cancel
Showing results for 
Search instead for 
Did you mean: 

Problems stabilizing FX-9590 on Crosshair V Formula Z

2dcomplex
Level 7
Hey there folks! I understand many of you have been able to get the 9590 running just fine on the Crosshair V Formula Z, so I'm reaching out for some help.

My current configuration is as follows:

Motherboard Rev: 1.01
BIOS Rev: 2101

Power Supply:
Corsair AX1500i

CPU Cooler:
Corsair H110

Video:
SAPPHIRE R9 295x2
Driver: 14.10.1006-140417a-171099C

Memory:
G.Skill F3-1866C8-8GTX
Size: 8GBx2 (16GB total)

Disk:
SSD: Samsung 840 EVO MZ-7TE500BW

OS: Windows 8.1 64bit

BIOS Settings all default except the following:
Ai Overclock Tuner: D.O.C.P
DRAM O.C. Profile: Profile #1
Memory Frequency: DDR3-1866MHz
DRAM Voltage: 1.60000

While idling in Windows with the above configuration, the system temporarily pauses/freezes for up to 15 seconds every couple of minutes, and then mysteriously recovers. Upon further investigation, I found that these periodic freezes were related to Turbo Core.
It appears that when the system increases frequency to 5GHz, the system is likely to temporarily hang until it drops back down to 4.7GHz. I have tried the following BIOS settings to resolve this issue with no change in behavior or improvement:

Setting CPU voltage to 1.493750 with Load Line Calibration set to both High and Ultra High.
Setting CPU voltage to 1.50000 with Load Line Calibration also set to both High and Ultra High.

Disabling Turbo Core and leaving all other CPU and power related BIOS settings at defaults resolves the temporary hangs/pauses, however the system will not run Prime95 for more than an hour before the system hard locks.
During the Prime95 run, the CPU frequency bounced between 4.0GHz and 4.5GHz, and the temp hovered in the mid 50s, so I'm completely baffled as to why it would crater.

My last ditch effort to get stability was to set everything manually. Here's what I set the CPU settings in the BIOS to:

Turbo Core: Disabled
CPU Ratio: 23.5
CPU Manual Voltage: 1.493750
Load Line Calibration: High

With the above settings, the system is almost stable. It will run prime95 for three hours before a random thread or two will die with a rounding error, but the other threads continue running. Temps during this run did get a little on the hot side, peaking in the low 70s.

Given all that, I decided that the CPU or the PSU were probably bad, so I RMA'd both and received brand-shiny-new ones. Sadly, I'm having the exact same problems with the new 9590. Why me? 😞

One last, and highly interesting data point is that I obtained a new FX-8350 for comparison sake, and it runs perfectly stable at 4.7GHz 1.493750v with LLC set to high. I was able to get a 12 hour prime95 run in with all threads humming away perfectly, albeit slightly hot. During the 12 hour run, CPU temp would reach into the low 70s. I was not able to push it any higher than 4.7GHz without it crashing a few seconds or minutes after applying load.

How odd that the board freaks out when I run the 9590, but seems more than happy to run the 8350 at 4.7GHz.

So 9590 + Crosshair V Formula Z people, what in the name of all that's good am I doing wrong? Could some of you stable guys share your BIOS settings and/or ideas as to why this might be happening? I'm at a loss.
114,020 Views
136 REPLIES 136

2dcomplex
Level 7
Just for the heck of it, here's a picture showing how everything is connected up, physically anyway. 🙂

38219

Hey 2Dcomplex, just wondering I see where you are having the high heat issues, not sure if you still are but from what I can see in your pic it looks like you have your fans set to exhaust on the CPU and GPU Radiator Fans? If so try setting them to intake as pulling cool air from outside instead of pushing hot are through the already hot radiators you should get a big drop in temp as I did this when my Temps where that high on my old system.
Just checkin. Good Luck, Nice build almost as Purdy as mine 🙂

MeanMachine wrote:

Have you tried to increase CPU ratio to 25x at same voltage?


Just tried that at my current settings, and while windows boots and loads basic programs, applying any load at all results in an immediate hard lock.
I then increased the CPU voltage in one step increments 3 times, re-testing between each adjustment. All three attempts resulted in the same immediate lockups under load.

Lastly, I adjusted the CPU/NB voltage in one step increments 3 times, re-testing in between as before. Same lockups again. Bollocks.

chris635 wrote:
Not sure if this helps but.....for my system to be stable with my ram configuration I need about 1.3v on the cpu/nb with the llc for it set to extreme which will pump it to 1.38v depending on load


Wow, that's quite the jump. I can certainly give that a try if it's safe.


