i7s throttle... so that's probably why it was showing 103

 

This is me running prime 95 with 1.55 volts... practically instantly hit the throttling temp

 

[ame=

]
[/ame]

 

You way OC your memory (great job by the by!) so I would expect your voltages and current draw to be higher. I have never figured the gain from that would be more than 1% so after trying it in years past I abandoned it . For me an increase would probably mean an i7 or a second R9. Time will tell.

 

Not by much... stock is 1375 and I run it at 1500, it can go up to 1620 without an increase in voltage

 

You can usually raise the stock clocks quite a bit with stock voltage... stock voltage is 1.175 at 920mhz, I can do 1.175 at 1050mhz

 

Might as well either turn the clock speed up or turn the voltage down... otherwise it's just wasted heat and power

 

I can do 1.05 volts at 920mhz... 100% stable to at least about 90°C

 

Turned the voltage down about 12%, which would be about 25% less heat and still have the same stock clock :)

 

I was running 3.7ghz at 0.968 volts on my processor for a while... stock voltage was 1.125 at 3.5ghz

 

They almost always give way more voltage than you need for the default clock speed

Link to comment
Share on other sites

yeah they do tend to run voltage a bit high - called "stability" LOL.

 

I made a partition image and updated the display driver and it failed right on queue. Found a "hotfix" but that failed as well, so just restored the image and will forget it. The Asus drivers seem to work fine for this Asus board. I see maybe one person on google complaining about it. Maybe it will be fixed in the future but don't care as I don't need the features. Yes, it might be fixed by getting another mfr's board but then again, it might not. Asus drivers work, AMD's do not(as far as this 1 failure). This does not bother me particularly as I have run into it before. No matter which video card you have you're always bitching about the drivers! :p:

 

The 103 was actually from a short furmark run and seen on the gpu. If I run it longer i could toast marshmallows! With prime, again short - several minutes - temps were only 77 or so(4 cpu cores).

 

In summary, basically the thing seems to be solid. Maybe I'll abandon Linux for now and look at folding...

Link to comment
Share on other sites

The 103 was actually from a short furmark run and seen on the gpu. If I run it longer i could toast marshmallows! With prime, again short - several minutes - temps were only 77 or so(4 cpu cores)

 

Wow, your GPU throttles really high

 

I guess my 7850 throttled and forced fans to 100% at 102, and my 7970 forced fans to 100% at 97 and if it was at 97 or above for more then like 10 seconds then it would start throttling and my 7850 shut off at 106 and my 7970 shut off at 104

 

Not sure why they set the shut-off temp for the 7850 higher

 

And on my 7970 it starts throttling if VRM temp hits 110 and shuts off if it hits 135

 

If I run it longer i could toast marshmallows!

 

True dat

Link to comment
Share on other sites

they run the r9 hot. don't know if that's good at all...

 

I think there may be a GTX 960 in my immediate future Only question is evga or gigabyte. not liking that halt in heaven. not at all...

see what sleep does?????

 

:D:

 

got 12,352 folding points last night. At least all this computing power id doing something functional

Link to comment
Share on other sites

got 12,352 folding points last night. At least all this computing power id doing something functional

 

Only about 12,000?

 

You entered a passkey for bonus points, right?

 

I get about 140,000 PPD with just the video card WU, and my video card is only a 7970

 

If you didn't enter a passkey then 12,000 in ~8 hours for a r9 sounds about right

 

Also, the r9 is made to run hot, so it's fine

 

Of course running it at that would shorten the life span, but even at 103 it would most likely still last months

Link to comment
Share on other sites

Wow, your GPU throttles really high

 

I guess my 7850 throttled and forced fans to 100% at 102, and my 7970 forced fans to 100% at 97 and if it was at 97 or above for more then like 10 seconds then it would start throttling and my 7850 shut off at 106 and my 7970 shut off at 104

 

Not sure why they set the shut-off temp for the 7850 higher

 

And on my 7970 it starts throttling if VRM temp hits 110 and shuts off if it hits 135

 

 

When you say it throttles, what are you looking at?

 

gtx960 seems to run cooler (see pics) draws less power, too. seems to run at about the same framerate. and doesn't crash on the latest driver from Nvidia (364.72)!!! :D:

1009084516_hwinfo960heaven.jpg.ffda47d11eab88792f1672d8f7ce1f18.jpg

1064017861_hwinfo960furmark.jpg.3f8c65555f8e59e9b8370c680097a543.jpg

Link to comment
Share on other sites

When you say it throttles, what are you looking at?

 

The core goes from whatever clock speed it's set to to 500mhz

 

It reduces the power consumption and thus reduces heat

 

Usually the GPU will also force fans to 100%, you'll hear the fans spin up for a split second then go back down then turn back up for a split second and etc. etc.

 

Just enough to keep the GPU right at 103, or whatever temp it does it at

 

 

Your GPU is probably throttling while running Furmark, I can see that it reached 100% of the TDP

 

Mine doesn't throttle so I could use up to about 700 watts if I wanted to lol

 

If I set my video card to 1.45 volts with 1300mhz core clock while running OCCT then yeah, it would use over 700 watts

 

Not sure my video card can even do 1300mhz core clock, no matter how much voltage you give it...

 

 

Just realized you were running your 960... heh

 

Well, those probably throttle at about the same temp... but the shut-off temp is usually excessively high, like at 120 or something, I think

 

 

 

Yeaaaaaah, nvm.. I think it varies from video card to video card

Link to comment
Share on other sites

Must not have VRM temp sensors... that's dumb

 

I know for sure that it has a VRM

 

Surely they wouldn't be stupid enough not to include a VRM temp sensor, though...

 

I got mine up to 135 EASY and would of gone up to 160-170 if it didn't shut off my video card... not having a temp sensor is an easy way to blow the VRMs

Link to comment
Share on other sites

Don't you imagine that they use it (or something) to control fan speed and simply use fans to not allow it to go above a certain temp? That's what the R9 seemed to do, and that's what this one seems to do.

 

My fan speed doesn't change with VRM temp, it only throttles the video card

 

Right now I no longer have fans connected to my video card, though...

 

Apparently the case fan wouldn't plug into the video card fan slot... the slot was too small

 

Actually, I think it would of fit if the case fan didn't have a big chunk of plastic on the connector... it's almost like they didn't want me plugging the case fan into the video card :(

 

 

 

Actually.... I don't think it would of worked.. I think the fan controller is broken on the video card... with the stock fans they kept going back down to a really low rpm, and if I stopped the fan for a split second then they would go back up but then they'd go back down after a few seconds...

 

Took me a bit to figure out the trick while recording but near the end I figured it out, lol

 

[ame=

]
[/ame]

 

Yeah, dunno

 

 

I guess it was the fans themselves that were messed up, not where they were plugged into the video card

 

It was like this from the start... only one fan would run unless the fan speed was 80% or higher... but then one day it just started turning all the way down... guess it finally decided to break, lol

 

 

 

http://i.imgur.com/mRY2XwF.jpg

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

×
×
  • Create New...