High TMPIN1 temperature? Weird smell coming from my case? Bad PSU?

What is the exact model number of your motherboard and power supply?

There are many 760G chipset MSI motherboards, all of which have different specs, quality, power phase and capabilities.

Knowing your labeled specs for your PSU if it's a low quality unit doesn't help, as low quality units can rarely if ever meet those listed specifications, so the model number becomes extremely relevant in determining if this is likely to be a relevant issue or not.

EXTEMELY terrible power supply, not just bad, but probably one of the worst Great Wall platforms they've ever used.

Also, there is no MSI 7641 motherboard model. The following MSI models with the 760G chipset are known.

MSI 760GMA-P34(FX)
MSI 760GM-P23 (FX)
MSI 760GM-P34(FX)
MSI 760GM-P21 (FX)
MSI 760GM-P34 (FX)
MSI 760GM-E51
MSI 760GM-P33
MSI 760GM-P35
MSI 760GM-E51 (FX)

Motherboard model should be printed directly on the board itself, or you can install and run CPU-Z which should list the motherboard model.

No, I am not misreading anything. That might be the part number of the board, but it's not the model number of the board. GA-78LMT-S2 IS the model number. The only model number that matters as you won't find any information about that board using MS 7641.

The power supply IS that bad, and refusing to understand that, no matter whether it's been working for two years or not, won't help you to solve your issue. As I said before, the PSU BEING a lousy model might not be THE problem, but it is A problem, no matter what. There are still people with running Ford Pintos, but it doesn't change the fact that 90% of them were dead and in the salvage yards within five years of being on the roads due to engine, power train or safety issues. They were lemons. Throw away cars. So are a lot of power supplies, including that one.

Here is What Jonny Guru says about that unit:

http://www.jonnyguru.com/forums/archive/index.php/t-8398.html


The fact that it powers on the system, and the system runs, does NOT mean it is adequately, reliably, powering that system in a healthy fashion OR that damage isn't happening, or already happened, to some or all of the hardware that is connected to it. Poor voltage regulation, high ripple, bad transient response, noise, etc., are all factors that can damage hardware over time, or in some cases, instantly, and you'd never know the power supply was to blame without expensive testing equipment. Which is why we rely on the opinions and reviews done by those who have this equipment and can perform these kinds of tests.

You can ignore the fact that your hand is cut off, but you will eventually either bleed to death or die of infection, whether you admit the hand is gone or not. Same thing here. Ok, enough on that subject, you already know it's a poor unit.

Just as important is probably the fact that your motherboard is extremely poor quality. I understand you have limited availability on hardware where you are, but it doesn't change the fact that the GA-78LMT-S2 is very poor quality and isn't fit to be paired with a 6 or 8 core FX chip.

Tier Two: Poor quality. Bad power phases and no heatsinks on VRM make this tier more like ‘lowest’ tier for AM3+ boards, not advisable for FX 6 or 8. No Crossfire/ SLI capability. In all probability, you’d not want to have them unless you must get one of them.

M5A78L-M LX PLUS
GA-78LMT-S2P
GA-78LMT-S2PV
GA-78LMT-S2
760GM-P23 (FX)
960GM/U3S3 FX
760GM-P34(FX)
880GM-LE FX
980DE3/U3S3
760GMA-P34(FX)
M5A78L-M LX
M5A78L- LE

The combination of poor motherboard and poor PSU are almost certainly cumulatively responsible for the problems you are having and that board is probably getting hot because it's power phase is not designed for use with a high TDP chip like yours, plus it has no heatsinks on the VRMs. Additionally, it's entirely possible that due to the low quality and age of your power supply, the poor voltage regulation and ripple are adding to the problem. I'd replace both, however, in the meantime, download and run HWinfo, NOT HWmonitor, and run "sensors only". Take screenshots of all the sensor values, which will take two or three screenshots to capture, and post them here. I'll take a look and see if it seems there is anything to be immediately concerned with.

HWinfo: http://www.hwinfo.com/download.php


HWmonitor and Open hardware monitor are both unreliable in my experience.

Electrical "noise", not audible noise:

https://en.wikipedia.org/wiki/Noise_%28electronics%29

Electrical "noise", not audible noise:

https://en.wikipedia.org/wiki/Noise_%28electronics%29

Temps look ok in those screenshots. Not seeing the 5v sensor readings from the PSU there though. Could be any number of reasons for that. Probably a problem with the 5v rail.

