OC ASRock Z97 Extreme6 - Question

chiptouz

Distinguished
Sep 12, 2006
194
0
18,690
Does this make sense? I was, what I thought, stable at 4.6ghz at 1.234 volts. I added a second fan to my coolermaster 212 evo and I have had to add 0.004v to become stable (still testing). Does this make sense, or was my initial conceived oc not stable, like I thought. I am using P95 26.6 to test. All tests using Aida64 are fine. at 1.234v tested stable for 3+ hours then I turned off waiting for the second fan to arrive. Specs are in sig below.

Thanks,

Chip

 
Solution
I highly recommend the Noctua NH-U14S, Noctua NH-D14, Noctua NH-D15, Scythe Mugen 4, Phanteks TC14PE, Cryorig ultimate or universal, Thermalright Silver arrow, Thermalright Macho rev.B, any of the Corsair, Cooler Master, NZXT or Swiftech 240mm or larger liquid coolers.

For the best price versus the size of the cooler and the performance you get out of it for the price, I like the Noctua NH-U14S which is why I use it on my own rig. It's not so huge that you can't fit anything else in the case and it handles large overclocks admirably well without making a ton of noise.
Sounds like adding the second fan pushed the voltage vdroop beyond what the cpu or motherboard could tolerate. Most likely due to either a poor PSU or not having the LLC enabled or set high enough. Could also have been that it wasn't actually stable before or was borderline. 3 hours of Prime is not enough, not by far, to establish stability. 17hrs minimum and most preferably 24hrs. I've seen rigs stable as can be up to 17hrs and then error out between 17 and 24 hrs. If it can't run 24hrs on Prime small FFT's, it's not stable IMO.

What is your PSU brand and model number. This will tell us a lot.
 

chiptouz

Distinguished
Sep 12, 2006
194
0
18,690
Darkbreeze - Thanks for the response. Here is my PSU: Corsair PSW750XT. I will definitely try to run p95 for 24 hours straight. I am now at 1.240 volts and have run it for over 22 hours, but not straight. I went 12hours, upped my FSB 0.1 and then another 10. I will try to run it for 24 hours straight. Wouldn't think it was my psu, but it was something else. with the additional voltage it seems ok, for the moment.
 
Ok, so what you posted above is not the model number. The model number is CMPSU-750TX which makes it a Tier 2A unit. Pretty damn good power supply but if yours is old it could still be an issue of the PSU being "tired". Honestly it's not highly likely, but it is highly POSSIBLE. Can you download and run HWinfo and post screenshots of the voltages in the sensors window both at idle and under load?
 
PSU voltages look ok, very good in fact. I am somewhat concerned though that it shows only .952v on the CPU core voltage. This is very likely to cause instability at that OC. In your OP you said it was at 1.234v. Did you change it to drop it down or is it possible that some further BIOS settings are needing to be changed, like disabling speed step or C states, as the system may be dropping the voltage down. Weird though that it's showing that voltage even at load though. You may need to address this if it's not something you did intentionally.

Also, your temps seem pretty high at load and perhaps the system is throttling due to exceeding the thermal limits. You don't have to get to shut down temps in order for there to be an adverse affect as the cpu will throttle to stop itself from getting to that point. Perhaps you can post screenshots of the BIOS window to show all settings. Flash drive in usb port and hit F12 will generally take a screenshot of the current BIOS window on most UEFI motherboards.
 

chiptouz

Distinguished
Sep 12, 2006
194
0
18,690


Darkbreeze,

For some reason CPUID HW monitor shows something different for CPU VCORE that what is shown in the VID votages under the Intel Core i7 4790 section. Those numbers match what cpuz shows as my core voltage 1.243. I will take some pics of the uefi and post.
 
Ok, I typed this whole fracking thing once already, and then got the inevitable Tom's hardware fracking error page, so, here we go again.

Couple of things. Your CPU temps seem way high, high enough to cause throttling or worker errors.

