View unanswered posts    View active topics

All times are UTC - 6 hours





Post new topic Reply to topic  [ 19 posts ] 
Go to page 1, 2  Next

Print view Previous topic   Next topic  
Author Message
Search for:
 Post subject: Picture gone crazy
PostPosted: Tue May 03, 2011 6:46 am 
Offline
Joined: Mon Aug 14, 2006 8:24 pm
Posts: 320
Location: Melbourne, Australia
While quietly watching a recording tonight the picture froze but sound continued. I restarted the frontend and the picture is permanently unstable with squares and other symbols flashing on and off no matter what the picture. I then rebooted and the frontend end would not start and remained black. Even the bios flash screen at boot was unstable which makes me think the graphics card had failed.

I was able to ssh into the box and checked some of the logs. The Xorg.0.log and /var/log/mythtv/mythfrontend.log showed nothing peculiar but the /var/log/errors.log showed lots of errors saying "Corrupted low memory". I tried rebooting to another OS via CD and again the image is unstable. I rebooted to gparted disk and ran MEMtest for a short while. It was difficult to work out what was happening because of the image instability but no errors were identified. I then rebooted to LinHES but even the backend would not start. Trying to boot again the reset button would not reboot but I managed to shut it down by holding down the power button. When I restarted it was all back to normal and the errors were gone.

I have had spurious freezing of the frontend before. Is my graphics card on the way out? Could it be something else?

I know I need to give memtest a good run but it is difficult to find a 48 h period with no machine usage. Anyway, I thought I would have both backend and frontend problems if it was memory.

_________________
Intel DG965WH, Dvico DVB-T Lite x2, HDHR, Gigabyte GT220, KingstonSSD, WD20EARS version=latest


Top
 Profile  
 
 Post subject:
PostPosted: Tue May 03, 2011 9:49 am 
Offline
Joined: Wed Feb 08, 2006 6:13 pm
Posts: 480
Location: IN
First check the fan on your video card (if you have one) and make sure it's still doing it's job. If you have a spare video card I would give it a shot. You problems definitely apears video card related but a system memory problem is certainly still well within the realm of possibility.

I'm leaning toward video card though as system memory shouldn't affect the BIOS screen.

_________________
ABIT NF-M2 nView | Athlon 64 X2 3800+ | 2GB DDR2 800 | HDHomerun | GeForce 6150 (onboard) | WD 640 GB SATA HD | DVD-RW (sata) | StreamZap IR receiver with Logitech Harmony remote

Vizio 37" LCD HDTV (1080p)


Top
 Profile  
 
 Post subject:
PostPosted: Tue May 03, 2011 5:03 pm 
Offline
Joined: Mon Aug 14, 2006 8:24 pm
Posts: 320
Location: Melbourne, Australia
Thanks for that. The video card is fan less and no I don't have a spare. I think I will just have to bite the bullet and buy one. Fortunately the machine is working again and it appears a total powerdown fixes the problem so I have time to do some research on what card to buy.

_________________
Intel DG965WH, Dvico DVB-T Lite x2, HDHR, Gigabyte GT220, KingstonSSD, WD20EARS version=latest


Top
 Profile  
 
PostPosted: Wed May 04, 2011 12:18 am 
Offline
Joined: Mon Apr 23, 2007 1:45 pm
Posts: 405
Location: Fargo, ND, USA
nicom
I see you have two pci-e capture boards and a fanless video card. I had something like this about 3 years ago and it was heat related. I would ether get frozen screens or black screens and the logs never gave any useful information to work with. I opened the case and placed a 80mm fan over the video card because all the logical information pointed to that. The problem went away until I closed the case up (with the added fan sill in the case) so instead of running a half hour before problems it now run 2 to 3 hours. The next step was to swap video cards with my desk top and it also run 2 to 3 hours before problems set in. What I finely traced it down to was a heat damaged capture card closest to the video card. I pulled my two capture cards and put the old video card back in and every thing was fine. Most of these fanless video cards will run fine at 80 deg C but the capture cards top out at about 30 deg C and I think the higher temps did some damage to the lower temperature rated capacitors on the capture card. I did have three pci slots on the motherboard and the cards where in 2 and 3 but that one slot wasn't enough distance to protect the capture card. My final solution was to swap positions of the capture cards and place a permanent 120mm 800 rpm fan over top of the three cards. I am still running my same capture cards today in a faster box but they are still fan cooled. I am not a believer anymore in the fanless video card concept. I have since added fans or ducting to provide forced air movement to all my fanless cards.

