Dynaverse.net
Off Topic => Engineering => Topic started by: NJAntman on July 02, 2009, 10:25:57 pm
-
Still having issues with overheating system; haven't been able to pin down where HWMonitor reads TMPIN2 from but when it gets over 225F it's shutdown time.
Figured I'd better check the DDR2 to see if still OK so I loaded up MemTest86+ and let it run. About 3 minutes in right about 16% into 1st pass, shutdown. Let it sit an hour and ran it again and whamo, right about 15% another shutdown. Let it sit 1.5 hours and ran and again it shutdown at about 16%. No errors listed just before shutdowns. Loaded up normally to WinXP a half hour later and the HWMonitor reported TMPIN2 as 215F and CPU Core #0 at 192F and Core#1 at 216F, a half hour later! As i write this it is a half hour past bootup and the temps are down about 135F all across the three.
Since I can't get through a full pass of MemTest before shutdown I don't know if the chips have errors but it seems they sure as hell are overheating. I'm thinking TTMPIN2 might be located right around the DDR slots.
So the question for you computer gurus out there is do DDR2 chips overheat that quickly (3-4minutes) when tested? If these chips are heat damaged shouldn't I be getting errors right out of the starting blocks? Obviously time for some replacement RAM.
-
Here's an idea..just throwing it out there
http://www.farayandmea.com/Productsviewer.aspx?rand=HmvTP+1gLKsqEg1Ocb3Sjg==
Mike
-
Interesting find, and only about $15 at Newegg. If something like that would work maybe I can just leave of the side cover and blow some compressed air on the chips next time I run a memory intensive app like a virus scan. If I see a big drop in temps maybe I've nailed down what the problem is. Assuming I don't screw the chips by cooling too fast?
-
What temperatures does your BIOS report?
PU Core #0 at 192F and Core#1 at 216F,
I find the CPU temperatures to be extremely high and wonder at the accuracy of the reports.
-
Are the voltages set to auto?
-
What temperatures does your BIOS report?
PU Core #0 at 192F and Core#1 at 216F,
I find the CPU temperatures to be extremely high and wonder at the accuracy of the reports.
To answer that question I rebooted from WinXP to BIOS to see what i could see but the only available measurement is Current System Temp (never budged from 33C) and Current CPU Temp which started at 70C and gradually went up to 83C over the space of less than five minutes just sitting there in BIOS, before crashing. A an hour later as I type this CPU core#0 is floating around 56C and core#1 at 54C; the max values when I opened the HW Monitor were 72C & 74C.
-
Are the voltages set to auto?
The only voltage parameters in the BIOS that I could mess with are all set to"Normal" vs. specific voltage choices. I haven't changed a setting since I got this computer.
-
It crashed while in the BIOS screen? I don't think I've ever heard of that happening let alone seen it. Let us know what the final solution is.
-
Two things came to mind:
As nemesis, the accuracy of the readings from HWMonitor (bad thermocouple? - The BIOS suggests not.) Crashing from the BIOS? What year and make is the machine? Does it have an EFI or traditional BIOS?
Second thing that came to mind is I seem to recall reading about the potential emergence of hardware thermal protection exploits in malware. Run RootKitRevealer, study the results carefully. Do you happen to have a virtual sound device driver installed that you never installed yourself? (A very versatile and powerful rootkit I've been seeing around lately....)
Are these memory sticks original to the machine or new? If new, who installed them?
-
It crashed while in the BIOS screen? I don't think I've ever heard of that happening let alone seen it. Let us know what the final solution is.
1st time I've ever seen that also, but i had just been in WinXP and logged off, shut down, and booted to BIOS without much pause so the CPU must have been hot. While I was in BIOS I was jumping between sub menus so maybe that let the CPU heat up past the threshold (set at 80C). Not really a crash, more like a self shutdown once past the threshold? Still I wonder if I'm just jumping between options in BIOS menus without selecting anything is that going to heat the CPU, the RAM chips, both?
-
Two things came to mind:
As nemesis, the accuracy of the readings from HWMonitor (bad thermocouple? - The BIOS suggests not.) Crashing from the BIOS? What year and make is the machine? Does it have an EFI or traditional BIOS?
Second thing that came to mind is I seem to recall reading about the potential emergence of hardware thermal protection exploits in malware. Run RootKitRevealer, study the results carefully. Do you happen to have a virtual sound device driver installed that you never installed yourself? (A very versatile and powerful rootkit I've been seeing around lately....)
