Computer Hope
Hardware => Hardware => Topic started by: zolton35 on April 10, 2016, 10:06:02 AM
-
I posted this on another forum but no help. I will post a copy of my post from there as it has all the information and what i have tried so far to find the root cause of the problem:
Ok so what i had before was this build:
CPU
AMD Athlon II X2 270 3.4GHz Dual-Core Processor
Motherboard
MSI 760GM-P23 (FX) Micro ATX AM3+ Motherboard
Ram
2 G.Skill Value Series 4GB (1 x 4GB) DDR3-1333 Memory
HDD
Western Digital WD Blue 500GB 2.5" 5400RPM Internal Hard Drive
Video card
Asus Radeon R7 240 2GB Video Card
Power supply
Antec 450W ATX Power Supply
Disk Drive
LG GH24NSB0 DVD/CD Writer
Windows 10 64
Now i upgraded my cpu to an FX 8300
I have no plans to over clock and bought the cpu on sale as an upgrade so went with the stock fan that came with it. I installed it and had it on just web surfing for a long while to see how it does. I launch a game Ark Survival Evolved to be exact and i was waiting for my mods to install so i could play when my monitor goes black but the pc is still running and would not power off via its power button so i unplugged it checked my video card and all was well.
I also double checked my cpu fan was secured so i powered it up again all seems fine so i started monitoring the temperatures using cpuid hw monitoring. I noticed my temps are all over the place.
TMPIN0 min 24 max 38
TMPIN1 min 37max 47
TMPIN2 min 29 max 31
AMD FX 8300 Min 14 Max 46
AMD Radeon R& 240 min 32 max 36
Now all seems ok that is at idle while posting this. When i stressed it with Cpuid TMPIN1 hit up to 80 and my cpu was hitting closer to 60+
I was not sure on its accuracy so i tried out AIDA64 extreme. And not long into it less then 3 minutes screen went black and i had to unplug the pc again to shut it down as again the power button did nothing.
So this time i took my pc apart. Completely cleaned off the stock paste and put on some arctic silver to the cpu and put it back together. Not sure what is going on i tried AIDA64 Extreme again and when i saw it jumping to high this time stopped it. The weird thing? The temps kinda lingered there a while. Even when i went into task manager and killed the program. Eventually they did drop back down to normal and i reset the temps to monitor them as i post this to see the changes.
I hope it is not the stock cooler. But i guess it could be the cheap MOBO. I kind of suspected it may be the dual core optimizer that was installed playing havoc on the temps so i uninstalled it and will run another test after i post this after a pc restart to be sure it is gone from the equation. Any ideas on a specific cause? I'd hate to opt for a better cpu cooler when i've no plans to oc the cpu. But i will if that is the problem. Or i guess i could try the older heat sink i had on my other cpu if need be. So any ideas?
Edit to add just tried cpu z stress test while monitoring with cpuid hw monitor and temps were as follows:
TMPIN0 min 24 max 56
TMPIN min 37 max 70
TMPIN2 min 28 max 31
FX 8300 min 14 max 57
AMD Radeon R7 240 min 32 max 36
My case has 5 fans all set to exhaust 2 on rear one on side one on top and one in the front. I stopped the stress when it hit 70 myself on TMPIN1. Not sure what temp that is monitoring but i do not like it at all.
Ok i tried the cpu fan that i had in it before and stress tested it with CPU z and let TMPIN1 hit 80 my cpu hit 65 under that stress so i then just stopped it out right. I know some are you are probably thinking that it is the cpu fan. Which is why i tried the other one i wanted to see if it was that cpu fan that was bad. But apparently doesn't seem to be the case. I just can not figure out what TMPIN1 is. It seems to be the mobo or a component of it. Could it be the onboard graphics on this mobo doing this running that hot? Or is it just a bad mobo? Or is it the cpu heat sink just not being up to the task?
Also on CPUID HW monitor there is a system fan 2 but for some reason it is only running at 50%? I'd gladly buy a new heatsink if that is the problem. But i'm not sure if its the mobo not up to snuff and overheating or if its the cpu fan not doing its job.
I tried running it with the case open. And tried the stress test again. And it is still running high. TMPIN1 hit 81 before i was able to stop it but at least it did not shut down this time. Cpu still did not hit above 65 when TMPIN1 increased to 81. But that is not to say it would not have done so. I'm at a loss on what is causing this and the best fix. I mean if i buy a new heatsink and that is not the cause then i'm wasting money. I just want to figure out the exact problem so i can fix it. One would think the heatsink that comes with the cpu would be adequate enough to handle the temps with it not being overclocked. And it may be that it is and the cpu socket on the mobo is running hot so mobo needing replaced. Any help would be greatly appreciated.