The first place I'd look is the center pressure screw on your EVO cooler. In case you're not aware, which most people who use this cooler are not, the center screw applies all the pressure to firmly seat the heatsink base against the CPU lid. The four outer screws do not, they are mainly just there to leverage the center screw fulcrum point. There is a loose and a tight position. The bottom half of the center screw has a slot in it which is supposed to catch on a matching raised slot in the center hole of the heatsink base so you can finger tighten the center screw.

The tab in the hole sometimes is not there or breaks off. It's only aluminum after all and not that sturdy. My recommendation is, if you are not 110% certain that you were previously aware of the center screw being adjustable, and did so, correctly, you remove the entire heatsink, clean the cpu lid and heatsink base, familiarize yourself with how the center screw works, repaste and reinstall the heatsink.

You can hold the bottom half of the screw in place to tighten the upper half into it, effectively getting it fully tight prior to installing the hold down bracket, but it makes it much harder to get the four outer screws in place and started.

Alternatively, if you can just loosen the four outer screws about halfway, press down on the center screw to hold the bottom half in place against the bottom of the bore in the center of the heatsink and tighten the upper half of the center screw fully in, then re-tighten the four outer screws to spec, it would make it much faster. The center screw not being adjusted correctly can account for as much as a 10-20°C difference in temps.

The next thought is perhaps too much TIM was applied to the bottom of the heatsink, or not enough. A rice grain sized amount is about perfect in most cases, applied directly to the center of the cpu lid prior to mounting the cooler.

Lastly, case cooling. Good case cooling requires a significant volume of cool air be replenished every second in order to allow the CPU and GPU fans to do their jobs. Otherwise, you're just trying to cool the CPU, or GPU, with warm air, which doesn't work well. If you don't have good internal case cooling then the CPU heatsink cannot exchange heat well which leads, once again, to high temps. All case fan locations should be populated on any overclocked rig. Both intake and exhaust fans, preferably multiple of each, are necessary to accomplish this. If you have good case cooling then one of the first two items is likely the problem.

I think your voltages are all fine just going by the sensor values which isn't as accurate as using a meter but is accurate enough to not indicate any glaring issues.
 

chiptouz

Distinguished
Sep 12, 2006
194
0
18,690


Darkbreeze,

No I was not aware of the center screw and I hoped that I did not place too much TIM. I usually use about two grains of rice size on the cooler and spread it out as evenly as possible. I guess I used maybe a bit too much. I could re-apply to check. I have used this amount on all pc's I build and haven't had any issues, but if it is supposed to be only 1 grain...Do you just place it in the center and spread? With two grains I think it is thin, but maybe too thick. Any good links to pics of what it should look like would be great.

Same with the evo cooler center. I never knew about that. I would definitely like to take a look. Pics would be good too.

Thanks for all your help!!!!

Chip
 

chiptouz

Distinguished
Sep 12, 2006
194
0
18,690
Darkbreeze - I went and reset my cpu cooler, reapplied TIM (dot pea size method), and checked the center nob on the cooler (fully tightened it). All look good. CPU temps are comparable to before. I have several case fans. All of them are running full. two 200mm fans blowing out on top. 1 230mm fan blowing in. 1 120mm fan blowing out, 1 200mm fan blowing in. I have two 120mm fans in push pull on the cpu cooler. Maybe I just need better cpu air cooling. Recommend me the best air cpu cooler. I have a hafx case.
 
I'll be honest. I really hate the 212 EVO. I know it's highly recommended as a budget cooler, but seriously about two out of every ten people, if not more, that come here and post about temp issues when overclocking have an EVO cooler and when they install a higher end model usually the troubles disappear. I used a Hyper 212 initially with the build I have now and had temperature issues as well that were immediately resolved when I installed a Noctua NH-U14S. I believe the problem to be 100% related to the poor mounting system, not necessarily the cooler.

