cancel
Showing results for 
Search instead for 
Did you mean: 

Asus Strix Z270i - Issues, Red CPU LED..

Malkan
Level 7
Hey,

So i built a new PC a few days ago and today when turning it off and moving it to a new desk, issues occured. (Rarely turn my pc´s off). After having it moved to the new desk, i plugin all the cables and start the pc - Wont POST and CPU_LED is lightning solid red. Pressed reboot button, took 5-10 sec then it rebooted and started flawlessly into windows.

I can reproduce the issue 8/10 times (tested) by turning PC off, pull power-cord out and wait 3-5 min each time, then it most likely wont post and getting stuck with red cpu led. It did happen once when i tested just rebooting system a few times aswell. I have reseated CPU + Cooler, no joy. Cant spot any bent pins, looks fine to me. I have no OC, only XMP turned on (3000mhz Corsair) rest is stock + fast boot turned off. The PC works flawlessly in Windows, been running stress-tests (Realbench) for 2 hrs. Will run overnight aswell.. Works fine in games, been playing Wildlands all week without hiccups.

I tried to reproduce the error with XMP turned off = Bios completely stock. Only tried 3 times, it booted every time fine, i will try more times to reproduce the issue with XMP off. So far with what i have tested, it seems to be a issue with xmp/bios, any suggestions? Been using google most of the day and it seems to be the exact same issue that people HAD with Asus z87-a, been reading quite a few topics about that which match my issues very well.

Specs:
Asus Strix z270i-gaming @ latest bios 0607 - Same issue on both available bios.
Intel i7 7700k (Delidded with Rockit88 delid tool) - works fine, had to be done because of 85+ temps @ stock in games, 90+ Aida64.. Now around 65c.
16GB Corsair 3000mhz ddr4 - CMK16GX4M2B3000C15
Asus GTX 1080Ti
Samsung 960 EVO 250GB M2 + 2x 250GB 850EVO.
Corsair AX1200i PSU
Corsair H100i v2 cooler
Corsair Obsidian 250D

Hopefully someone have a answer 🙂 Thanks!
741 Views
96 REPLIES 96

I stumbled upon this thread as I have been experiencing the same thing. While browsing this forum, I noticed this sticky: https://rog.asus.com/forum/showthread.php?90663-Kaby-Lake-overclocking-guide

Skimmed through it and on page 3 I read this:



Apply changes, then retest for stability. Also be sure to check the system can POST consistently from a power cycle and warm restart. Memory training occurs during these periods, and if these voltages haven’t been tuned properly, it will result in the system being unable to POST or BOOT.


I was like ok, never really tinkered with these and I've got nothing to lose. I did a cold boot with a 4.7 OC, manual timings, and VCCIO/SA voltages on Auto. No POST. Did my tried and tested revival method by unplugging the power cord, hit the power button a few times, plug it back it and power it back on. Pressed F1 and decided to put to 1.1V on both (actual values 1.12 VCCIO, 1.128V SA) as the OC guide suggests for the minimum. My memory is a Corsair Vengeance LPX DDR4 3000 2x 16GB (CMK32GX4M2B3000C15W) rated 15-17-17-35 timings @ 1.35V.



Cool, it booted. Then I thought, what if I changed it a bit. How high or low can I go? I hit the + button while at 1.1V on both and it had an input of 1.11250. Saved changes and sure enough gave me a RLOD. This is what it looked like after the reset:



Notice it said 0.968V VCCIO and 1.072V SA, I was like what if I put these values? Again using + on the keyboard, I went as close to these values as possible with having a bit more instead of less:



It booted, although I noticed the amber light stayed lit for like 3 seconds before it finally went through. So at this point I wanted to try the lowest values for both. From Auto, pressing the + sign resulted to 0.9V VCCIO and 0.7V SA. This time the LED was a solid amber.



So I figured out the lowest (0.975V input/0.992V-1.0V actual, this value was changing back and forth in real time) and highest VCCIO value I could go (1.1V input/1.12V actual). The SA on the other hand has a bigger room for error. An input of 1.0 to 1.2 booted no problem. Didn't test any higher than 1.2 though. It would've probably still booted at 1.3V.

Going back to the chart in the beginning, I'd like to stay at these values as close as possible on 3000mhz. I settled at 1.0625V VCCIO (1.088V actual) and 1.125V SA (1.152V actual).



As mentioned earlier ITT, this mobo needs a BIOS update to fix/stabilize the Auto setting for the VCCIO and SA, particularly VCCIO, as it has a smaller room for error. Maybe that explains why sometimes the system will boot and sometimes it doesn't. I need to read up more on this and find out where the author got those values from. For now, these are 100% stable. I'll report back if I see the RLOD again.