-
Welcome to the forum, hopefully we can help you out :)
Bear in mind that readings from Speedfan don't always correspond with working/active sensors, so they can't always be trusted. Having said that, I'm suspecting that your VRMs are overheating - VRM = Voltage Regulator Module, they're basically what supplies power to the CPU. Some cheaper AMD boards, especially MSI, are known for issues with power hungry CPUs such as the FX8300, and your symptoms sound somewhat like they're overheating. That said, I don't think TMPIN1 is measuring your VRM temperatures directly as they usually don't have sensors and can usually take around 100-105C anyway, but it may well be nearby hence the correlation.
There are two fairly easy ways to test this theory. One - if you happen to have a spare fan, or even if you use a case fan, just connect it up and point it to the VRM area on your board (this is to the left of the CPU socket, they're the small black square things) and run the test again, see if it lasts any longer with some cooling directed at the VRMs. Two, you could underclock and undervolt your CPU a little to reduce its power consumption, this will reduce stress on the VRMs causing them to heat up less - again if it runs for longer this would potentially point to a VRM overheating. Alternatively if you can disable turbo mode in the BIOS, give that a try too.
The VRMs on your board are specced to 95W which is in theory the FX8300's TDP, however there is evidence out there that many "95W" VRMs aren't truly capable of this, and AMD's FX CPUs don't always seem to adhere to their published TDP either, both of which obviously cause issues.
I'll stress that this is just speculation at this point, but at least it gives you a few things to try, to hopefully narrow down the issue a little.
-
I did try it with the case open. And temps did not change. I tried a second cpu fan i had that was in it before for my athlon cpu and its in there now with no change. I tried it with arctic silver thermal paste doing away with the cheap stuff that came preloaded on the heat sink with no change. My case has 2 80mm fans on the rear 1 80 mm on top and 1 80 mm on the front and a 120 mm fan on the side. I'm up for adjusting the voltages if they would help. Although i'm no overclocker so never fiddled with voltages before so would need some advice on how to do it. And i am using CPUID HW monitor not speedfan to monitor things. While not 100% accurate it is much more accurate then speedfan is. And the program i am using to stress test the cpu is call cpuid.
I chose this board as it was cheap and seemed to fit my basic needs. But i see now in hindsight you generally get what you pay for. I had to build this rig in a hurry though to replace an older dying build hence why i had an athlon 270 instead of an fx 8300 to start with. I do suspect some thing is over heating the cpu and it is some component within the mobo. But confused as to whether an aftermarket heat sink would fix it or if it is just best to just replace the mobo. If turning down the voltages would help and do away with the over heating and allow me to play my games until i can get a replacement board i'm definitely down with that option. I can check my bios and see if i can disable the turbo and see how that plays out. And i'll post back after a short test with that.
-
Ok now that just scared the crap out of me. I adjusted bios removed the turbo mode and adjusted bus from 200 to 190. I ran a stress test here are the readings as i stopped it before things got really bad:
TMPIN0 min 25 max 64
TMPIN1 min 40 max 95
TMPIN2 min 29 max 32
My cpu on the other hand? min 16 maxed at 206 at which point i stopped the test
hdd min 32 max 33
GPU min 31 max 36
So seems some thing is very wrong here. And i just got this cpu yesterday.
-
In BIOS enable Cool'n Quiet this will drop the clock to around 1400Mhz idle for that CPU. I have the FX8350 and stock cooler is not the greatest and it runs very warm at 4Ghz without Turbo. I found that with cool'n quiet enabled the CPU temps are much lower as for the 8 cores dont have to run full tilt for most applications.
Also I'd set everything in BIOS back to defaults before enabling cool'n quiet and this way nothing is old config from the prior CPU. If your bios somehow has the voltages ramped up it could run it hotter than normal.
I am guessing your using the stock cooler that came with the new CPU where there was already thermal compound on the bottom of heatsink and it was mated to the CPU with a clean bond, and the heatsink wasnt removed again and reinstalled after initial installation? If heatsink was locked down and then removed and then placed back down again the thermal compound might be trashed and you need to clean CPU and Heatsink surface of the prior compound and add new.
-
He re-did the paste above.... 8)
-
I tried redoing the paste still getting those numbers. I just put on cool and quiet with turbo still disabled and hit these temps:
TMPIN0 min 23 max 65
TMPIN1 min 37 max 95
TMPIn2 min 29 max 32
CPU min 14 maxed again at 206 that is celcius so 402 f
So that did not help sadly. The mobo is less then 2 years old and the cpu is less then 2 days old. I am very worried about these temps.
-
I dont trust those temp readings...
It'd be on fire if those were actual.
-
I dont trust those temp readings...
It'd be on fire if those were actual.
This, that CPU temperature definitely isn't right. Either the CPU temperature sensor is misreporting or Speedfan is being Speedfan. Try HWMonitor and see what that comes back with.
-
I am using cpuid HW monitor lol I said so before that is what I have been using I never use speed fan. A and like I said it has been confusing unless the sensor is bad maybe?
-
I am using cpuid HW monitor lol I said so before that is what I have been using I never use speed fan. A and like I said it has been confusing unless the sensor is bad maybe?
