How do you test voltages on your motherboard specifically? Thank you for your help!

[KopeAcetic]

Hello everyone,

I've been having issues for a while and done multiple things to try to figure it out. My computer used to reset quite a bit, I eventually noticed the CPU was running hot and blew out the heatsink of the cpu and the gpu with condensed air (they were caked with some dust) which stopped the constant resetting, but this caused some issues such as in a game I was playing when I get to a character select screen the graphics look hyper accelerated, and it won't actually let me choose to enter the game, the buttons indent like I'm clicking on them but nothing happens. Now I can't really stress my system much (Prime 95 locks up my system after the cpu gets to about 75 degrees, but I can run 3dMark 11 demo, cpu goes between 70 and 74 degrees). Also when running 3dmark 11 the graphics look fine, it gets down to 9 fps on a couple of areas but it actually handles things quite well surprisingly. One kind person on these forums has been going back and forth with me a bit and given me a few tips that I needed to check out.

Below is my HWMonitor capture after running 3dMark 11 demo:

The kind person on these boards said one of the voltages looked like it was bad, and the motherboard sensor wasn't calibrated? So I needed to run a voltage monitor to check it out. I went out and purchased a voltage meter today . After reading up I assume it's the 3.3v that is a bit off? I'm a bit lost to be honest.

How exactly do I test what needs to be tested? Is it a specific capacitor on the motherboard? I've checked the manual but I can't find the specific capacitors I'm looking for...if that's even what I"m looking for. I'm assuming I either have an issue with my motherboard or psu. Another issue I've been having is sometimes after going into my computer the graphics card isn't being detected. I have to re-seat it a few times and try to unplug and replug the wires back into the graphics card a few times to get it working (I've tried different PCIE slots, same thing happens).

System specs:

CPU - AMD FX 8350
MOBO - ASUS Sabertooth 990fx r2.0
Vid Card - Nvidia GTX460 (this is changing soon to a gtx 970)
Audio - ASUS Xonar Essence STX
Ram- G.SKILL Ripjaws X Series 8GB X2
Seagate 2tb HDD
Samsung 840 pro 256 GB SSD
Case - Antec 900

I'm sorry to be a bother 🙁 thank you for any help you can provide.

 

[InvalidError]

The closest thing to measuring voltages on the motherboard would be to back-probe the 20/24-pin ATX and 4/8-pin ATX12V/EPS connectors to see what the voltages measure at there with a multimeter. Even $10 multimeters from hobby shops are usually accurate within 1-2% and should be good enough to check for gross out-of-tolerance issues.

What is more likely to be problematic is transient response and other semi-random events and for those, you would need either an oscilloscope or at least a multimeter with fast min/max logging to see how clean the supply rails are.

 

[Dark Lord of Tech]

HWMonitor is poor use HWINFO or AIDA64.

 

[larkspur]

You need to keep the stress test MAX temps under 70c on your 8350. Your chip is running too hot - that's at least one reason why you were getting resets and lockups. So after you blew the dust out you are still getting 70c and more in 3dMark11 - that's still too hot. What's your ambient (room) temp? You should remove the cpu cooler, clean off the paste, and apply new paste. That being said, if I was running an FX8350 I would use an aftermarket heatsink like a Coolermaster 212 EVO or better. Also make sure your case fans are blowing enough fresh cool air into your case.

I doubt your 3.3v rail has anything to do with it. But what exactly is your PSU make/model - cheap PSUs cause all sorts of problems.

 

[KopeAcetic]

HWMonitor is poor use HWINFO or AIDA64.

Not sure if this would help more 🙁. I'm really new to monitoring all this information, and I literally just downloaded HWM a couple days ago and got used to it so I'm a bit lost with the new program and all the new info, maybe this will help.

