PicoPSU systems shuts off unexpectedly
Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee
PicoPSU systems shuts off unexpectedly
Hey Everyone,
I've been having an issue that popped up last May that seems to be getting worse. Here's my System:
CPU: i5-2400
Mobo: GIGABYTE GA-H67N-USB3-B3
GPU: AMD 7750
PSU: PicoPSU 160W with Dell DA-2
Case: MI-100
It originally started with the Diablo 3 release, while playing the game the system would shut off unexpectedly then power back on. I attributed it to the GPU overheating and made some adjustments to improve airflow (Case cover was already removed due). I later replaced the GPU that was previously a 5670, issue persisted when gaming. I eventually spoke with a ME friend and he was certain it was the CPU getting too hot. I was using the Stock heatsink as I couldn't find one that fit the case I have. With his recommendation I went out and tracked down the Samuel 17 which I installed back in April.
After that it appeared the issue had been rectified, however I'm seeing occasional shutoffs for no apparent reason. In this instances I'm barely working the computer, the last one I was browsing the web while lightroom imported photos.
Anyone have advice on what's going on and how I might correct it? I've been looking into getting another case, one that's a little taller/wider so I can get proper airflow to the GPU/CPU with the cover on. Considering what I've already replaced I'm wondering if it's the PSU/Power brick.
I've been having an issue that popped up last May that seems to be getting worse. Here's my System:
CPU: i5-2400
Mobo: GIGABYTE GA-H67N-USB3-B3
GPU: AMD 7750
PSU: PicoPSU 160W with Dell DA-2
Case: MI-100
It originally started with the Diablo 3 release, while playing the game the system would shut off unexpectedly then power back on. I attributed it to the GPU overheating and made some adjustments to improve airflow (Case cover was already removed due). I later replaced the GPU that was previously a 5670, issue persisted when gaming. I eventually spoke with a ME friend and he was certain it was the CPU getting too hot. I was using the Stock heatsink as I couldn't find one that fit the case I have. With his recommendation I went out and tracked down the Samuel 17 which I installed back in April.
After that it appeared the issue had been rectified, however I'm seeing occasional shutoffs for no apparent reason. In this instances I'm barely working the computer, the last one I was browsing the web while lightroom imported photos.
Anyone have advice on what's going on and how I might correct it? I've been looking into getting another case, one that's a little taller/wider so I can get proper airflow to the GPU/CPU with the cover on. Considering what I've already replaced I'm wondering if it's the PSU/Power brick.
Re: PicoPSU systems shuts off unexpectedly
How about you run a memtest, some benchmarks and monitor the temperatures? That should help diagnose the problem.
Re: PicoPSU systems shuts off unexpectedly
Have you checked your 12V rail? If you have a multimeter, measure the voltages during idle/load on a molex connector.
I've had stability problems in the past with a similar system. The 12V rail was down to 11.4V and below. Fixed it by doing some cable work; got a better connector between the Dell DA-2 and the system and also had the brick feed the 4 pin 12V-cable directly, bypassing the picoPSU.
I've had stability problems in the past with a similar system. The 12V rail was down to 11.4V and below. Fixed it by doing some cable work; got a better connector between the Dell DA-2 and the system and also had the brick feed the 4 pin 12V-cable directly, bypassing the picoPSU.
Re: PicoPSU systems shuts off unexpectedly
I ran prime95 for 30 minutes and watched the temp, no lock ups or weird behaviors.Cistron wrote:How about you run a memtest, some benchmarks and monitor the temperatures? That should help diagnose the problem.
Great suggestion, I'll check the levels later today. I made an adapter for the DA-2 so I wouldn't have to hack it apart, it's entirely possible that's at fault. Here's a pic of it:Vicotnik wrote:Have you checked your 12V rail? If you have a multimeter, measure the voltages during idle/load on a molex connector.
I've had stability problems in the past with a similar system. The 12V rail was down to 11.4V and below. Fixed it by doing some cable work; got a better connector between the Dell DA-2 and the system and also had the brick feed the 4 pin 12V-cable directly, bypassing the picoPSU.
Re: PicoPSU systems shuts off unexpectedly
I measured the voltage coming from the DA2 and going to the 4-pin motherboard connector. Both measured at ~15.3V
Re: PicoPSU systems shuts off unexpectedly
15.3 volts? Is the system running? The tolerance according to spec is +/- 5% so 15.3v is a little high..
