Interesting Random Restarts

[Farkas]

Ok, going to start this off by saying "Yes, I have search the forums here for solutions to my restarts." My situation seems to be different from what I've seen here on the boards and so I'm hoping that maybe this has happened to someone else. First off, my system stats:

 

RS482-Infinity

AMD Athlon 64 X2 4800 (2.20 ghz)

2 1 gig OCZ DDR sticks

ATI X1900 (Overclocked using ATI's built in Overdrive tool)

Western Digital 500gb SATA HD

D-Link Airplus DWL-G520 wireless PCI Adapter

580 watt HIPER Modular Power Supply Unit

WriteMaster DVD Drive

Running Windows XP Pro

 

All this is packed inside of an ASPIRE form factor case. When the restarts first occured I had assumed it was a heat issue seeing as how whenever it happened it was while I was playing Oblivion on full settings. But then it would restart while I surfed the net, the amount of time I spent on it didn't seem to matter, the power would just blink out and the computer would reboot. When my new, slightly larger, ASPIRE case came in the mail I moved all the components to it and replaced the default power supply that I had been using up to that point with the HIPER modular supply. However the restarts kept coming.

 

I updated the BIOS with the latest version from the DFI site, set the BIOS to Optimized Defaults like I'd read elsewhere, but it doesn't seem to go away. I even thought it could be the socket at my house, currently stationed in Japan and thus stay at a Japanese house, so I switched it around. Still nothing. So I ran the Western Digital HD checks, ran the Power95 stress test(and it ran for about 30 hours till I came home and stopped it).

 

I've read around and heard alot about the Libery power supply problems, which i don't have, and since I have gone through 3 Power supplies I think I can count that out. The temp inside the system itself never gets too high, the highest I've see the CPU is about 50 celcius, and the graphics card tops out at 80 celcius when playing games with extremely heavy graphics. Since it restarts despite workload, and the CPU/graphics card temp, I have dismissed the heat issue. I'd heard that RAM could be the issue but my ram type is on the list of those approved from DFI. I've even told windows to give me the blue screen of death, instead of restarting, should an error to that extent occur. Nothing shows up in the event logs either, a highly irritating problem. Since it still restarts I have to assume that it is a hardware problem.

 

With all my options exhausted I hope that one of you out there might have an answer. I've sent a mail to DFI but have yet to receive and response, seeing as how it has been more than a week I don't actually expect one. Any help would be appreciated. Thanks.

[DuckDodgers]

Farkas,

 

I picked up a DFI CFX3200-DR about a month ago and have been fighting instability problems with my AiW x1900. I could never get the system to run after installing my D-Link DWL G-520 WiFi card. I did finally get the system running but it was never real stable. Also I could never get my X2 4200+ over 2.3Ghz.

 

After fighting with it for 2-3 weeks I finally removed the D-Link WiFi card and install the onboard Marvell NIC yesterday and cranked the CPU up to 2.64Ghz stable (240x11). My suggestion is to try and remove the D-Link G-520 card and see if your instability problems go away.

 

Cheers.

 

DD

[Poisonsnak]

Hi Farkas, welcome to DFI-Street and please put that system spec listing in your signature if you can.

 

Also why are you running your 4800+ at 2.2GHz just out of curiosity is it not stable at 200x12? Also what voltage are you running it at?

 

Also what voltage and timings are you using for your RAM? Do you know your RAM's part number? If you have the EL Platinum sticks your system may be very similar to my RS482 based system.

 

A couple things I can suggest though:

 

-Set RAM timings to 2.5-3-2-5, Vdimm to 2.7V

 

-Set CPU voltage up, I wasn't stable at stock (200x11 1.30V) but I am stable now (erm well just running OCCT but looks promising!) at 218x11 1.40V which translates to 2.4GHz the stock speed for your chip also.

 

-Set NBSB 1.8V up - I don't really know if this is a big stability thing but when I set it to 1.8V I found I was running 1.72V or so, so I set it to 1.9 in the BIOS and now I'm at around 1.81V.