Good Luck
TVBox

_________________
TVBox
LinHES R8.6.1


Top
 Profile  
 
 Post subject: Re: Fanless video cards
PostPosted: Wed May 04, 2011 8:51 pm 
Offline
Joined: Thu Mar 25, 2004 11:00 am
Posts: 9551
Location: Arlington, MA
I think the problems people get into in this area come from forgetting that machines need to breathe and failing to monitor the thermals for the system before there is an issue.

Of my 3 currently active machines, all of them have passively cooled video cards, and all of them idle in the 30-33C range. The new media box has a fanless GeForce 210 (GT218 GPU) with a half height heat sink, and the tweaked RRD monitoring graphs tell me that it's currently idling at 32C and under max load has peaked out at 43C.

The key is that while the card is half height, the case isn't, and rather than being passive it'd be more accurate to say that the cards are indirectly cooled. The PS and optical drives are in a separate bay, and there are two 120mm exhaust fans for the relatively small remaining volume. Even with both fans at low speed they're still pulling roughly 60CFM out of the case, which given the compartments volume of roughly 0.6 cubic feet, means ~100 complete air changes per minute. As a result the video card gets plenty of airflow.

Frankly the case has so much air flow, and such good air path geometry, that you could turn off the fan on the CPU heat sink and run _it_ in passive mode too. So far the 43C that the video card recorded is the highest temperature for any component in the box, which gives us a 25C rise over ambient. This summer when the ambient in the house starts hitting 40C I'll probably turn up the fans to medium which will nearly double the airflow and should keep the max load temperature under 60C.

BTW - Any component that can't take 30 or 40C probably doesn't belong in your PC or just about any other type of consumer electronics. Very few places outside of high latitude or high altitude don't see some summer days in the 80s and 90s F (~27-37C). And that's ambient so add 10-20F (5-10C) at the very least for the temperature inside any kind of case.

TVBox wrote:
I am not a believer anymore in the fanless video card concept. I have since added fans or ducting to provide forced air movement to all my fanless cards.


Top
 Profile  
 
 Post subject:
PostPosted: Thu May 05, 2011 4:46 am 
Offline
Joined: Mon Aug 14, 2006 8:24 pm
Posts: 320
Location: Melbourne, Australia
I made my box very spacious as well to limit heat build up so I hope you are right tjc.

I will keep an eye on the temperatures in rrd. How do I get it to record GPU and tuner card temps? I looked in the forum but all I could find is modifying python scripts which don't exist on my machine.

_________________
Intel DG965WH, Dvico DVB-T Lite x2, HDHR, Gigabyte GT220, KingstonSSD, WD20EARS version=latest


Top
 Profile  
 
 Post subject:
PostPosted: Thu May 05, 2011 10:15 pm 
Offline
Joined: Thu Mar 25, 2004 11:00 am
Posts: 9551
Location: Arlington, MA
nicom wrote:
How do I get it to record GPU and tuner card temps? I looked in the forum but all I could find is modifying python scripts which don't exist on my machine.

The script below demonstrates the basic utilities I used to customize the rrd_mbtemp.pl (perl) scripts. If you don't have that you can install it using pacman under 6.04 like this:
Code:
pacman -S rrd_stats

You only need lm_sensors if mbmon doesn't work for your board. The nvidia-utils package provides nvidia-smi.

Here is my test script and it's output:
Code:
[root@black3 ~]# cat temps.sh
#!/bin/bash

echo "========================================================================"
echo "Motherboard fan and temperature sensors"
echo
sensors | egrep 'temp|fan'

echo "========================================================================"
echo "Video card temperature sensor"
echo
/usr/bin/nvidia-smi -g 0 | egrep 'Product|Temperature'

echo "========================================================================"
echo "Hard drive temperature sensors"
echo
hddtemp /dev/sda
hddtemp /dev/sdb

echo "========================================================================"
[root@black3 ~]# ./temps.sh
========================================================================
Motherboard fan and temperature sensors

fan1:        936 RPM  (min =    0 RPM)
fan2:          0 RPM  (min =    0 RPM)
fan3:          0 RPM  (min =    0 RPM)
temp1:       +30.0°C  (low  = +127.0°C, high = +127.0°C)  sensor = thermistor
temp2:       +23.0°C  (low  = +127.0°C, high = +127.0°C)  sensor = thermal diode
temp3:       +64.0°C  (low  = +127.0°C, high = +127.0°C)  sensor = thermistor
========================================================================
Video card temperature sensor

        Product Name            : GeForce 210
        Temperature             : 30 C
========================================================================
Hard drive temperature sensors

/dev/sda: WDC WD6400AAKS-65Z7B0: 30°C
/dev/sdb: WDC WD10EADS-65P6B0: 28°C
========================================================================

Note that temp3 is bogus and the reported number appears to be offset by ~30C from the real value. I've verify this a couple different ways now.