Are these memory sticks original to the machine or new? If new, who installed them?
What is an EFI?
system specs , a downgraded Gamer Ultra 7500 SE purchased from CyberPowerPC in March 2007:
CoolerMaster Stacker 830 Tower W/ Side-Panel Window
(Socket AM2) AMD Athlon™64 X2 3600+ Dual-Core CPU w/ HyperTransport Technology
FAN: AMD ATHLON64 CERTIFIED CPU FAN & HEATSINK + 3 EXTRA CASE FANS
HDD: Single Hard Drive (250GB SATA-II 3.0Gb/s 8MB Cache 7200RPM HDD)
MOTHERBOARD: (Socket AM2)GigaByte GA-M55SLI-S4 nForce4 SLI Chipset DDR2/800 SATA 16x PCI-Express MBoard w/GbLAN,IEEE1394,USB2.0,&7.1Audio
MEMORY: (Req.DDR2 MainBoard)1GB (2x512MB) PC6400 DDR2/800 Dual Channel Memory
XP Home Edition w/ Service Pack 3
POWERSUPPLY: Apevia 680W BEAST POWER ATX Power Supply
SOUND: HIGH DEFINITION ON-BOARD 7.1 AUDIO
VIDEO: NVIDIA GeForce 9600 GT 512MB PCI Express x16
16X DVD ROM
Tried RootKitRevealer but had to abort 10 minutes in as scanning pics in the My Documents folder had the temps up to 226F on Core#1 and 240F on TMPIN2.
-
Ok, two things I would suggest is first remove the case side panel and aim a large fan directly at the motherboard and set the fan on high. Then turn the system on and go into Bios, the main Bios screen should have a temp reading that is constantly updated. Once your in the Bios, just set and watch the temp. If the temp goes from zero to overheat/shutdown in about 2 seconds with a fan blowing on it then I would suspect the motherboard sensor (yes they do go bad). If the temp build up is slower then the CPU cooler may need to be reseated or replaced. Did you use thermal paste? If you did it could have went bad. If you didn't it is going to be he|| getting that CPU cooler off.
Second while you are watching the Bios, take a look inside the case and see if the CPU fan (or any fan for that matter) is spinning correctly. If the fan looks like it is catching, not spinning, or spinning really slowly compared to the other fans, you could have a bad fan or the fan is not getting proper power from the power supply or motherboard depending on how the fan is connected. So if the fan does not look right, try hooking it up to the power supply directly if it is connected to the motherboard (vice versa if the fan was connected to the power supply) and see if the fan was getting bad power. That is not usually the case but I want to check and see if there is a bad power lead somewhere.
This should get you started, hope this helps.
-
What is an EFI?
Short answer EFI is an alternative to having a standard BIOS. EFI was created by Intel and currently is used by Apple. (EFI = Extensible Firmware Interface)
Long answer (http://en.wikipedia.org/wiki/Extensible_Firmware_Interface)
-
Don't AM2 & AM2+ chips interface directly with the memory? Could what he's seeing actually be something overheating in the CPU?
-
Don't AM2 & AM2+ chips interface directly with the memory? Could what he's seeing actually be something overheating in the CPU?
There are lots of things that could cause the CPU to overheat. This problem could be anything from a bad CPU to poor air intake/exhaust. My best guess based on what I've read so far is a bad CPU/cooler or a bad power supply. Won't know until we start eliminating the possibilities.
-
Don't AM2 & AM2+ chips interface directly with the memory? Could what he's seeing actually be something overheating in the CPU?
Ack! As if the possibilities weren't daunting enough I hadn't suspected something like that! When MemTest86+ bombed out I figured must be the RAM overheating, no I have to wonder if its the CPU. Running Avira virus scanner, SpyBot Search & Destroy and SuperSpyware in scan mode will almost always cause a shutdown; are those aps CPU intensive or RAM intensive, or both?
I'm using the laptop now as I just watched the CPU temp start at 37C and go to 81C in 10 minutes just sitting in BIOS watching the temp display; it crashed when I left the BIOS to start up XP.
I'm going to go with the following plan. I'll get some thermal paste and pull the CPU and reseat it. If that doesn't lead to a big improvement I'll get new RAM, stuffs cheaper now and I'll need it in time anyway. After that if things haven't improved I'll start messing with getting fans in there.