Re: PicoPSU systems shuts off unexpectedly
I thought it was a bit high too, that measurement was taken with the system running. When powered off I measured ~15.8V coming from the DA2. All measurements were done at the PicoPSU.
I'm wondering if it's GPU related, I know firefox uses hardware acceleration and the unexpected shutoffs tend to occur when I'm using that browser (but it's not consistent, I'll have to test).
I'm wondering if it's GPU related, I know firefox uses hardware acceleration and the unexpected shutoffs tend to occur when I'm using that browser (but it's not consistent, I'll have to test).
Re: PicoPSU systems shuts off unexpectedly
Probably not GPU related, I ran a GPU stress test for 1 1/2 hours without any issues, GPU got up to 62C without any sudden shutoffs.
-
- Posts: 1839
- Joined: Mon Feb 05, 2007 2:10 pm
- Location: Northern New Jersey
- Contact:
Re: PicoPSU systems shuts off unexpectedly
If you pulled the GPU, could you test your PICO with a different power brick?
I'm wondering if your DA-2 is bad. Voltage that high above what it should be, even going through the PICO could be a sign of a bigger problem, either related to the PICO or the DA-2. I would see if there's a way to cut off the adapter and DA-2, as long as the PICO operates normally without those two things, you've narrowed your issue down to either the DA-2 itself, or the adapter. If you cut one of the 12v lines, and just looped the remaining line to the other 12v wire, does the system become stable again?
Lots of questions..
I'm wondering if your DA-2 is bad. Voltage that high above what it should be, even going through the PICO could be a sign of a bigger problem, either related to the PICO or the DA-2. I would see if there's a way to cut off the adapter and DA-2, as long as the PICO operates normally without those two things, you've narrowed your issue down to either the DA-2 itself, or the adapter. If you cut one of the 12v lines, and just looped the remaining line to the other 12v wire, does the system become stable again?
Lots of questions..
-
- *Lifetime Patron*
- Posts: 2269
- Joined: Sun May 21, 2006 9:09 am
- Location: Northern California.
Re: PicoPSU systems shuts off unexpectedly
did you rule out bad ram by running memtest for at least 12 hours yet?
Re: PicoPSU systems shuts off unexpectedly
Memtest did not stress the ram nearly enought. Use Prime95 with the following settings for ram stress:
http://dfiles.eu/files/a6yj5ct5d
http://dfiles.eu/files/a6yj5ct5d
-
- *Lifetime Patron*
- Posts: 2269
- Joined: Sun May 21, 2006 9:09 am
- Location: Northern California.
Re: PicoPSU systems shuts off unexpectedly
it is entirely possible to pass prime95 with flying colors and still have memtest find a bad ram stick. 0_o
(its also possible that memtest wont find the offending stick of ram in a few hours, or even 12 ... i like to run it for 24 hours at least.)
(its also possible that memtest wont find the offending stick of ram in a few hours, or even 12 ... i like to run it for 24 hours at least.)
Re: PicoPSU systems shuts off unexpectedly
Maybe on multicore systems this can happen, but I tested several times memtest and find it to be very light in terms of stress for the rams. Anyone can test that. Just crank the overclock of rams to the point, where memtest start showing errors. Then back down a 2, 3Mhz and let memtest run overnight.
No error? Great.
Fire up Prime95, start with the Min FFT size of 1792 and you get error under 1 minute.
That test basicaly prove, that when you want to do quick test and win is crashing, then memtest is priceless. When you want to really know that your system IS stable and reliable, there is no alternative. Maybe the OCCT for more cores, but that thing require NETframework higher that v2.0, witch is not good IMHO. Just v2.0 SP1 refuse to install on my Win2k SP4, so screw this sowtware
No error? Great.
Fire up Prime95, start with the Min FFT size of 1792 and you get error under 1 minute.
That test basicaly prove, that when you want to do quick test and win is crashing, then memtest is priceless. When you want to really know that your system IS stable and reliable, there is no alternative. Maybe the OCCT for more cores, but that thing require NETframework higher that v2.0, witch is not good IMHO. Just v2.0 SP1 refuse to install on my Win2k SP4, so screw this sowtware