cancel
Showing results for 
Search instead for 
Did you mean: 

Maximus VII Hero throttling issues, please help

rwchiz
Level 7
So, I've been having run away heat issues since putting together my Maximus VII Hero and i7-4790K rig a month ago. I originally had a H80i which would cause IBT to go to 99C within two seconds and OCCT to fail within minutes. Then tried my old CM Hyper 212+ with similar results. I just purchased a Swiftech H220-X with no marked improvement. All this is by running optimized default settings in BIOS 2012 and only changing from AHCI to RAID.

The only way to improve run away temps is by using a manual voltage which helps but then my voltage stays high when even though multi drops. Tried an offset which seemed to work but would BSOD under load transistions and gaming.

I then proceeded to disable the Asus core enhancement and set it to act like a stock 4790K with varying boost bins. I checked Intel XTU and it had the default Core TDP set to 88W which was a good sign. I ran the XTU stability test and temps were good, around 75C. However, I noticed the CPU never got to default of 4.2GHz fully loaded, only went to 4.1GHz (40x1). I then noticed that the CPU was constantly being current throttled. It was not being thermal or power throttled, only current throttled.

So here's the weird part. In XTU, if I set processor current limit to 255A or less, it will always throttle due to current when stress testing and temps stay under 82C but never maintain max frequency. If I set the current limit to 256A or higher, I then get run away thermals within a second or two so I have thermal throttling. The only way I can get it to power throttle is if I set turbo boost power max to something stupid low like 40W. Vcore peaks at 1.223V until throttled to 1.1V. Only time voltage gets to 1.289V is under Prime or Aida64 FPU.

I have updated BIOS to 2201 with no change. Does anyone else have this problem or a solution. Seems weird that setting current limit to anything more than 255A just seems to open the flood gate but 255A or less slams it shut. I'd love to be able to get my rig thermally stable at 4.4GHz so I can eventually run 4.6GHz as my daily driver. Currently running the optimized default, I record peak temps of 78C when playing BF4 with GTX 970 and cooled with a H220-X which seems too high to me. It would be one thing if I was running cheap air cooling but the H220-X is supposed to work well.
7,022 Views
9 REPLIES 9

NemesisChild
Level 12
Welcome to the forum!

What case are you using and tell us about your case fan configuration.
What TIM are you using and how are you applying it?
In the Control Panel, under Power Options, set it to performance and see how it performs.
Intel i9 10850K@ 5.3GHz
ASUS ROG Strix Z490-E
Corsair H115i Pro XT
G.Skill TridentZ@ 3600MHz CL14 2x16GB
EVGA RTX 3090 Ti FWT3 Ultra
OS: WD Black SN850 1TB NVMe M.2
Storage: WD Blue SN550 2TB NVMe M.2
EVGA SuperNova 1200 P2
ASUS ROG Strix Helios GX601

NemesisChild wrote:
Welcome to the forum!

What case are you using and tell us about your case fan configuration.
What TIM are you using and how are you applying it?
In the Control Panel, under Power Options, set it to performance and see how it performs.


Case is NZXT H440. Three 120mm stock NZXT fans in front for intake, 140mm stock NZXT fan in back for exhaust, Swiftech H220-X with 2 Swiftech 120mm fans exhausting at top. Also tried with two 140mm fans on top as exhaust when using the Hyper 212+ and H80i. When I was using H80i it was as intake. All thermal testing is done without the restrictive top and front panel of my case on.

Currently using Swiftech Tim-Mate 2 with pea sized amount. Previously tried Thermalright TIM in pea sized and IC 7 diamond with lentil sized.

Currently using performance mode with 5% min and 100%max cpu.

Windows 8.1 Pro 64

Backfire1
Level 7
Reset your bios. Hit F10 save and exit.
Go back into your bios and set your xmp profile. Hit F10 and exit.
Use AID64 to stress your system and watch your volts and temps and see what happens.
If you are using prime 95 to stress your system you will get temp and throttle issues.
Let us know how things turn out for you. Also here is a good read. http://www.overclockers.com/3step-guide-to-overclock-intel-haswell

Backfire1 wrote:
Reset your bios. Hit F10 save and exit.
Go back into your bios and set your xmp profile. Hit F10 and exit.
Use AID64 to stress your system and watch your volts and temps and see what happens.
If you are using prime 95 to stress your system you will get temp and throttle issues.
Let us know how things turn out for you. Also here is a good read. http://www.overclockers.com/3step-guide-to-overclock-intel-haswell


Thanks for the suggestions.