Any update on this issue yet Asus? Bios coming soon?

Lots of others having pretty much same issue on Asus boards; http://forum.corsair.com/v3/showthread.php?t=149795

+ the people in this thread!

No changes here, still the same random colboot bugs and i check asus bios page atleast once a day 😞

but im curious, what happened to the previous guys post, he posted a bunch of bios (screenshots) settings he said worked, did he (as i did) realize it doesnt matter and removed his own post? 🙂

Good Thread good Information Thanku Admin

kingkajsdhaskasda wrote:
Our mem is listed in the QVL pdf: https://www.asus.com/Motherboards/ROG-Strix-Z270I-Gaming/HelpDesk_QVL/ which should mean it was tested and supported. I ran AIDA64 with CPU/FPU/CACHE for 1 hour no errors displayed that i could see (do the errors show in a report of some kind?), i forgot to mention i ran memtest86 on Saturday for little more than an hour (3 passes) no errors.


Alright, now try running only System memory without CPU/FPU/CACHE, should produce error after a while. Yeah the test goes all red.

Raja@ASUS wrote:
XMP is overclocking, so it's not plug-and-play for all CPUs. You may need to tune SA and IO voltage. Would suggest reading the Kaby Lake overclocking guide here as it has some guidelines for voltage tuning:

http://edgeup.asus.com/2017/01/31/kaby-lake-overclocking-guide/


Ok, sounds strange that they should not work as is but i will try. When you say relax timings, is that putting them higher or lower as they are in the XMP? My SA and IO voltages are both higher on Auto then what they say in the Guide, you think by putting them at lower voltage, it could help?

Malkan wrote:
I got the same memory kit CMK16GX4M2B3000C15, should not Corsairs memory be supported on a Asus board? Seeing how popular Corsair are.. Any suggestions about voltages and timings to try?

Edit; Just got home from work, before i left for work this morning i left Aida64 - System memory test ONLY running, it failed after 6 min.


XMP is overclocking, so it's not plug-and-play for all CPUs. You may need to tune SA and IO voltage. Would suggest reading the Kaby Lake overclocking guide here as it has some guidelines for voltage tuning:

http://edgeup.asus.com/2017/01/31/kaby-lake-overclocking-guide/

kingkajsdhaskasda wrote:
For me it happens with or without XMP these mems is what im using which should be supported; CMK16GX4M2B3000C15



When you say without XMP, do you mean the board will not POST after AC power cycle if you've loaded optimized defaults?

Raja
Level 13
Yeah, a lot of people fall into the trap of believing that XMP = plug-and-play, 100% of the time. It simply doesn't.

With SA and IO voltages, too much can induce instability, just as too little can. You'll have to tune slowy and see if it helps.

For the timings, to "relax" them, you need to increase them. Aside from Refresh Interval (tREFI), which is more aggressive as you increase it. You can try increasing the secondary timings by a few clocks each and seeing if it helps. If it doesn't, you'll need see if spacing the third timings helps.

An alternative is to enable the Fast BOOT settings lower down in the DRAM timing page, but I'd only use that as a last resort. When enabled, it bypasses the re-training processes and reloads previous values from NVRAM. The issue with that is you need to chance upon an initial training pass that has sufficient margin to cope with system variance. If the training isn't good to start with, it isn't going to help.

Yeah i can imagine that most people fall in that trap, myself included ^^

Testing 16-18-18 with 1.15v SA/IO, 40 min passed so far. Will try turning pc off completely and removing the power cord for a few mins and see if i can post without the CPU-LED turning red. Why does the motherboard light up the CPU LED for both of us when there is a RAM LED above it? That´s what got me thinking that my cpu was broken in the first place but it really seems to be a memory issue

While we got you here Raja, i got a white sticker on the backplate-screw of my 1080Ti which is from Asus, do i void my warranty by removing this sticker? I dont like having a white sticker in a all-black build

Edit: never mind

I too fell in to that trap, thought XMP was a guaranteed setting. 🙂

But after fiddling around i finally found a setting that works for me, i also ran memtest86 for 1h22m with 31 tests all passed ASWELL as 'the boot post screen after a cold start error' i had, even with/without ERP enabled!

So currently i have XMP enabled BUT i tweaked the DRAM voltages from 1.3530v -> 1.3860v and that solved all my issues sofar!

Thanks for pointing me in the right direction Raja@ASUS! 🙂

edit: btw i tried running a Auto profile and selecting 3Ghz mem freq which had the mems running at 1.504v, thought that was a bit high 😛