 

-Run memtest86+ (make a bootable CD www.memtest.org), loop test 5 about 50 times or so and then run the full suite of tests overnight and look for RAM errors that way.

 

-If memtest passes try SuperPi (use 32M), OCCT (30 min test), and Prime95 (use SP2004 - two instances on Blend, set each to about 850MB of memory)

 

What temps does Smartguardian report? I am running CPU-52 Chip-49 Sys-50 and it seems to be happy (this is at full load).

 

Also while I agree you are probably in the clear temperature wise DFI motherboards are very picky about their power supplies, do you have a voltmeter you can use to get accurate readings from your PSU under full load? (run 1 instance of ATITool for full load on 1 core and the video card, and 1 instance of SP2004 for full load on the other core)

[Farkas]

This took me a bit of a bit with our underway schedule (yeah for sea trials lol). Anyway, here's what I've done. As for my CPU speed,

it runs at 2.4 ghz, 2.2 was just a typo(my bad yo).

1) I've removed my wireless card.

 

2) Set RAM timings and voltages for CPU, RAM, and NBSB.

 

4) Ran Memtest86+ for 10+ hours without any errors reported.

 

5) Ran Prime95 for a bit(can't remember duration) without any incidents.

 

6) I haven't taken temps with the smartquardian, as of yet.

 

This is really getting a bit irritating. I'm worried that this constant restarting is going to **** with my harddrive and then, well I'll have to buy a new one when this one takes a .. Any other advice to give?

 

I've been giving considerable thought to just buying a Duo Core CPU and motherboard and leaving DFI behind. The next week or two will determine that I suppose. Thanks for all your help thus far.

 

Added Info: I was tkaing apart the computer again, looking for stuff that might be amiss, when my friend pointed out that the mother board was slightly warped(it bulges up slightly on the edge). This seems to have been caused by the case's odd design(I have an ASPIRE form factor case -the second edition- that is slightly longer by 1-inch to accomodate normal power supplies). In any case, I'm sure that such a bend in the board is bad but I'm not sure if it could be causing the restarts. Any opinions on this?

[noneed4me2]

I and a couple others had the same issue with this board and after changing PSUs it did go away for me, but I am not sure it was resolved with everyone; some gave up for different boards, but my restart problem went away and damned if I know positively what ever caused it. I did everything you did, bios flash, driver updates ect. but it just seemed to go away and has been stable running 24/7 ever since. It really is a fun board to mess with and overclocks really well for me. One thing on the PSU I tried 3 and ended up with an Antec Smartpower 2.0 500w and although I am not a 100% that was the problem, the restarts never came back. DFI just seems really picky with what hardware works with there boards. Nice build by the way I hope you get your problem fixed.

[iamsynthetic]

I'm having this restart issue as well. I'm fairly new to building computers, meaning I know my way around them, but all the testing of the temps and memory is new to me. I've never had any problems with any of my builds except this one. I've also never overclocked or messed with detailed bios settings like what you guys are talking about. I really don't think this is an overheating problem as I have the case open and a large house fan blowing directly into it does nothing to fix restarts. I am using a rosewill 450w, and people are suggesting that i use a bigger power supply. Can anyone give a list of working power supplies that they are using or seeing as this is a common problem, is there all ready a list out there? I believe i have the correct ram.. thanks in advance.

 

infinity rs482 -latest bios

AMD Athlon 64 4000+ San Diego 2.4GHz 1MB L2 Cache Socket 939 Processor

corsair - xms xpert twinxp1024-32000c2 2-512 sticks

rosewill 450w

onboard video and sound

sata hdd - 320gb seagate/unknown hdd 30gb

win xp sp1

[junediy]

I have seen this restart issue resolved on some forum somewhere before I started using this micro-board for most all the builds I do. I can't remember what or where but I Incorporate all the tips I've learned on each build and have never had an issue.