I have performed clear CMOS and reset my BIOS to optimized settings with F5. Ram with XMP profile (9CAS) and without (11CAS) 1.5V. If I run default BIOS settings max current limits and power limits are stupid high. Prime 26.6 will get into low 90'sC but be stable for at least 5 minutes, voltage 1.223V. More recent Prime shoots up to 95C+ within a second at 1.289V. Aida64 with CPU, cache and FPU checked off gets up to 99C when the FPU portion is run and voltage gets to 1.289V. If I run just CPU in Aida64 temps get to low 90C and voltage 1.223.

I am focusing on Intel XTU since its sanctioned by Intel and shows a lot of useful info. I think the root of the problem is why this mobo acts one way with current limit set to 255A and completely different when set to 256A. There has to be some happy place where I slowly creep to a throttle point whether thermal, power, or current. As it is, if I set current limit to 256A, it hits the thermal wall. If I set current limit to 255A, it hits a current limit. There is no gray area. I've read everything I can about this issue and Intel's answer is to use XTU and set power and current limits which is BS.

I'm almost thinking of throwing in the towel and changing mobo but I hate to have to buy a new one and buy a new Windows license.

jab383
Level 13
Hi rwchiz, and welcome

Without delidding, with MX-4 TIM under a Hyper 212 EVO, I get 90C at 117 watts in OCCT (using Aida64 to monitor during OCCT tests) without throttling. With water cooling, that drops to 77C. Temperatures are similar in XTU without throttling. Your 4790K is a better overclocker than mine and should be getting you into a good frequency range.

With core clock at 4.8GHz plus a bclk push and Vcore of 1.50, I get 83-85C in XTU. Of course, that's with a delidded CPU in which I replaced the Intel **** with liquid metal.

I've never used anything but default power and current limits, which are essentially unlimited. XTU says 4096 amps.

I truely think you and the motherboard are fighting a bad CPU. I've seen some pretty bad applications of TIM internal to the Haswell CPUs, but this may just be the worst. The TIM between the silicon and the heat spreader (IHS) is poor on 4770K and is supposed to be better on 4790K. Better doesn't mean good. Moreover, I'm thinking that this one 4790K has a particularly bad problem - missing TIM or large air gap over the silicon.

I'll suggest you pick one of two alternatives depending on how adventurous you're feeling:
1. Delid the little darlin' and replace the ****y TIM.
2. Contact Intel and RMA the CPU. They're under warranty for such things.

Jeff

Backfire1
Level 7
You definitely have a problem if you are running default settings @ stock speeds and you are hitting 99c. I have a h100 and only hit 100c when running 4.7 ghz & running prime. Seems very odd that Intel would release a flagship possessor without it passing certain test. Good luck and let us know how things work out.

rwchiz
Level 7
I am thinking it must be a mobo issue because I had a i5-4690k two weeks ago with similar but not quite as bad issues. I am regretting giving my son the trusty AsRoch Z77 Extreme 6 and i5-3570k rig I had previously. Just bought a cheap g3258 and MSI Z97 PC Mate combo from Microcenter for $99 so I can rule the CPU or mobo out.

Well, put the CPU in the MSI mobo and temps are so much lower. I actually have a system that works normal under load. I am power limit throttling like I should at 88W when stress testing. On the Hero, I only got to a maximum of 60W fully loaded.

When I was getting ready to box up the Hero to get ready for RMA, I did something stupid by trying to put the socket cover under the clamp like on previous mobos. Needless to say, some pins were bent. So now I am going to have to buy a new board since the MSI PC mate is not going to cut it.

Any suggestions on a new motherboard or should I get another Hero. I'd like it to be stable with a decent feature set and look good in a white NZXT H440. If I get a different Z97 with Intel NIC, will I be able to avoid reinstalling Windows?

Backfire1
Level 7
If you get another MB you will have to re install windows. I sometimes wish I would have purchased an ASROCK MB. Pic one you like and go with it. Not sure what is going on with ASUS lately but I had to return a Z97-a because it just quite working. I could not get it to post and I worked on it for 2 days. I have read were others are having stability problems. I came from a rock solid AMD 1090-t and Gigabyte MB. I have had a heaping amount of issues and days of stress. I thought it would be a good time to upgrade and to be honest I really can't tell that I have upgraded as far as day to day use. The performance difference is not that noticeable. I like my HERO but I guess I had higher expectations for stability and performance. I to had some issues with temps and volts. I am so tired of messing with it that I just run it at stock settings with XMP profile #1. Seems like anything else causes a blue screen or it gets to hot for every day use. I thought I had it figured out but it seems that I was mistaken. Good luck on your quest for a stable fast performer.