If you can, take another series of screenshots with the system under a load. Run a stress test like Prime95 on Small FFT, use version 26.6, and take your HWinfo screenshots WHILE it's running.


At this point I'm going to venture a guess and say it was either a fluke, or maybe a cap is leaking in the PSU and what you smelled was something in there burning off. Could be many reasons. I'd get a quality power supply and go from there, even if everything else SEEMS ok at this point.

Your configuration sounds like it's either pooched, or the system is entirely unstable as it can't handle that CPU. Could just be HWinfo, but not being able to run Prime95 is a rather dead giveaway. Try resetting your bios to default settings and then try it again. Maybe use CoreTemp instead. If your CPU is running at 80°C under load, then you are either drastically lacking in cooling, the heatsink isn't seated correctly or you have way too much voltage. Resetting to default bios settings should take care of the voltage, and ANY CPU should be able to run Prime95 v26.6 at it's stock settings.

If it won't RUN prime95 at stock settings, then there's a problem. Probably either PSU or motherboard. Doubtful that anything would be wrong with the CPU itself.


Might also try monitoring CPU temp using AMD Overdrive, and watch the CPU tab. That will show thermal margin, rather than any actual temp, as AMD chipsets use a different method of determining thermal values than Intel. That distance will not be the actual temp, but will show how far you have left before it's "too hot". If thermal margin is more than 10°C, then core temps are ok, but motherboard temps OR VRM throttling can still be an issue even when the CPU is staying within specs.

CPU 0 is core temp. CPU package is socket temp. Both are separate from motherboard/VRM temp. CPU 0, CPU package and motherboard are the only ones I'd be worried about. Like I said, I'd run AMD overdrive and if your CPU thermal margin under load is more than 10°C, I wouldn't be too worried. Chances are good, since that motherboard has no VRM heatsinks, that motherboard temps are going to be rather high at stock speeds, especially if turbo is enabled, when under a sustained load, but not much you can do about that until you get a better board.

Are you using the stock cooler or an aftermarket cooler. If it's an aftermarket cooler, what is the model? What is your case model and how many case fans are installed, and in what orientation are each of them?

That's totally wrong on the fan configuration, but before I go further on that you still didn't say what case model you have? Do you have a top or bottom mounted PSU in your case configuration?

As far as board models go, the UD3 would be fine for your CPU, as would any of these would be fine for your CPU:



970 Performance
970 GAMING
970A SLI Krait (USB 3.1 supported)
GA-970A-UD3P
M5A97 or EVO or PRO (R.2 as well)
GA-970A-UD3
GA-78LMT-USB3
Extreme 3 R2.0
M5A99X EVO (R2.0 as well)
GA-990XA-UD3
990XA-GD55
Extreme6 (Termed as 990X chipset board on some sites)
Extreme9
Fatal1ty 990FX Professional
Crosshair V Formula-Z
Sabertooth 990FX R2.0
GA-99FXA-UD5
MSI GD80V2 (See NOTE1 below)
M5A99FX PRO R2.0
GA-99FXA-UD3
GA-990FXA-UD7
MSI GD65V2 (See NOTE1 below)
990FX Killer
990FX Extreme4
990FX Extreme3


The UD3 would be fine for that chip, even for some overclocking. Personally though, MYSELF, I think the PSU is the larger issue and should be addressed first. And although there are a lot of PSUs that would be better than your current one, by far, but would still not be good choices as far as quality and reliability, I'd try to make sure you get one that's either Tier 1 or 2 and is at least 550w.

http://www.tomshardware.com/forum/id-2547993/psu-tier-list.html


Tier 3 units are probably MUCH better than what you currently have, but are not likely to be good investments and probably won't offer you a long lifespan when paired with a 125w CPU and a high end discreet graphics card like your 380x.

Take the RAM back out, and insert it lightly in the slot with both ends equally down in the slot, but not far enough where the locks "click" into place. Then, using one hand to hold each end of the module, push straight down on one end until it clicks into place while making sure the other end does not "raise" back up. Then push the other end down until it clicks into place. Make absolutely certain that the cutout on the bottom of the memory module lines up with the "key" in the memory slot.

Sorry you don't feel my advice is helpful. I guess you don't need my help, you seem to think you have it well in hand so more power to you. Good luck.

Also, the CM B units are lousy and use 85°C caps and seem to use a low end CWT platform.