-
LOL, your bad luck is rubbing off on us Antman. I woke up this morning to an awe full racket. Turns out one of my Big 120mm Thermaltake fans went out. I still have 5 more in the case, and adjusted it properly, but it was a shock to say the least.
<sniff> my system doesn't quite seem load enough now.
Stephen
-
Run it with the side of the case open and you can probably eliminate ventilation.
-
I've thought about that, only problem is, We have some "Critters" we have been fighting this early summer. Once Orkin takes care of them, it will be ok.
Stephen
-
I've thought about that, only problem is, We have some "Critters" we have been fighting this early summer. Once Orkin takes care of them, it will be ok.
Stephen
I've thought of running with the side open as well but I can't call Orkin to control my critters. Yesterday my critter help me clean the van and somewhere in the blink of an eye managed to stuff the dash power outlets with pennies. I shudder to think what how he could modify an open tower case.
Anybody with CPU experience want to recommend a good thermal paste for reseting the heatsink? Local Radioshack sells one brand only at $12 per tube but there are several computer shops nearby I'm going to try tomorrow.
Sorry about your fan Sirgod, I can't imagine the racket but it reminded me that this whole mess started about a year ago when the small vid card fan went with an annoying warbling. That led to running two SLI cards and also about when these system temp problems started. I should've been content to just kick the friggin tower when I could.
-
Anybody with CPU experience want to recommend a good thermal paste for reseting the heatsink? Local Radioshack sells one brand only at $12 per tube but there are several computer shops nearby I'm going to try tomorrow.
Last time I did so a couple of years ago the top brand was Arctic Silver.
-
Anybody with CPU experience want to recommend a good thermal paste for reseting the heatsink? Local Radioshack sells one brand only at $12 per tube but there are several computer shops nearby I'm going to try tomorrow.
Last time I did so a couple of years ago the top brand was Arctic Silver.
OK, I'm seeing different kinds of thermal grease/paste/ceramic as well as thermal pads. Anybody use a thermal pad? Seems like that might be a more even application, and less mess, but since they're cheaper I wonder if they're inferior?
Moved the two DDR modules over to the next set of slots, figure it gives another half inch of space from the CPU. No difference in temps seen as a result.
Alrighty, just finished re-greasing with some Artic Silver. What fun, I loosened the heat-sink/fan but couldn't get it off the CPU, so I look for the CPU lever and its no where to be seen. That is when I noticed that the whole heat-sink sits on this orange bracket that seems to be an add-on to the board. Well the lever couldn't be inside the bracket, that makes no sense, or does it? So i pulled a little harder, a little more, and POP! out comes the heat-sink/fan/CPU and there is the lever - inside the bracket?! WTF? I guess I'm damn lucky that the lever jumped to the release position and I didn't break the CPU pins. Or is it supposed to work like that?
Anyway, cleaned the CPU and heat-sink, applied two rice grain size dabs of Artic Silver and put it back together.
Also took an old 60mm fan and rigged it to the open air grill about even with the CPU and DDR modules. It's at least four inches away but i figure it can't hurt to try it.
Booted into BIOS and notice a difference: before re-greasing after re-greasing
startup 37 33
+1 minute 43 36
+2 minutes 46 41
+3 minutes 49 40
+4 minutes 59 42
+5 minutes 62 43
+6 minutes 66 44
+7 minutes 69 45
+8 minutes 73 44
+9 minutes 77 44
+10 minutes 81 44
HWMonitor is showing Core#0 at 50C & Core#1 at 44C, so it looks like I got back 5-10C with re-greasing. I will try a virus scan later to see if it makes a necessary difference where I can let it go for the needed half hour without overheating. Then a Rootkit scan. Then a MemTest+86 so i can get a read on the DDR.
Crossing fingers.
-
OK, I'm seeing different kinds of thermal grease/paste/ceramic as well as thermal pads. Anybody use a thermal pad? Seems like that might be a more even application, and less mess, but since they're cheaper I wonder if they're inferior?
HWMonitor is showing Core#0 at 50C & Core#1 at 44C, so it looks like I got back 5-10C with re-greasing. I will try a virus scan later to see if it makes a necessary difference where I can let it go for the needed half hour without overheating. Then a Rootkit scan. Then a MemTest+86 so i can get a read on the DDR.
Crossing fingers.