 

[KopeAcetic]

You need to keep the stress test MAX temps under 70c on your 8350. Your chip is running too hot - that's at least one reason why you were getting resets and lockups. So after you blew the dust out you are still getting 70c and more in 3dMark11 - that's still too hot. What's your ambient (room) temp? You should remove the cpu cooler, clean off the paste, and apply new paste. That being said, if I was running an FX8350 I would use an aftermarket heatsink like a Coolermaster 212 EVO or better. Also make sure your case fans are blowing enough fresh cool air into your case.

I doubt your 3.3v rail has anything to do with it. But what exactly is your PSU make/model - cheap PSUs cause all sorts of problems.

How do I keep stress tests under a certain temp?

I actually just purchased a coolermaster hyper D92 (I looked at the 212 EVO and it just seemed too huge to me). I should get it tomorrow and be able to install it tomorrow.

My ambient temp is probably about 15-18 degrees C.

I have a corsair GS700 power supply (bronze rating). I think it's actually quite beefy for my system (I brought it from my previous system I built about 5 years ago).

So, you don't think the rails are bad? Could it just be the cpu overheating? I'm a bit confused as to the graphical glitches I am getting. I can play other games fine, and I"ve uninstalled and reinstalled that game to no avail 🙁.

 

[larkspur]

Your cpu should be much cooler with the D92. To lower your temps during stress testing and any other use, you simply need better cooling which you have wisely acquired and will have tomorrow. Technically you may be able to undervolt the CPU to reduce the temps, but that requires some knowledge and you probably don't want to go to the trouble. Just install that D92 and things should be better.

Also keep in mind that PSUs don't last forever and we can't rule out that yours isn't beginning to wear-out. That being said, your problems appear to be more related to overheating.

I also want to point out that if we assume that the HWMonitor 3.3v value you listed is correct, then your 3.3v rail is still within the 5% tolerance established by the ATX standard. In other words 3.147v on the 3.3v rail is still within ATX specifications. But as InvalidError mentioned, the voltage reading isn't the only important measurement of a PSU's stability - but you'd need more expensive equipment and knowledge to fully test your PSU.

Try that new cooler - make sure to clean off ALL the old paste off the cpu using rubbing alcohol (as pure as possible). Denatured alcohol works too.

The game you are having problems with might just be a bug or maybe you need to update your GPU driver. If it only happens in that game then it may just be a driver issue or a bug. Also make sure to watch the temps on your GPU too! GPU-z is a nice little utility that provides info and logging of values specifically for the GPU.

 

[KopeAcetic]

Your cpu should be much cooler with the D92. To lower your temps during stress testing and any other use, you simply need better cooling which you have wisely acquired and will have tomorrow. Technically you may be able to undervolt the CPU to reduce the temps, but that requires some knowledge and you probably don't want to go to the trouble. Just install that D92 and things should be better.

Also keep in mind that PSUs don't last forever and we can't rule out that yours isn't beginning to wear-out. That being said, your problems appear to be more related to overheating.

I also want to point out that if we assume that the HWMonitor 3.3v value you listed is correct, then your 3.3v rail is still within the 5% tolerance established by the ATX standard. In other words 3.147v on the 3.3v rail is still within ATX specifications. But as InvalidError mentioned, the voltage reading isn't the only important measurement of a PSU's stability - but you'd need more expensive equipment and knowledge to fully test your PSU.

Try that new cooler - make sure to clean off ALL the old paste off the cpu using rubbing alcohol (as pure as possible). Denatured alcohol works too.

The game you are having problems with might just be a bug or maybe you need to update your GPU driver. If it only happens in that game then it may just be a driver issue or a bug. Also make sure to watch the temps on your GPU too! GPU-z is a nice little utility that provides info and logging of values specifically for the GPU.

Thanks a lot Larkspur 😀.

I definitely don't have the knowledge and/or expertise to undervolt the cpu or test the psu without some serious hand holding, I'm even scared about taking out the stock heatsink mounting unit and putting in the other one (probably watched 7 or 8 videos of people doing it with a few different coolers), but I think it's going to help quite a bit .