 

The Rosewill 450 watt PSU is more than enough when using on-board graphics, the 4000+ SanDiego with the XMS RAM is using only 70 to 75 watts at the wall. And each rail is never running low, it is a very well made PSU for its budget target and can handle the 2600xt and 2600pro very well, not the 2900 and not the 1900xt.

 

Make sure your RAM is fully seated in the first two same colored slots, or if clearance is required for CPU heat sink then use the second two same colored slots, remove, look for dirty connectors, and reinsert, same with all power sockets.

 

Always be sure to clear CMOS on any new motherboard before installing, I think that is the reason for the restarts.

 

Be certain to disable Cool and Quiet as that can cause many problems including restarts. Use the Smart-guardian and set it to load at windows and remove the check-mark for always on top. set it to Default then set it to software and raise the fan speeds to full on for all 3 chips. Set the warning temp to 48C for all and you will never reach that when ambient room temp stays below 85F.

 

Never use any of those cool and quiet drivers for any manufacture, maybe in 5 or ten years when they get around to fixing it.

 

If you are playing a full screen game and you hear the small beeps for temp you will know you are at 48C, try setting them to 45C on a hot day just to hear what it sounds like, I think the default is 52C and you will not reach this unless your heat sink was not seated when you clipped it on.

 

Good luck, I have had good luck with over 30 of this DFI RS 482 infinity and plan to do about 4 more before I switch to the Nvidia 6150 version of the DFI micro board for variety.

[ReelFiles]

Did you try a clean install of windows?

[iamsynthetic]

reelfiles - yeah i did do a fresh install of windows (on a new hard drive too). I get the same results with both hard drives in windows.

 

junediy - i'm going to try to do all the things that you've mentioned. my ram came with these led kits that set on top of the ram. it didn't fit into the case at all, so i'm just not using the led modules. i seriously doubt that could be causing the problems though. i didn't setup the cool and quiet option, however i did notice that it was still set to auto in the bios, so i disabled that.

a couple quick questions for you...

 

when you flash a bios to update it do you have to flash it twice or is that just a myth?

 

when i clear my cmos, do i need to update drivers in the bios again? i guess what all will that affect and what all do I need to redo?

 

i'm a bit lost under the smart guardian. i set the 3 slider options under the tab "cpu fan" to 100% and on that same tab at the bottom i clicked on software programmable mode, and changed the temp alarm setting to 49C. I think thats what you meant. If i'm wrong just let me know. the results for my temps are as follows...

 

not running anything but IE *with case off*

cpu - 48C

chipset - 39C

system - 37C

fan speed - 3125

 

voltage:

cpu - 1.34V atx+12V - 12.13v

nbht+1.2V - 1.16v nbsb+1.8V - 1.81V

atx+5.0 - 5.17V atx+3.3V - 3.3V

dram - 2.62V vbat - 3.37V

 

i started diablo 2 (video game) up and my temps on my cpu went up to 55C and the chipset and system went up to. before i could check them and write them down it restarted. can anyone explain what temps are ideal in a stable system? 55C seems low to me.

 

thanks for the help. the reason i bought this psu is because someone on newegg had said it was working fine for them. maybe it is something with the psu, but i am going to try to do everything that you guys said first. OH.. junediy.. what psu did you use that worked well?

[iamsynthetic]

*bump*

[junediy]

iamsyn: I wish I could remember exactly where I found the cure to your exact problem as it was written much better than I can.

But I now think yours is heat related, didn't know yours was running that hot.

I think 55c for this board is much too high. Mine would lock up at that temp, the highest I've seen on any I've installed is 38c using the stock cooler set low RPM on the 3700+ and 4000+ CPU during Rise of Nations. Setting to full RPM cured that.

I think only the cheap no name mobo needs double flashing of the BIOS. I only flashed one of these DFI boards and used the Diamond flash and the owner said it improved, I couldn't tell a difference, he was using only media center OS.