I don't know of any programmatic way to get the temperatures for the tuner cards.


Top
 Profile  
 
 Post subject:
PostPosted: Fri May 06, 2011 7:30 pm 
Offline
Joined: Mon Aug 14, 2006 8:24 pm
Posts: 320
Location: Melbourne, Australia
Thanks for that. I have not tried your script in total but I did try the commands on command line. The hard drive works but not the lm_sensors because for me output from 'sensors' is different and minimal. This is all I get.
Code:
[mythtv@Silverstone ~]$ sensors
coretemp-isa-0000
Adapter: ISA adapter
Core 0:      +44.0°C  (high = +78.0°C, crit = +100.0°C) 

coretemp-isa-0001
Adapter: ISA adapter
Core 1:      +43.0°C  (high = +78.0°C, crit = +100.0°C) 
Is there a way I can modify the configuration of sensors? The man page does not help much.

Ultimately I want to be able to graph the GPU temp to look for trends especially when a freeze occurs. I have enabled rrdtool and monitorix but am struggling to find how to add GPUCoreTemp to them.

_________________
Intel DG965WH, Dvico DVB-T Lite x2, HDHR, Gigabyte GT220, KingstonSSD, WD20EARS version=latest


Top
 Profile  
 
 Post subject:
PostPosted: Fri May 06, 2011 8:25 pm 
Offline
Joined: Thu Mar 25, 2004 11:00 am
Posts: 9551
Location: Arlington, MA
Here's the hack I made. Your MB and CPU stuff will be different, but if you have a modern Nvidia GPU that should work as shown.
Code:
[root@black3 rrd]# diff rrd_mbtemp.pl.old rrd_mbtemp.pl
20,22c20,23
< &ProcessHDD("mbtemp", "T 1", "Motherboard");
< &ProcessHDD("cputemp", "T 2", "CPU");
< &ProcessHDD("ambtemp", "T 3", "Case ambient");
---
> &ProcessHDD("mbtemp", "temp1:", "Motherboard");
> &ProcessHDD("cputemp", "temp2:", "CPU");
> # &ProcessHDD("ambtemp", "T 3", "Case ambient");
> &ProcessHDD("gputemp", "Temperature", "GPU");
32c33,39
<         my $temp=`/usr/bin/mbmon -c 1 -$_[1]`;
---
>         # my $temp=`/usr/bin/mbmon -c 1 -$_[1]`;
>         my $temp;
>         if ($_[0] =~ "gputemp") {
>                 $temp=`/usr/bin/nvidia-smi -g 0 | grep $_[1] | cut -d' ' -f2`;
>         } else {
>                 $temp=`/usr/bin/sensors | grep $_[1] | cut -c'15-16'`;
>         }


Have you installed the nvidia utils package mentioned above and tried the nvidia-smi command?

Oh and have you run sensors-detect to let it do the hardware probing?


Top
 Profile  
 
 Post subject:
PostPosted: Sat May 07, 2011 2:06 am 
Offline
Joined: Mon Aug 14, 2006 8:24 pm
Posts: 320
Location: Melbourne, Australia
Maybe my problem is the age of my card. My nvidia driver is 96.43.16 and I was of the opinion that was appropriate for my 7300GS card. If I try to install nvidia-utils it says it conflicts with nvidia-96xx-utils and asks if I want to delete it. I aborted at that point. So I do not have nvidia-smi.

Yes I have run sensor-detect but the only one it finds is
Code:
Just press ENTER to continue:

Driver `coretemp':
  * Chip `Intel Core family thermal sensor' (confidence: 9)

I should mention I got scared and said no to probing the I2C/SMBus. Is that important?

_________________
Intel DG965WH, Dvico DVB-T Lite x2, HDHR, Gigabyte GT220, KingstonSSD, WD20EARS version=latest


Top
 Profile  
 
 Post subject:
PostPosted: Sat May 07, 2011 8:28 am 
Offline
Joined: Thu Mar 25, 2004 11:00 am
Posts: 9551
Location: Arlington, MA
Worth a try, I've never seen it do anything bad. Just make sure that the system isn't in the middle of anything important and worst comes to worst you can power cycle it.

