• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

SOLVED Help: BSODs and freezing issues - and a undetected H100i GTX (new rig)

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

Niteshade

Registered
Joined
Dec 25, 2011
Location
Sweden
I really need help with my new rig.

Hardware list at the bottom of this thread.

The first problem I got was an undocumented beep signal during POST: one long, four short, one long. After connecting a screen to the PC I saw that I had a CPU fan error. Appears common with the H100i. The PSU fan did also sound like a farmers old tractor, but this noise stabilized and went away after a time. Didnt find any solution to this H100i problem though - but found some inconsistencies in ASUS installation guides: My H100i does NOT have a SATA connector. It only has a 3-pin CPU-fan cable, a fan cable splitting into two meant for the radiator fans and the Corsair link mini-USB. The Corsair link cable is optional (plugged to a USB2 internal header) but the other cables are stuck to the pump. My mobo has a 4-pin header so I attached the 3-pin anyway into the CPU_FAN header (to the far right - and the only position it can be placed in). I googled a lot and couldnt find much info about this. Someone said that Corsair stopped using SATA for power and just went with it via the CPU fan header for later versions of the H100i

Also during this time, the onboard diagnostic LED-lights on the mobo turned on, waited a sec and then turned off - one at a time, ending with the red BIOS-failed LED. On top of this I´ve also experienced situations where I didn´t get past POST - or there didn´t seem to be a POST: sometimes nothing really happend when pressing the power button (other than fans running). This situation also happend a few times when the system performed an automatic reboot - such as after changing BIOS-settings.

Either way, I went into BIOS and disabled the Q-fan stuff that I read others had done. I also put the CPU fan monitoring to IGNORE. Now the beep code went away but now I got a "Overclocking failed" error during POST. I resolved this by turning off XMP on the motherboard hardware switch. (It appears that the hardware switch overrides software settings in BIOS).

Ok so now the PC boots, and I am able to install the mobo chipset and other drivers, as well as the drivers for the graphics card +Corsair link software for the H100i.

Now I experienced other errors: Random BSODs (always different error codes) and screen freezing. I knew this wasnt due to overheating because out of sheer paranoia I made sure the CPU temp was fine in BIOS before booting into windows. The H100i LED light is shining, and the pump is working around the water!

Worth noting here: I did not perform a clean install of windows. I migrated it from my previous PC setup. (I was actually surprised the OS thought it was fine with this new hardware.)

Other thoughts and comments:
-It´s not the RAM which caused the POST-errors. I´ve successfully booted with one stick of RAM into windows. 1 stick or 4 doesnt matter - I still get freezes.
-The fans on my graphics card are not running. Possibly normal, as some say the DirectCU-II fans only start running at a GPU temp of above 50C (or maybe 60C?)
-Corsair link reports that the H100i fans are not running. These are the ones at the radiator which goes into the splitter (which goes into the pump). So neither the CPU header nor the radiator fans are detected by the mobo.
-Corsair link reports that my SSD drive is exactly 100C hot. I gently touched the SSD with first a pen, and then a finger. It´s definetly not a 100C.
-My temps are good. The CPU is at a stable 27-30C idle which is 6-10C above ambient temp. Mobo reports a steady 25C.

What I need help with:
-Getting the H100i fully functional
-Stop the random freezing and BSODs

My suspicions:
-Faulty PSU as it is 4 years old. Will check voltages via some hardware monitor. Will also try booting without gcard and run with the internal one on the mobo.


------- HARDWARE LIST:

Memory:
Corsair Vengeance Pro DDR3 2400MHz 32GB 4x8GB CL11
New

Graphics:
ASUS GeForce GTX 970 4GB STRIX DC2 OC
New

Motherboard:
ASUS Z97-A
New

CPU:
Intel Core i7-4790K 4,0GHz
New

CPU cooling:
Corsair Hydro H100i GTX
New

PSU:
OCZ 700 MXSP - 700W, modular
Almost 4 years old

SSD:
Some PNY @ 240gb
Almost a year old
 
Last edited:
Ok so I removed the gcard and plugged a display into the integrated graphics card DVI-port.
Booted up, and the failed boot -LED was shining for a few seconds on the mobo but then the PC booted normally (no beep code here).

Opened up chrome and a youtube video and just let the computer be. This time, instead of freezing within a couple minutes, the computer was active for about 10 minutes before I got another BSOD with a nonsensical error code. Computer did NOT reboot by itself after this!

Corsair Link still says that the H100i fans and the CPU fan is dead at 0 rpm. However, I can control the pump and fans just fine - tuning them to 100% speed gives off a noticable noise increase. It´s still annoying though that the mobo doesn´t detect the fans, but it's not system critical.

