New Build

[quaxo]

Here's the cooler with the side cover off. You can see it comes right up to the side. It blows directly out through the newly replaced rear fan.



With the cover on, you can see the cooler comes right up to the window with about 3/16" to spare. The fan, despite being a slim model (1/2"), is too big to go on the inside, so I removed the old fan guard, screwed it to the outside of the fan, then screwed the fan to the outside of the case. The cooler blocks a little over half of that hole though.



So much better now. That cooler is soooooo quiet.

[patio]

Nicely done...hope she stays chilly for you.
BTW how much was the Coolermaster ? ?

[quaxo]

Normally, I think they run about $50 retail, I got this one from an acquaintance's shop for about $35, which I think is what he pays for them. He builds more high-end gaming systems than I do, so I send a lot of people his way who are looking for that and he sends me people who are looking to build low-end budget computers and people who need repairs and what-not, so we help each other out when we can.

The case fans I already had in stock (they're all Cooler Master, only fans I sell now), but I don't normally keep large CPU coolers around since I don't get a lot of requests for them. I keep a couple stock coolers for Intel and AMD around for replacements in systems that have them fail, but those I usually get for free or next to nothing.

-=EDIT=-
Just checked online. Looks like they're $60-70 most places. Guess I saved more than I thought. 

[quaxo]

Ok, there's something else wrong now. I don't know if it's a heat related problem or what. Because it was running so hot before, I assumed it was, but now after making these changes, I installed CoreTemp and let it log the temperature changes over the last 12 hours. It takes a reading every 10 seconds, noting the time, current temp, record low, record high, CPU load, and CPU speed. It saves it in a CSV file (readable by Excel, OpenOffice, and other programs). These readings are without air conditioning, this is the computer idling at 35-38C (good enough) with room temp at about 30-35C.

At about 1am, the processor kicked in for 2 minutes and the temperature went up to 46.

At about 4am, the processor kicked in again for about 2 minutes and the temperature went up to 47 (the max for the night).

Code: [Select]

Time    Temp   Min Max Load Speed
04:15:27,38°,,,32°,46°,2%,803.55 MHz,
04:15:37,38°,,,32°,46°,0%,803.55 MHz,
04:15:47,43°,,,32°,46°,55%,3415.10 MHz,
04:15:57,43°,,,32°,46°,63%,803.55 MHz,
04:16:07,44°,,,32°,46°,71%,3415.10 MHz,
04:16:17,45°,,,32°,46°,40%,3415.10 MHz,
04:16:27,44°,,,32°,46°,46%,3415.10 MHz,
04:16:37,44°,,,32°,46°,60%,3415.10 MHz,
04:16:47,43°,,,32°,46°,49%,803.55 MHz,
04:16:57,43°,,,32°,46°,48%,3415.10 MHz,
04:17:07,44°,,,32°,46°,60%,3415.10 MHz,
04:17:17,43°,,,32°,46°,42%,3415.10 MHz,
04:17:27,46°,,,32°,46°,49%,3415.10 MHz,
04:17:37,46°,,,32°,47°,55%,3415.10 MHz,
04:17:47,46°,,,32°,47°,78%,3415.10 MHz,
04:17:57,40°,,,32°,47°,8%,803.55 MHz,
04:18:07,40°,,,32°,47°,2%,803.55 MHz,
I assume it was either an AV update or some other minor thing.

Here's the last minute or so before it stopped recording the log:

Code: [Select]

Time    Temp   Min Max Load Speed
09:06:11,39°,,,32°,47°,6%,803.55 MHz,
09:06:21,39°,,,32°,47°,6%,803.55 MHz,
09:06:31,38°,,,32°,47°,8%,803.55 MHz,
09:06:41,40°,,,32°,47°,12%,803.55 MHz,
09:06:51,39°,,,32°,47°,8%,803.55 MHz,
09:07:01,39°,,,32°,47°,14%,803.55 MHz,
09:07:11,39°,,,32°,47°,9%,803.55 MHz,
09:07:22,39°,,,32°,47°,8%,803.55 MHz,
09:07:32,39°,,,32°,47°,6%,803.55 MHz,
09:07:42,39°,,,32°,47°,8%,2209.77 MHz,
09:07:52,39°,,,32°,47°,12%,803.55 MHz,
09:08:02,39°,,,32°,47°,11%,803.55 MHz,

09:08:02 being the last entry, but nothing unusual temperature wise, so I think this rules out my initial theory that the computer was overheating while I was away. Something within 10 seconds of 09:08:02 caused the computer to freeze, and it has done this about 5-6 times now if I leave the system on all day or night, but I've never had it happen while using it and therefore never see it actually occur.

Here's something particularly odd I noticed: This is a 16:9 screen. When receiving a 4:3 signal, it (the TV) colors the outside edges a dark gray (not black, this is normal and how this TV functions). When it freezes like this, it seems to have changed to a 4:3 signal as these outer gray parts are showing and the TV indicates in its status that it's 4:3. The 4:3 section that is showing is black, with a slightly distorted mouse cursor stretched wide slightly. The computer is running, but the mouse and keyboard are unresponsive.