First, yes thermal pads are inferior. They don't cool as well and they are a total pain in the @$$ to get off if you need to reseat the CPU cooler. To my knowledge Artic Silver is still the best thermal grease.
Second, from what you are telling us it sounds like it looks like re-seating the CPU cooler did the job. Have you tried to run the system without the 60 mm fan? My guess is the temp won't go up much if at all. If it does go up a lot without the extra fan then I would suggest buying a superior CPU cooler. Your motherboard manual should have information on how to remove the CPU cooler without damaging the CPU or motherboard. Also try putting the ram modules back in their original slots and see how much the temps rises before you're done.
I do find the 50 C temp on core #0 a little troubling but it is a AMD so I would expect it run a little hotter. Not to mention most of the processing load is probably being handled by that core since a lot of software still doesn't take advantage of multi-core CPU.
Let us know how the software tests run I'm really curious to know how it works out. Looks like you've done a really great job so far and more importantly you didn't have to pay someone to do it.
-
OK did full scans with SpyBot S&D, SuperAntiSpyware, and then Avira, no problems.
Did a Rootkit Revealer scan, got through it and had these results:
HKLM\SECURITY\Policy\Secrets\SAC* 3/24/2007 11:52 PM 0 bytes Key name contains embedded nulls (*)
HKLM\SECURITY\Policy\Secrets\SAI* 3/24/2007 11:52 PM 0 bytes Key name contains embedded nulls (*)
HKLM\SOFTWARE\Microsoft\Cryptography\RNG\Seed 7/8/2009 5:47 PM 80 bytes Data mismatch between Windows API and raw hive data.
HKLM\SYSTEM\ControlSet001\Services\sptd\Cfg 4/11/2007 11:15 AM 0 bytes Access is denied.
Good, bad, ambiguous?
Ran MemTest86+ this morning, saw it go through about 30% of 1st pass test but had to leave, when I came back this afternoon the computer was shutdown so it looks like the DDR is still a problem. Going to run that again and check on it every 15 minutes to see when it crashes out. Temps as I type Core #0 50C & Core#1 45C and TMPIN2 a balmy 37C. That right there is the biggest change, TMPIN2 used to get up to 240F just before crashing (pre re-greasing).
Ran MemTest86+ 3 times last night, each time it gets about halfway through 3 complete passes without any errors but the sytem always shuts down. I'm thinking overheating but MemTEst86+ doesn't give temp readings so don't know for sure. Brings me back to my question and Nemisis' comment, on this system does running MemTest86+ put a load on just the RAM, or both CPU & RAM?
Have a feeling I'm going to be buying some DDR2 for peace of mind.
-
Brings me back to my question and Nemisis' comment, on this system does running MemTest86+ put a load on just the RAM, or both CPU & RAM?
Have a feeling I'm going to be buying some DDR2 for peace of mind.
I don't think that was my comment.
An AMD chip would likely stress some CPU components as well as the memory as it integrates the memory controller and has the hypertransport links to the memory.
-
I'm enjoying this banter, as I will be getting a bonus check soon, and hope to learn..
I've built mine and my wife's last three computers, and haven't had any problems other than the occassional hard drive crashing...
My wife does push her system, having three different browsers open with 5 to 7 sites in each one (mozilla, IE, and some other one)..6 or 7 programs running (usually solitair, mah jong and other members of that genre), word, excel, and outlook and something else..and then she asks (in the Glenn Beck wife's voice) "Why is it so slow all of a sudden" :smackhead:
Well, good luck Nick
Mike
-
Well, just a thought, but what are your shutdown temps set to in the BIOS? Are they too low?
-
BIOS has "Warning Temps", which as I've never seen a warning displayed I'll take as the shutdown temps, and are set to 80C on the system and 70C on the CPU with any values between 60C - 90C being the selectable range.
-
BIOS has "Warning Temps", which as I've never seen a warning displayed I'll take as the shutdown temps, and are set to 80C on the system and 70C on the CPU with any values between 60C - 90C being the selectable range.
Any way you can get a temp reading while memtest is running? I haven't run memtest in ages myself. Does Bios display the correct amount of Ram? Some people overlook the amount of Ram that Bios can see, that's why I ask. Have you tried running memtest with only one stick of Ram at a time? Just because one stick of Ram is bad doesn't mean that both sticks are bad. I'd hate to see you waste money replacing perfectly good Ram. Have I totally annoyed you with all these questions yet??!? I seem to be on a roll the last couple of days... ;)
-
The version I have is MemTest86+ and there is no temperature display. The BIOS is showing the correct amount of RAM. Hadn't thought of running one RAM stick at a time to isolate the problem, will try that this weekend. Never get tired of questions, they lead to the best discoveries.