EdTheHead wrote:
Hey 2Dcomplex, just wondering I see where you are having the high heat issues, not sure if you still are but from what I can see in your pic it looks like you have your fans set to exhaust on the CPU and GPU Radiator Fans? If so try setting them to intake as pulling cool air from outside instead of pushing hot are through the already hot radiators you should get a big drop in temp as I did this when my Temps where that high on my old system.
Just checkin. Good Luck, Nice build almost as Purdy as mine 🙂


I was thinking about doing this, but I would then be faced with problems keeping the northbridge and VRM cool on the motherboard, as well as problems cooling the pcie bridge chip and power section of the 295x2.

The air that comes out of the 295x2's radiator is extremely hot when gaming or doing opencl work; far hotter than the CPU radiator in fact. When I first noticed the sheer volume and temp of the air coming from the GPU radiator, I likened it to a hairdryer. lol

If I configured both radiators to pull in air from the outside, it would be up to the two 140mm fans in the front of the case to exhaust the hot air.

Currently, under gaming load my temps hover in the mid to high 50s, and idle sits around 38~40c. Only if I really beat the hell out of the processor does it shoot as high as 70c, which is still technically within spec. With the case wide open and a floor fan blowing right into the case, the temps are just about the same. Maybe a degree or two lower at best.

Cool your good to go, just checking man 🙂 I thought I might be able to heap someone with there issues since I can't help myself with mine 😞 Good luck bud.

Hi, I'm having similar problems, and Asus tech support is not helping much at all. I'm kinda sad about this, since I dumped a bunch of money into this rebuild. Years ago their tech support would actually walk through the BIOS with the customer, doesn't look that way anymore, seems as though alot of them are reading from a help window on a computer and have little knowledge of the BIOS settings on this board. . In the course of one week, of on and off calling several times every few days nothing has progressed? I've built lesser systems with only one SSD and the boot time is almost non existent. This quad SSD raid 0 benchmarks with ASSSD close to 1,000 read/write. I can't imagine with those throughput scores that this would boot up slower then lesser machines? Also when the computer is doing something my mouse becomes lagy or looses responsiveness for a few seconds? I've read somewhere in these forums that auto settings do do so well with this board/CPU combo. I really wish tech support would stop trying to lead customers away from these issues and at lease try to help implement these SLIGHT voltage increases that people are saying in THIS FORUM are required for this CPU/mobo combo to function properly. So sad, the people that REALLY have the know how are hidden away from callers only to be reached through this method of communication, if that....

Several suggestions have been brought up including possibly a bad SSD in the bunch or an oddity with the Southbridge or mobo in general. Another tech mentioned that the onboard RAID controller will not work well with 4 SSD's in RAID 0 and it would be too much to handle. Funny part is that I've spoken with LSI tech support and they actually have a Sabertooth rig for testing purposes. They explained to me that 4 SSD's in RAID 0 will definitely not cause these kind of problems I'm experiencing and that there may be other issues to investigate. Again, I was told twice that in the course of a week that some higher up tech support would actually call me to help, but alas not even a smoke signal did I see 😞 Good grief what a let down...

If anyone can please help me figure this out that would be great....

Here's my setup:

Crosshair v formula z

FX 9590 stock

OCZ black edition 4x2 1600MHz 1.65v stock

H100i CPU cooler

HAF 932 case

RAIDMAX ultra gold 1200w PSU

4x PNY Optima 240GB raid 0

Windows 8.1 64-bit all updates

2x Sapphire HD 6950 2 GB x-fire stock.

i had some problems too... asus mb too... same beast processor... you need to set 1.5v on it, or it crash on heavy load every time... More often and almost everytime with turbo on...
the first solution was to fix all and disable all and fix it at 4.7ghz...
then i realised to try the voltage that come automatically sometimes agfter a crash.. 1.5 instead of 1.48-1.49... then.. at load it was going to 1.3 +- and crashed sometimes,, with 1.5 it go to 1.4 and stay fine..
then i tryed to turn on turbo and worked much better also... but then... (with fixed ghz was better with standard llcs and 100% normal volt on digi+.) with turbo on was better using more on both things.. like 110%+ and an above std option for llc.
BUT, it still crashing sometimes when it changes from 1 trad 5ghz to all cores at full.. then all time it crash....
its a pain this thing.. never got something that had so many problems to use at stock parameters...
So....
Im going to try at 1.5 maybe 1 tick more,, and turbo on with agressive llc...
the problem is that this more than std setings on digi+ make my processor lazy and less productive.. so i dont know yet... 1.5 is a huge voltage... im afraid if this will burn all after 1-2 years and goign to make more and more problems... dont know..
hope not..