When you erase the CMOS with the jumper, with power disconnected, after erase hit the delete key on restart and go first to set "default minimum" and say yes, then set to "default optimum", and say yes, then go to PNP and enable to reset, then reset date and time and any thing you need different, I leave most all on Default Optimum and auto detect, get it to boot and go back later if I want to bump things up. In System Monitor make sure your cooling is either Smart Guardian or Full, disable cool and quiet. Then save settings and exit. I don't even install Cool & Quiet drivers ever.

In windows you can set your guardian to Fahrenheit if that makes more sense to you, mine always runs about 87F to 97F depending on room temp.

What is your room temp? and what CPU fan? You may need to reseat your CPU cooler, make sure your wires are away from air flow to CPU, it will be very difficult to remove the heat sink after it has been very hot and not harm the pins on the CPU and socket, some people have pulled the pins right out of the CPU trying, clean the old paste off with alcohol and reapply a thin coat of diamond paste or artic silver according to instructions. The stock paste would have been fine if there was any, but yours seems to have a seating problem or not flat maybe, something is wrong to run that hot. Make sure your resolution is at least 1024x768 to get the load off the CPU.

I can't know for sure but really think you have a cooling problem, with side cover off I can't reach 37C when room temp is 95F no matter what game I'm running.

I also use the Rosewill 450 because of reviews on NewEgg, except when it is for 1900 and above add in video card, then I use a 500 watt of most any make like Ultra, or Corsair when someone has the money for that.

But anything better is more for SLI or Crossfire and not needed on single slot. I have one friends DFI RS482 running a no name 300watt PSU just fine.

It depends on CPU cooler but a stock one should be around 3444 RPM set to full when temps are above 37C.

You seem to have smart guardian under control but maybe set to default first then back to software at full, I think SG is a bit funky like written in Taiwan maybe.

Are you sure you need to replace RAM, all your hardware seems top notch including PSU, it seems more likely a cooling problem at the heat sink, did you use some kind of thermal paste interface?

I'm no pro at all, but have had much help through the years from sites like this and manufactures through email and a genius nephew has really helped.

I hope you can fix your heat problems then I bet it will run perfect for you.

Keep searching all the forums. I try to twist a tiny bit back and forth on an old heat sink after releasing the clamp before any lifting pressure, but know I may break both the board and CPU trying. Try to keep any old or new paste off of anything so it won't short out the board. You may need to re-lap the heat sink if it is not flat and smooth, I use 4000 grit sand paper on glass using a little water and being very careful to keep flat, checking often if it shows not touching some where, (Not Flat). Or just get a better cooler that fits well. I don't think you need to replace RAM or PSU. You won't find much use for that high end DDR RAM except this board as they all are DDR2 now. Try some extra case fans, I run five 120mm & 140mm case fans and cut holes for two and run filters to keep the dust out of the heat sink fins and fans.

Heat is the enemy and yours is hot I think. My Nephew always runs seven or more case fans, try to keep equal intake and exhaust and know you have exhaust in the PSU already.

[iamsynthetic]

this is the cooler i'm using...

ASUS P5A2-8SB4W 80mm Sleeve CPU Cooler

http://www.newegg.com/Product/Product.aspx...N82E16835101012

 

when i was installing the processor and the asus cooler i didn't use any paste. the manual said i didn't need to if the cooler had a sticky backing to it. this may be my problem. the cooler did have a sticky back to it, but i think i may need to use some paste. by my question is this...

 

if my cpu is running at 121 F, and i use this paste, then my chipset - 110 F and system - 110 F are still going to run hot. correct? I'm not doubting you at all, just asking about the other two temps. my room temperature at these reads is about 95 F.

 

i did install another case fan recently, but it didn't drop the temperature. my psu sits almost directly over my processor's cooler. literally almost touching each other. I'm thinking about rerouting some things to see if i get some different temps (after i use some paste).

thanks junediy.