First build, POSTs fine, but frequent freezes/reboots, should I try underclocking RAM?

Pyutis

Honorable
Nov 22, 2013
5
0
10,510
TL;DR:

  • - System reboots and freezes often while starting windows or running applications.
    - More rarely, freezes occur in BIOS settings.
    - Memtest freezes at test 6 or 9, but no errors.
    - Still going through different RAM stick/slot combinations and memtesting, thus far no errors, only freezes.

    Will underclocking the RAM damage it in any way?
    I'd like to try underclocking to see if it works better, since manually adjusting to manufacturer settings slightly alleviated the freezing (only memtest, not windows).

The build

This is my first time building a PC, after having been a laptop person for almost a decade.
To prevent compatibility issues I decided to follow a recipe and take a build from somewhere else.

The build is the 600$ build here:
http://newbcomputerbuild.com/gaming-pc-builds-of-the-month-2/october-2013-600-1000-1500-budget-gaming-pc-builds/#600-pc-build
The only difference is the case (slight upgrade) and optical drive (the one listed was hilariously expensive in my country).

Just to keep it neat, here are all the items:

Motherboard: MSI ATX Motherboard 970A-G43
CPU: AMD FX-6300
GPU: Gigabyte AMD Radeon HD 7870
RAM: Crucial Ballistix Sport 8GB Kit (4GBx2)
HD: Western Digital Caviar Blue 500 GB
PSU: Antec VP-450 450 Watt
Case: NZXT Source 210 Elite
Optical Drive: LG GH24NS95

All parts are new and bought within the last 2-3 weeks.

All encountered issues (however minor)

I will list all issues I noticed, in case it helps with troubleshooting.
- Build was posted on US-based site, I am in Germany; PSU is the same model, but for European voltage/frequency settings
- Case had minor, almost unnoticeable bends where PSU installs, had to apply some force to screw it in properly, but otherwise no problems
- CPU heatsink had one slightly bent fin, not sure if it was there from beginning or if it happened during installation
- CPU is at 45°C right after system start, and heats up to 55-59°C, where it stays stable; not sure if this temperature is too high; heatsink came with pre-applied thermal compound and instructions said to use this pre-applied compound
- Rear case fan cable too short to reach motherboard, does not spin when plugged to PSU, regardless of which molex connector I use
- Plugged top case fan to PSU; I could plug it to the motherboard, but short cable and multiple connectors create cable managing issues, I wanted to prevent cables touching motherboard
- Multiple reboots and freezes, occur almost always during windows installation/boot/function, rarely in motherboard BIOS settings, frequently while running USB-bootable DBAN, sometimes and always at specific parts when running Memtest86
- When the system reboots after saving BIOS settings (save and reboot option), there is no video output, and I must manually reset

POST and Windows install: Major issues with freezing and random reboots

Built it on Saturday, took anti-static measures in the form of wristband, frequent case touching, and cotton clothes.
The build turns on and POSTs fine (although I don't have a motherboard speaker), CPU fan spins, upper case fan spins, rear case fan does not spin (DOA? tried different connectors, no dice), the motherboard recognizes CPU and RAM correctly.
Graphics card obviously works otherwise I'd have no video output (no built-in graphics).

When I tried installing windows 7 64 bit, it rebooted at the part where it's supposed to, but then I got a "BOOTMGR missing" error. After tinkering around with DBAN and boot order and several attempts at windows installation, I managed to pull off a clean windows install.
However, all my attempts were plagued by freezing and random rebooting of the system, which I at first believed were software issues (corrupt install, DBAN freezing/getting non-fatal errors).

Now that windows is installed, though, I still get random freezing and reboots. It often freezes when windows is starting, needing power/reset switch reboot. Often the screen goes black and the system reboots.

Troubleshooting

- I checked all the cables again, improved cable management; issue remains the same
- Switched SATA cables of HD and optical drive; issue remains the same
- Unplugged rear fan that was not working, removed optical drive to reduce PSU strain, in case that was the problem; issue remains the same
- Removed and reinserted RAM; issue remains the same
- Started running Memtest86 using different RAM stick/slot combinations; changes in freezing/rebooting behavior!

Memtest86

I'm using an USB-bootable Memtest86 (newest version), and found a reproducible pattern of freezes, but without errors. I'm still running tests, and will have to stop to go to work soon, but wanted to give an overview already and ask for advice.

As my motherboard manual instructs, I have installed both sticks into DIMM2 and DIMM4.
Windows almost always freezes at "starting windows" screen. If it manages to boot, it may reboot on its own.
Memtest tests 1-5 run fine, no errors.
Memtest test 6 freezes right after it starts, but gives no error. Tested twice with same result. Needs manual reboot.

Removed stick in DIMM4.
Windows boots fine, and appears stable if I don't touch anything. Gets slow, freezes, and reboots when trying to run an application (solitaire, system performance check).
Memtest tests 1-5 and 7-10 run fine, no errors.
Memtest test 6 freezes with no error. Needs manual reboot.