It's really a badly designed mounting bracket and at least ten times now I've had people with issues that were resolved by finally getting the center mounting screw fully tightened prior to tightening the four outer screws. It shouldn't BE that difficult though, and IMO is a poor design. But it does work for a lot of people and I've used it successfully on many builds as well. Perhaps it's as things often are with budget products and not all are equally efficient for whatever reasons like poor tolerances here and there. Pinholes in the solder of the heat pipes could easily allow the coolant inside to escape. It could be anything but if it's not the paste job and it's not the center screw, and you've got good case cooling with all fans facing the proper directions, then it's either just not a good chip or the cooler isn't sufficient for the task.
 
I highly recommend the Noctua NH-U14S, Noctua NH-D14, Noctua NH-D15, Scythe Mugen 4, Phanteks TC14PE, Cryorig ultimate or universal, Thermalright Silver arrow, Thermalright Macho rev.B, any of the Corsair, Cooler Master, NZXT or Swiftech 240mm or larger liquid coolers.

For the best price versus the size of the cooler and the performance you get out of it for the price, I like the Noctua NH-U14S which is why I use it on my own rig. It's not so huge that you can't fit anything else in the case and it handles large overclocks admirably well without making a ton of noise.
 
Solution

chiptouz

Distinguished
Sep 12, 2006
194
0
18,690
Darkbreeze-I am going to go the NH-D15 route. What TIM material do you use? I have some prolimatech PK-1, but am thinking about buying some Gelid gc extreme. Not sure it is worth it for 0.7K decrease. Your thoughts?

Edit: Was looking at more reviews/comparisons and the cooler you recommended actually cools about a degree better with two fans than the d15. It would be about $10 cheaper with your recommendation with two fans.

Edit 2: Now not sure what I am going to do...
 
That degree, or even five degrees, probably doesn't matter in the least. You will never do anything with your computer, for 99% of users, that will ever, ever, even come close to stressing the CPU with a load even remotely similar to what a full Prime run does. During very intense gaming or rendering, conversion of raw audio or video, you might run the numbers up a bit, but it won't be a full time 100% load and your rig will be able to handle it with any good cooler.

The TIM doesn't really make that much difference either. It's all propaganda. From the absolute best thermal material you can buy for any amount of money, to the average stuff that comes with most coolers, you're looking at about a 5°C difference, for most configurations under laboratory conditions. Honestly, I've used the liquid metal TIM and Arctic Silver, both, on the same rig, and only saw a 3°C difference. Clearly it wasn't worth the 16 bucks including shipping that was spent on it. Plus, the margin for error when applying high end thermal interface materials is generally higher than using a mid-grade product as it's much harder to get a perfect application.

I'd just use the prolimatech you already have or some Arctic Silver, or the paste that comes with Noctua coolers is pretty fair too.
 

chiptouz

Distinguished
Sep 12, 2006
194
0
18,690
Darkbreeze,

the cooler seemed to be it. With the noctua NH-U14S my max temps seem to be a couple of degrees c lower with the max being 80 now on prime 95 26.6. I am much more happy with this cooler than the EVO. Great recommendation.

Thanks,

Chip
 

chiptouz

Distinguished
Sep 12, 2006
194
0
18,690
Darkbreeze - Got another one for you. Ok, so when I run prime95 26.6 no issues, no blue screens, etc. While prime 95 is running and I surf the web, within a few minutes I get a blue screen with the 0x124 error, which I believe is that the CPU isn't getting enough voltage. Should I not do anything else while prime95 is running? I have a feeling that I could run prime95 for days and it will be stable, but for some reason when I run prime95 and then do something else concurrently I have an issue.
 
Do you have two fans on the CPU again, as you did before? Try taking the rear fan off and using only the fan that came with the unit on the front, in push configuration. I think your temps are still a little bit high. Then again, it might be ok. What is your ambient temperature where you have the computer?
 

TRENDING THREADS