Is it possible the two video cards (or even one of them) are overheating? This program doesn't log video card temps, do you know of any that do?

[Salmon Trout]

20 minutes of gaming and it seems to be peaking at 48-52C, so should be alright now.

A few thoughts -

I know my hardware is not directly comparable - In my little Shuttle SN78SH7 I have a Phenom 945 running at 3.0 GHz, (it's the 95 Watt part, there is also a 125 Watt part which runs a bit hotter) cooled by the stock Shuttle heatpipe cooler (with a radiator which a vertical fan cools) and with Cool 'n' Quiet enabled it idles at around 25 C and ramps up to around 38 - 40 on load. The 965 has (or had) C2 stepping (140 Watt) and C3 stepping (125 Watt) variants - if it's a newly bought part I would guess it's a C3 (the "Revision" section of the Core Temp app should indicate this)

I have seen reports of the 955 BE 9 (140 Watt) running at around 25-30 idle up to around 48 load and this being considered OK so I think your temps are reasonable especially if your ambient temp is on the warm side.



Speedfan should be able to measure the GPU temp(s). It has a Notification Area temp readout like Core Temp's but not such an attractive one (weird font) and also can do logging and show a temp chart in real time.

[quaxo]

Alright, I'm pretty sure it isn't heat at all now. The air conditioner is on, but the computer just did it again when I was out of the room.

Windows 7 was still using the default "Balanced" power setting. If it's putting the something to sleep and failing, I'll find out soon enough. I've just shut off any feature that allows it to do that.

CoreTemp does indicate it is a RB-C3.

[Salmon Trout]

Alright, I'm pretty sure it isn't heat at all now. The air conditioner is on, but the computer just did it again when I was out of the room.

Windows 7 was still using the default "Balanced" power setting. If it's putting the something to sleep and failing, I'll find out soon enough. I've just shut off any feature that allows it to do that.

CoreTemp does indicate it is a RB-C3.

Have you looked at the event logs?

I edited the "power saving" plan because it can set the CPU to a lower power level than the "balanced" one.

Require a password: never
Hard disk: Turn off hard disk after "9999 minutes"
Desktop background settings: "enable"
Wireless Adapter settings: Power Saving Mode: "Maximum Performance" (I don't have a wireless adapter)

Sleep:
   Sleep after "Never"
   Hibernate after "never"
   Allow wake timers "Disable" (I'm not sure if this makes a difference if sleep/hibernate are set to "Never")

USB settings: USB selective suspend setting: "Disabled"

Power buttons and lid:
   Power button action "Do nothing"
   Sleep button action "Do nothing"

PCI Express: Link State Power Management "Maximum power savings"
I have never really thought about this; when you hover over it an infotip says "Attempt to use the L1 state when the link is idle"
The other states I can see are:
Moderate power savings - "Attempt to use the L0S state when the link is idle"
Off - "Turn off ASPM for all links"

Processor power Management
   Minimum processor state "0%"
   System cooling policy "Passive" (slow the processor before increasing fan speed) The alternative is "Active" - increase the fan speed before slowing processor)
   Maximum processor state (100%)
   
Display: Turn off display after: "Never"

Multimedia settings
   When sharing media: "Prevent idling to Sleep"
   When playing video: "Optimize video quality"   

[quaxo]

Yeah, went through the logs. The next entry of note after the shutdowns is something to the effect of "The previous system shutdown at 3:20:22 PM on ‎4/‎30/‎2011 was unexpected." except the date and time of course changes based on when the system stopped responding. It is accompanied by another entry of "The system has rebooted without cleanly shutting down first. This error could be caused if the system stopped responding, crashed, or lost power unexpectedly."

The final message of note after those is this:
The computer has rebooted from a bugcheck.  The bugcheck was: 0x0000008e (0xc0000005, 0x922622a7, 0xa2cea794, 0x00000000). A dump was saved in: C:\Windows\MEMORY.DMP. Report Id: 043011-17160-01.

I'm reading online through stuff about this last message right now.

[quaxo]

0x0000008e (0xc0000005... seems to be tied to a lot of things. Bad memory, drivers, and even sleep mode failures.

Since disabling sleep and hibernation 12 hours ago, I left the system running and it's still going now. I suspected it was this because this never occurs while it's in use, only when I'm not using it.

I'm going out for the day, so I'll leave it and see if it's still running later.

[patio]

I'd run MemTest...

[quaxo]

If it happens again, I will. The memory is brand new, so there shouldn't be anything wrong with it, but things happen.

Since I changed the power settings (no sleep, no hibernate, never turn stuff off automatically) it hasn't happened. 16 hours now and no freeze.

[Aegis]

but the computer just did it again when I was out of the room.

So, don't leave the room!   

[patio]

    

[Salmon Trout]

I wonder if the right sleep setting is enabled in the BIOS? I believe some W7 systems have trouble coming out of S1 sleep in which case S3 should be selected in the BIOS.

[patio]

In regards to the memory being new also on 1 build last month i had a brand new stick of OCZ RAM that was bad....it happens.