Apologies, for some reason I thought you were using Speedfan... :-[
The sensor sounds screwy to me, it's not uncommon that they'll report stupidly low idle temps like yours is doing, but it's rarer for them to report totally incorrect max temps.
Whether the sensor reading incorrectly is related to the actual issue you're having, I'm not sure.
When you said...
I did try it with the case open. And temps did not change. I tried a second cpu fan i had that was in it before for my athlon cpu and its in there now with no change.
Did you mean you'd tried what I said, to have a fan blowing over the VRMs? If so then I'd say we can probably rule that out.
To confirm, did you have any issues before changing the CPU?
Are you running the latest BIOS for your board?
Have you checked your temps in the BIOS to see if they're still reporting weird numbers?
-
Putting a fan directly on it I have not done as my side case fan is a 120 mm and sits not but 4 to 5 inches from the CPU I am willing to try it or even switch it to intake if need be. I can not done remember if I did update my bios or not to be honest. Although I did watch the bios temps and they seemed fine.
I am concerned I may need to replace my mono for 2 reasons first being cost the second is Microsoft's toe that ties your windows key to the mono meaning if I replace the mono I need to get a new windows key if they refuse to reactivate my old one which is a pita. So I am hoping to find a fix for it that will not add further costs time then need be.
-
Putting a fan directly on it I have not done as my side case fan is a 120 mm and sits not but 4 to 5 inches from the CPU I am willing to try it or even switch it to intake if need be.
It's worth just holding it as close as you can, just as a simple test to make sure the VRMs have a decent amount of airflow for cooling. For an example, see here - http://www.overclock.net/t/1318995/official-fx-8320-fx-8350-vishera-owners-club/17710#post_20297011
I can not done remember if I did update my bios or not to be honest. Although I did watch the bios temps and they seemed fine.
Worth checking your current BIOS version and comparing it to the one on MSI's support site, just in case.
I am concerned I may need to replace my mono for 2 reasons first being cost the second is Microsoft's toe that ties your windows key to the mono meaning if I replace the mono I need to get a new windows key if they refuse to reactivate my old one which is a pita. So I am hoping to find a fix for it that will not add further costs time then need be.
You shouldn't have any issues reactivating with a new board, you may just have to use the phone activation. If you do have to speak to them it's just a case of explaining you repalced the motherboard and it shouldn't be a problem.
-
OK I had my wife hold a 120 mm fan on it and the temps on the CPU showed holding between 62 to 65 and tmpin1 was holding at 91 and below I did not test it for more then a few minutes as that was what help I was allotted lol I wanted to be at the PC to be able to stop the test if need be. The temps could have gone higher I am not sure but they were not going a lot lower that is for sure. I hope that info can help you to help me diagnose what is going on with this and a possible solution.
-
OK I was curios about lowering voltages and saw that most who over clock watch the vcore voltages and most never go above 1.4 in v core to of. I remember seeing my vcore at 1.92 at one point when stressed. So I checked my bios and lowered my voltage settings I could find to the lowest I possibly could. I then stress tested for 10 minutes straight some thing I could not do before. The results?
Tmpin0 min 25 max 59
Tmpin1 min 39 max 79
Tmpin2 min 28 max 29
CPU temp? Min 15 max 57
I am curious as to whether or not that fixed it? Could my bios have reverted to some built in of setting? Was I running my athlon on that? Now my vcore hits min 0.856 and maxes at 1.52 during a 15 minute stress test. And when I stopped stress testing temps seem to of went down a lot faster. So did that fix it I how? Ido not know.
-
I tried to launch the game and played just fine even though I hit a small bit of stuttering here and there but that could be my GPU. So this seems to have been my problem. Every thing seems much better and well within the norm for temps now. In case some one finds this in a search with similar problems do not leave your volt settings on auto as that seems to be the problem I was having as it was bolting up when stressed sucking more power then needed and by extention increasing the heat.
-
Very odd that auto which is default ramped up the voltages to cause the heat up as I stated as a possibility...
Also I'd set everything in BIOS back to defaults before enabling cool'n quiet and this way nothing is old config from the prior CPU. If your bios somehow has the voltages ramped up it could run it hotter than normal.
When I last dealt with this condition it was an ABIT motherboard with a socket 775 upgraded from a overclocked Pentium 4 3.0Ghz HT to a Pentium D. I assumed that the voltages would go back to normal with CPU swap. But it over voltaged the Pentium D causing it to run 80+ C even with cooling fan at full tilt and PC Speaker was sounding like a European Police Car as it climbed towards 90C the shut down temp. On my motherboard it didnt clear the overclock config in BIOS and I guess the Pentium D was too close of a family change for the BIOS to revert back to defaults. But your default that is set to auto in which auto cooks the CPU that doesnt sound right, but it appeared to be the fix. :-\
-
Are you able to check the power supply voltages with a multi meter. I have had over heating problems with a power supply either being noisy or to high in voltage, to check power supply noise you would need an oscilloscope the noise should be less than 100 mv . The power supply output should be as close to 5 and 12 volts but no more than 5.2 volts and 12.4 volts.