Removed stick in DIMM2 (stick #1), and installed stick that was previously in DIMM4 (stick #2) into DIMM2 slot.
Windows freezes at boot. Needs manual reboot.
Memtest freezes right before any tests load, no way to enter configuration. Needs manual reboot.

Left stick #2 in DIMM2 slot, and manually adjusted voltage to 1.492 V, and frequency to 1600 MHz (manufacturer default recommendation).
Windows freezes at boot. Needs manual reboot.
Memtest tests 1-8 and 10 run fine with no errors.
Memtest test 9 reproducibly freezes after a variable amount of time.

Will underclocking the RAM damage it in any way?
I'd like to try underclocking to see if it works better, since manually adjusting to manufacturer settings slightly alleviated the freezing (only memtest, not windows).
Should I try something specific next?

Sorry for the lengthy post, I appreciate any help!
 
Solution
To answer your main question, under-clocking the RAM will not harm it in any way, and if reducing the temperature doesn't fix your problem, this may be a reasonable alternative.

Yogi

Pyutis

Honorable
Nov 22, 2013
5
0
10,510


Thanks for the input! I'll check the heatsink as soon as I get home.
I have some Arctic Silver 5 lying around, which I used on my laptops. I might just use that.
 

Pyutis

Honorable
Nov 22, 2013
5
0
10,510
First, I reinserted the two RAM sticks in their original positions. Then I tried to work on the CPU cooler.
Since the cooler didn't want to come off easily, I nudged and twisted it slightly, to have it sit better on the CPU.
Turned on and still had problems with high temperatures, it was even freezing on the BIOS screen, which only occurred rarely before!

I opened the case again, removed the CPU cooler with a lot of patient light twisting and also thanks to rubbing-alcohol soaked dental floss. It came off cleanly, I wiped off all of the old thermal compound, and applied my Arctic Silver 5. I made sure to nudge the cooler a bit to have it adhere to the paste well once I installed it on top.
However, the temperature was still rapidly rising into the 50s, even though the case itself was 23°C according to the BIOS (roughly the temperature of this room).

Then I noticed that the BIOS had set itself to overclock mode and froze soon after. After rebooting I quickly reset to factory default, and after another reboot it was working as intended and did not freeze. I manually set the RAM voltage to 1.492 V, and underclocked to 1333 MHz.
Windows booted up without a hiccup, something nearly impossible before with both RAMs inserted! I haven't tested its stability, however.

Since I read elsewhere that this particular motherboard's BIOS is a bit unreliable on getting the temperatures correct, I will transfer Core Temp over by USB, and then do more testing.

This thermal paste has always worked wonders for lowering idle temperature on the laptops I've used it on, so I'm suspecting it might be insufficient fan cooling or bad airflow due to the dead rear fan. I'm going to contact the seller of the case for a replacement fan next morning, and am considering getting a third party CPU cooler, too.
Also, depending on how the system performs now, I may buy new RAM tomorrow.

I've also been wondering if I should get a more powerful PSU, since I may need to install more fans, and I believe this PSU hits the limit of what the system needs.
 

Pyutis

Honorable
Nov 22, 2013
5
0
10,510
UPDATE: The system is up and running now, apparently stable.

CPU
The temps in the BIOS would never go down. However, both Core Temp and HWMonitor show really low temperatures at idle (around current room temperature of ~16°C), and they seem to increase to a maximum of 45°C under load.
After digging a bit around I found multiple posts around the internet of people complaining about the MSI 970A-G43 and MSI 970A-G46 motherboards having a "wrong" BIOS temperature reading. The sensor seems to be placed in an unusual location. However, the temperature shown in the BIOS doesn't seem to control the CPU fan, which is controlled by the sensor that is read out by Core Temp and HWMonitor.
I guess I'm fine with the stock cooler until summer.

RAM
Underclocking to 1333 MHz (1.492 V) magically fixed all issues. Memtest86 passed all tests without a single error, and I was able to install drivers, windows updates, and run the CD that came with the motherboard.
After installing all the stuff, I tried setting the clock back to its intended speed of 1600 MHz, but the freezing returned. I gradually upped the voltage until everything was stable. I'm currently at 1.565 V, which seems to work fine.
I'm not sure if the increased voltage requirement is a normal, motherboard-specific thing, or if it speaks for a low quality RAM. In any case, the problem appears to be solved.

Thank you all for your help!
 


Glad to see that you are nearing a final solution!

Is this your RAM? http://www.crucial.com/store/partspecs.aspx?imodule=BLS2KIT4G3D1609DS1S00
If so, it is rated at 1.5 v., and 1.5 v. RAM can typically be run safely at 1.65 v., but you may want to confirm that with Crucial Tech Support. I would contact them anyway to see what they have to say about your problems at 1.50 v. Personally, I'm surprised that it gave you problems at 1.50 v. I have Crucial 1600 MHz RAM in my system and although I did increase the voltage, that was because I OC'd my RAM to 2133 MHz and CPU to 4.6 GHz.

I'm a big fan of Crucial RAM. Crucial is owned by Micron and Micron makes all of their chips, so when you buy Crucial RAM, it is all manufactured by the same company, and I like that.

Yogi