-
Whilst on NewEgg looking at the memory slot fans Mike suggested I saw these PCI slot case coolers and figured what the heck, they're cheap and easy to install.
So now starting at the bottom of the case I have a PCI slotted TV card, an open PCI slot, then over the first PCIE (x16) is this cooler combo http://www.newegg.com/Product/Product.aspx?Item=N82E16835114024 (http://www.newegg.com/Product/Product.aspx?Item=N82E16835114024) with its directional fan blowing over the next two empty PCIE (x1) slots and onto the hotspot of the SLI video card in the nest PCIE (x16) slot. Then on the other side of the SLI video card is this cooler http://www.newegg.com/Product/Product.aspx?Item=N82E16835119065 (http://www.newegg.com/Product/Product.aspx?Item=N82E16835119065) which seems to be effetively sucking air down from the CPU heatsink and out the back as the Core #0 & Core #1 temps are running 5 degrees cooler. Noise ain't too bad and I can feel a lot of warm air being pushed out the riser vents of each cooler. Only drawback is things are getting crowded but I'll accept it as a $25 solution.
I'm going to run it this way and install some new memory. I'm going to hold off putting the other SLI video card back in until I can get the CoolerMaster multifan bracket for the case and load that up with four 120mm fans to keep things breezy in the case.
-
Just my two cents, from what I've seen so far my suggestion would be to save money for a new case. Something that has good cooling a solution. A Cool Master HAF 932 (http://www.newegg.com/Product/Product.aspx?Item=N82E16811119160) case is something I've been looking at for my next build. Here's a picture (http://www.newegg.com/Product/ShowImage.aspx?CurImage=11-119-160-S01&ISList=11-119-160-S01%2c11-119-160-S02%2c11-119-160-S03%2c11-119-160-S04%2c11-119-160-S05%2c11-119-160-S06%2c11-119-160-S07%2c11-119-160-S08%2c11-119-160-S09%2c11-119-160-S10%2c11-119-160-S11%2c11-119-160-S12%2c11-119-160-S13%2c11-119-160-S14&S7ImageFlag=1&Item=N82E16811119160&Depa=0&WaterMark=1&Description=COOLER%20MASTER%20HAF%20932%20RC-932-KKN1-GP%20Black%20Computer%20Case%20With%20Side%20Panel%20Window%20-%20Retail) of the case. Just a thought.
Did you get a chance to test the Ram independently? What is the overall temp, the CPU temp, the Ram temp, and GPU temp if you have them.
-
Holly Fresh Breeze Batman! THREE 230mm fans!!!
Does it come with a "Beware Prop Wash" sticker?
Alas I'm still awaiting an opportunity to test the RAM sticks independantly although I already bought 2GB of new 800Mhz DDR2 (at $26 I couldn't resist). Figure if the original RAM is still good I can offer them up to the PC spare parts thread.
-
LOL
I'm sure you could plant one or two of those stickers on it's side if you really wanted to. Use some really good glue though...
Wow I didn't realize the Ram was so cheap. The Ram coolers you are buying or bought should probably cost as much. Speaking of which, Ram shouldn't get that hot unless the rest of the parts are getting hot as well. That's another indicator that the case is being overwhelmed by what's inside.
-
Here ya go Nick, if you get that case, get some of these from here...
http://www.aeroconsystems.com/misc/beforeflight.htm
-
Here ya go Nick, if you get that case, get some of these from here...
[url]http://www.aeroconsystems.com/misc/beforeflight.htm [/url]
lol, I use to keep one of those hanging from the rear view in my car. I have to say, it never left the ground, so it worked.
-
Well, reviving an old subject...
I finally got some time and built a D2 system, and moved the hard drives over to the new case...and wouldn't you know, the C: drive came up with an error...so, I tried to recover with the XP disk..well..it keeps asking me for the administrator password, which of course I don't remember...
Anyone know of a way to bypass this so I don't have to reformat and reinstall..I'm old, and forget where all the friggin' disks are since the kids have moved out and we remodeled spare bedrooms into offices...
Thanks ahead of time..
Mike