I was actually a bit scared that I damaged something when blowing out the dust from the gpu and cpu. I didn't hold the fans in place when I blew the air in, I read somewhere online you can damage them by letting them spin like that while cleaning them. Also I wasn't getting the graphical glitches on that game before I did that, they started immediately after I did that. I guess I'll see tomorrow (I'm also getting a gtx 970 with the cooler tomorrow).

If I continue to have issues I would assume it's the PSU or the Motherboard having issues 🙁.

I'll come back tomorrow (or the day after depending on how long it takes me to put everything in) and give an update .

I also need to learn better cable management. The case I have has a lot of fans, but I'm not sure how much airflow there are from the 2 front ones because of my bad cable management 🙁.

 

[larkspur]

I think you are going to like your new GTX 970 and D92 a lot! Even though your GTX460 is also Nvidia, you should still uninstall the old drivers and install the newest - a clean install is always best. Here's an article about applying thermal paste (and yes your D92 comes with some): http://www.tomshardware.com/reviews/thermal-paste-heat-sink-heat-spreader,3600.html

I've always used the 'bead' method as its very simple and effective. Good luck, go slow, don't get frustrated and 'force' anything, and have fun!

 

[KopeAcetic]

I have one other quick dumb question:

I originally thought I damaged my cpu with it overheating. I was wondering if there was any way to tell when I take my heatsink off if the cpu is damaged at all? I actually purchased a CPU and I'll be getting another one with my package today...just wondering if I can figure out if I can make sure my current cpu isn't damaged before I return the other one.

 

[InvalidError]

I was wondering if there was any way to tell when I take my heatsink off if the cpu is damaged at all?

Unless you have X-ray vision with 10 000X magnification to see through the heat-spreader and examine internal circuitry for anomalies, you are not going to be able to see any damage unless something went horribly wrong and burned a hole somewhere on the CPU's PCB.

The next best thing to X-ray microscopy would be to simply use a burn-in test and see if you get any mathematical accuracy errors or other issues.

 

[larkspur]

just wondering if I can figure out if I can make sure my current cpu isn't damaged before I return the other one.

Don't open the new cpu - leave all packaging intact. Install your new cooler and run prime95. Watch the temps while running prime95 - if you installed the new cooler properly then the temps will stay below 70c. Prime95 should be able to run indefinitely (like forever) without crashing but running it for a few hours should be enough. That is the best way to test your cpu. It is very unlikely that your cpu is damaged - usually a damaged cpu won't boot anything at all.

 

[KopeAcetic]

Thank you for your responses, sorry I didn't respond earlier, I was at work all day so I just had the time to check the thread, not enough time to respond adequately.

So, I was able to install the new graphics card (holy cow is it a beast, I had to take out part of my case so it would fit).

I also installed the coolermaster d92 which definitely seemed to lower the temperature of the cpu. I'm running Prime 95 right now with a blended test, it's been going for about 20 mins and hasn't crashed (progress yay!). With HWinfo the Vcores are the hottest part of the CPU it seems, there are also a couple of cores that are getting into the 50's which concerns me a bit.

Unfortunately the 1 game still has that graphical glitch, I'm not at all sure how to fix it 🙁.

So far it seems like all of the cores have passed all of the prime95 tests, is it obvious when something fails?

I want to reiterate, thank you very much for all your assistance Larkspur and InvalidError!

 

[KopeAcetic]

Here is after about 2 hours (I think a little over, I started hwinfo about 20 mins after I started prime 95)

Seems pretty good, is there anything in there that I"m missing that looks bad? I'm not 100% sure how to read hwinfo, but it looks like prime95 passed everything. I noticed a few of the workers did much less work than the others (lazy scrubs?)

When I installed the cpu cooler I couldn't get it to screw down with a manual screwdriver, so I used an electric one until it couldn't tighten it down anymore, hopefully it isn't too tight...I assume that's the way you're supposed to do it since it was ridiculously hard to get it manually.