Satability Factory setings :
Dont need to disable c1 c6 and turbo.
just do it:
1.5v if its not stock there... it will be after seting 1 time.. set like auto step with + and done.
leave c1 c6 on... c6 is needed for turbo modes when you have 1 job on 1 core.
disable high processing thing its bellow c1 and c6 thing... ( it alow the processor to trottle down if power/ temp get so high on the sockket ( Maybe you need to put a little fan behind the processor/digi+ processors on the back of the mb, it solves the socket high temp readings that arent from the processor itself, its from the digi+ temps that can reach 100+C with no colling behind..
so you still have protection, and it will not crash about inside each core temp.
set the nc/cpu and cpu digi+ to extreme.. and normal for nb... with 120% - try with 110 first.. - it will make the drops of voltages geting softly...
SET the SPEED of line... to more than std/normal - so the mb will react to fix the voltage faster....

this way you can set base frequency to 4.7 (23,5) or even 4.8 if you want ( meed real care on temps) .. and stock 5ghz turbo with total stability.
also you gona be sometimes cooler bcs it can set lower frequencies if it check that cant go with more, OR when youre not using it.
and done.

i made this after making it stable with non turbo... now i have stabilizeed turbo too.

Walblue wrote:
Satability Factory setings :
Maybe you need to put a little fan behind the processor/digi+ processors on the back of the mb, it solves the socket high temp readings that arent from the processor itself, its from the digi+ temps that can reach 100+C with no colling behind..
so you still have protection, and it will not crash about inside each core temp.

Could you please elaborate on this? Does the mainboard holder panel need to be cut?
Details of the fan size? A photo with the exact location perhaps?

Sorry about one more repply... but its necessary, i passed for all this path as others,, and i find that many are satisfied with fixed clock, or with over volting it... I made it ! there is:

The Main problem to stabilize it is the stepping on voltages, people try to solve this going with upper and upper base voltage.. its can solve the problem for a while... but i was afraid... of 1.5 and more that was the volts that my mb was doing... so....
THE Main problem is that, it set a higher voltage, and have a high drop... - it chashes when it get so low volt.. or when it put a high volt on the back turn... SO>.
thats was all the problem about..
FORGET Stepping Voltage,, turn off that "+" thing from asus mb...
set the cpu voltage, to 1,44-1.48(if you want to secure this more than 1.47, but in my testings more than 1.4/ 1.44 is enought for 5ghz !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! Understand ?
When you set it as manual, it not try to change it to f"fit" the clocks... and it will run fine on stock fixed voltage.. it almost cut the vdrops also. its a good MB that can reset and continue to give the same voltage ! not need to stepping all the time until it crash with a wrong combination...

So:
AI OC - MANUAL
23,5 (stock) ( or even 24 or what you want.. 24 is stable (4.8ghz)
T|urbo enabled.
turbo rate ... 25.... you car ser 24,5 to be total sure if you neeed total stability.. 😛 but 25 is stable!
bus 200
mem set your spd correctly..
ht link 2600
p saving disable..
spread sprectrum(s) disable.. not need.
THEN.... HERE !!! :
POWER thing bellow..
MANUAL MODE. in voltage seting.
just need to set the cpu voltage as 1.44-1.47 seems best)... etc.. in here was ok even with anything more than 1.350
This is what fix the fx9590 instability. ( the voltage will not more drop so much, and not return high when mb try to fix it again.. it get stable voltage everytime.
with non manual it will change voltage every ns trying to "fit" until it get a crash. thats was the problem on 2 fx9590 i had.

so.. then..
you can set cpu llc ultra... cpunb llc high... 120% both... power phase std... cpu freq volt, auto.. vrm spread... disabled...
duty- tprobe... cpu poewer resp FAST, cp nb pwer resp Medium... mem 100%
and on mem adv conf, south bridge i thing.. or north ? lol... set eec to off. as our memorys can work as non ecc and it is suported by the processor.

And...
DONE !!!
🙂 🙂
not need to fix the freq .. not need to turn turbo off.
not need to overvolt it... the hi
the higher voltages (1.5+) was for using(trying to solve ) the steping voltages with "+" thing there... its not needed as the voltage will not drop until crash anymore on manual mode.
i set on 1.4*v cpu and all others on auto.
and... done ! 🙂

Oh !!
also:
c1 ON !! ( much more cooler when not using. step down while not using... and its the same fast. its nanoseconds to change frequency again.. )
c6 ON !! ( turbo mode)
hpoc mode ( ON!! opcional, this with on or enabled will make the processor not step down, even while temp on socklet is high... ALSO make the processor dont step down when using all cores at total maximum, with c1 on and this off it will go at a little lower freq when full loaded.) you can disable to get it protected from overtemps etc.. BUT., it if get very high temp it wll trottle down the same way...
The soluton in here is to put a mini fan as i sayfd before on the back from mb on back of digi+, right side of the processor socked on the back side.

i think thats all... 😉

dont butrn the probessor with 1.5+ volts ! 🙂 lol dont need to disable turbo.... etc..