I'm surprised that that chipset isn't using a a more recent driver. My workstation box had a GT6600 in it for a long time and that worked great with the most current drivers until I replaced it a few months back. According to Nvidia you should be able to use right up to the latest drivers with that card. http://www.nvidia.com/object/linux-display-ia32-270.41.06-driver.html


Top
 Profile  
 
 Post subject:
PostPosted: Sun May 08, 2011 9:04 pm 
Offline
Joined: Mon Aug 14, 2006 8:24 pm
Posts: 320
Location: Melbourne, Australia
I updated the nvidia drivers but am still not having any luck with rrd data and even sensors is not behaving properly. I think I will try and solve that separately and start a new thread if necessary.

Getting back to the instability problem, I ran nvidia-settings and went to the temperature monitoring page and noticed the temp was about 65-70°C depending on what the machine was doing. On a static screen like the menu, it was closer to 70 but would drop when watching TV. I assume this is do do the CPU fan speeding up a bit and increasing the air flow.

I then opened the box moved the tuner card closest to the video card to the furthest slot to create more space around the video card. That seemed to drop temp by about 2°C. While watching a recording last night the system froze again. After a reboot I quickly went back to nvidia-settings but the temperature had not changed much.

I think my next step is to jerry-rig a cheap fan onto the video card. If that eliminates the problem then it might be time for a new card maybe with a fan. I don't think my box is as well ventilated as yours tjc. My PS sucks from the back and discharges out the side so that anything next to it is in a bit of a sheltered spot.

_________________
Intel DG965WH, Dvico DVB-T Lite x2, HDHR, Gigabyte GT220, KingstonSSD, WD20EARS version=latest


Top
 Profile  
 
 Post subject:
PostPosted: Mon May 09, 2011 6:50 pm 
Offline
Joined: Thu Mar 25, 2004 11:00 am
Posts: 9551
Location: Arlington, MA
I'd say that your box is substantially under ventilated, especially in the neighborhood of the video card. Your idle temperatures should be ~30C lower. Since it's May, you're well into fall there, so the ambient temperature shouldn't be a problem, which leaves poor ventilation as the elephant in the room.

Do you have an empty slot on the heat sink side of the video card that can be replaced with a perforated filler? (Something like these - http://www.amazon.com/SilverStone-AEROSLOT-Maximum-Vented-Covers/dp/B001NPEBEC

What do you have for exhaust fans on the box? Hopefully one or more 120mm fans (other than the PS) pulling warm air out of the box? At least one or more 80mm fan? Maybe a slot fan next to the video card? Anything?

65-70C is *much* to high for idle temperatures.


Top
 Profile  
 
 Post subject:
PostPosted: Mon May 09, 2011 8:54 pm 
Offline
Joined: Mon Aug 14, 2006 8:24 pm
Posts: 320
Location: Melbourne, Australia
You confirm my suspicions that 65°C is too hot. It's also cold for Autumn here so the heating is on so the room is 20°C.

My box is a Silverstone LC17 so it comes equipped with 2 x 80mm fans exhausting at the back. The CPU with its fan sits in front of these and a vent on the side allows air in. As I said in my previous post the PS exhausts out the side vent on the other side, so the space between the PS and the video card does not get cross flow.

By moving the tuner card I do now have a slot on the side away from the exhaust fans so those vented slot covers look a good idea. I might just make my own from an old slot cover with a drill and give it a try. If that does not work then a new card with fan is looking likely.

_________________
Intel DG965WH, Dvico DVB-T Lite x2, HDHR, Gigabyte GT220, KingstonSSD, WD20EARS version=latest


Top
 Profile  
 
 Post subject:
PostPosted: Tue May 10, 2011 6:38 pm 
Offline
Joined: Thu Mar 25, 2004 11:00 am
Posts: 9551
Location: Arlington, MA
Looking at pictures of the LC17 now... That case has some intake vents above the cards but it probably doesn't help much, and the side vent probably only helps the PS... You could probably mount another exhaust fan in the vent by the CPU if you had another intake... Are there any floor vents up under the drive bays?

Not a great case design in terms of air flow not matter how you slice it. Among other things, there's not much ventilation visible for the HDs either.

A ventilated slot cover next to the video card is probably your best bet. To test I'd just pull the blank closest to the HS side of the video card and see how much that drops the GPU temperature. It's hard to tell from the pictures I could find of the Gigabyte 7300GS cards, but if the HS under that cover will actually let air flow through from the components side to the back side, you might also want to play with some simple baffles above and behind the card to try to direct the airflow through it....


Top
 Profile  
 

Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 19 posts ] 
Go to page 1, 2  Next



All times are UTC - 6 hours




Who is online

Users browsing this forum: Google [Bot] and 26 guests


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Jump to:  
cron
Powered by phpBB® Forum Software © phpBB Group

Theme Created By ceyhansuyu