So, still having this BSOD/freezing issue. I do have an extra, non-modular PSU of 400 watts that I guess I could perform a test with. But I´m unsure if thats enough to power the mobo, i7-core, h100i, drives, fan-controller and DVD-RW drive (the latter can be pulled out with ease, but it's pretty crowded behind the fan controller!) Do you think 400 watts is enough for a quick test without a gcard?
 
What are your RAM speed, timings, and voltage set to in the BIOS?
 
Couple things here.
1) Corsair link software is known buggy.
2) Fresh install before you try to diagnose hardware, otherwise your chasing ghosts.
 
The other PSU turned out to be too old - doesn't even have an 8-pin power connector for the mobo. I guess I'll have to buy a new PSU tomorrow :( Been looking at this one (which has a superb warranty): EVGA SuperNOVA G2 - 650W. According to PSU power calculators that should be more than enough for my stuff, with about 100W extra spacing/padding in case I wanna OC the CPU.

But before I buy a new PSU, is there any other check I should perform?

EDIT: RAM and CPU stuff is set to "auto" in BIOS atm (default)
The frequency appears to be underclocked to 1300MHz (should be 2400MHz) with a voltage at 1.539-1.542 (yes, alternating - is this bad?)
Timings are 9-9-9-24 with a DRAM command rate of 1

The voltages on the PSU rails are:
+12V: 12.384V
+5V: 5.080V
+3.3V: 3.408V

CPU vcore: 1.088V - sometimes dropping to 1.072V

All of these stats are from idling in BIOS. The computer does not freeze here - only inside windows.

As for the Corsair Link software: I've heard others say the same, claiming the software is bad. But that doesn't explain the CPU fan detection during POST...
 
Last edited:
The SuperNOVA G2 is a phenomenal unit. You could even go for a 550W version and have plenty of room to OC.
A 650W gives you enough room to OC on SLI 970's with the 4790K.
 
As for the Corsair Link software: I've heard others say the same, claiming the software is bad. But that doesn't explain the CPU fan detection during POST...
If the fan/pump isn't above the minimum RPM that the board is looking for, it will still trip the warning even though they are running. You did the right thing disabling it.......as long as it's actually running.
 
After looking up that PSU, I concur. It's an old design that's very heavy on the 5V.
 
Okay I've now installed the new PSU and I think it is safe to say that the old one is to blame for these BSOD/freeze errors.
Now onto my next problem: BIOS underclocking my memory. Why is this even happening? Should I just set my frequency and timings manually and if so, where can I find these? Thanks in advance.
 
Definitely try setting it manually.
You can worry about the first few timings only and leave the rest.

What's the model number of your RAM?
 
Corsair Vengeance Pro DDR3 2400MHz 32GB (4x8GB)
Not sure if this is the model number but it's above the bar code on the back of the package:
CMY16GX3M2A2400C11R
 
Yep, that's the model!

So, in the BIOS, set 2400MHz on the DRAM Frequency, 11-13-13-31 on CL-tRCD-tRP-tRAS, and 1.65V on the DRAM Voltage.
 
Alright, I'll do so tomorrow. Gonna update the BIOS and chipset drivers as well. Now it's late and I need to get to work in a few hours.
(BTW: Mind explaining a bit what you meant when you said that my old PSU was "heavy" on the 5V? I'm not that techy about PSU's)
 
Alright, I'll do so tomorrow. Gonna update the BIOS and chipset drivers as well. Now it's late and I need to get to work in a few hours.
(BTW: Mind explaining a bit what you meant when you said that my old PSU was "heavy" on the 5V? I'm not that techy about PSU's)

Sounds good, let us know how it turns out :)

Older systems used to take a lot of 5V power, along with 3.3V and 12V, from the PSU while modern systems almost exclusively use 12V.
By saying it's "heavy" on the 5V means that it was designed with these older systems in mind instead of modern systems because it is rated for high amperage on the 5V rail.
Look at the 5V amperage rating on your old PSU and new PSU, then think about how close the overall wattage rating is between the two.
 
Yep, that's the model!

So, in the BIOS, set 2400MHz on the DRAM Frequency, 11-13-13-31 on CL-tRCD-tRP-tRAS, and 1.65V on the DRAM Voltage.

I find some conflicting info regarding this. NewEgg also mentions those timings, but on ProClockers, Hardware Slave and Corsair's official site the timings are 10-12-12-31 which - with my limited understanding of memory - is faster.
 
Back