1. Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

Weird problem

Discussion in 'General Hardware' started by tomba2k, Nov 4, 2011.

  1. tomba2k New Member

    Joined:
    May 11, 2009
    Messages:
    33 (0.02/day)
    Thanks Received:
    2
    Location:
    Croatia
    So, i had this system:
    ECS A780GM-A (1.1)
    AMD Athlon x2 4000+ (2.1GHz Brisbane with FSB clocked from 200 to 220MHz)
    (Phenom II stock cooler)
    2GB Geil BlackDragon
    Point of View 8800GTS 320MB EXO
    160GB Maxtor + 330GB 2.5" portables (80 + 250GB)
    Asus DRW-1608P2S
    Saitek Cyborg Keyboard
    SteelSeries Xai
    Canyon CN-WF511 (Realtek RT61) WLAN
    Frontier Achilles chassis with LC-8400BTX 400W PSU

    once, when i sold my radeon 4650, before i installed my (then) new geforce 8800, i tried to clock my integrated radeon 3200. best stable speed i had got was 650MHz.
    any speed above that made pc restart on windows logging in.

    then, few days ago i bought another 2gigs (2x1) of OCZ Platinum DDR2-800.
    first thing i tried was to clock my athlon from from 2310MHz to 2394MHz to get 399MHz on ram (CPU/6). same thing as when overclocking my integrated radeon happened.
    my pc restarted, sometimes while logging-in to windows, sometimes while booting windows, and sometimes even during bios post-message being showed.

    BUT, this time reverting fsb speed to 220, 210 and even original 200MHz didn't helped.
    first thing that crossed my mind was that mainboard is damaged (because of previous integrated radeon experience), so i changed it to Biostar A770 A2+ - problem continued.
    i've tried putting my friends old Gigabyte GeForce 7300GS (128MB DDR2-800 - passive cooling), installed XFX 450W (XFX-P1-450S-X2B9) PSU, tried removing and combining my rams in different ways (all 4 in all combinations, 1 ocz+1 geil, 2 geil's, 2 ocz's, 1 geil, 1 ocz), nothing helped until i discovered that underclocking my cpu to 1500MHz prevent the problem happening.
    also, setting cpu voltage to 1.4V (default is 1.25/1.325V) makes it stable on 1.8GHz.

    so, i changed cpu to athlon II x2 7750+ black edition (AD775ZWCJ2BGH).
    again, restarts happen - no matter is it windows xp or windows 7, after loading text mode drivers, boot screen shows up for a few seconds, then it restarts. xubuntu freezes on start of loading live user. i have also tried latest bios's for ecs board, biostar board, pointofview gpu and gigabyte gpu, no use.

    i presume that underclocking my cpu would again prevent restarts, but only components i didn't change are chassis, dvd-rw, hdd (if you count ram-setup variations) and ide/sata cables. i even tried unplugging my mouse and/or keyboard, no use of that too (restarts too), both old and new power-supply cables, and 3 different electricity jacks.
    again - i changed cpu, gpu, mainboard and psu.

    my current setup is:
    biostar a770 a2+
    athlon ii x2 7750+ be with phenom stock cooler
    1gb ocz platinum ddr2-800
    gigabyte geforce 7300gs
    160gb hdd
    asus drw
    saitek cyborg keyboard
    canyon wlan
    frontier achilles with xfx 450w psu

    bios info shows:
    CPU Vcore: 1.31V
    NV voltage: 1.10V
    +3.3V: 3.36V
    +5.0V: 4.94V
    +12.0V: 12.28V
    DDR voltage: 1.87V
    HT voltage: 1.21V
    5V (SB): 4.70V
    bios battery: 3.37V
    CPU temp: 28°C

    does anyone know would could be the problem, because only idea i currently have is my flat's electricity installations:S (but all other devices work fine, including my laptop!)

    P.S. i underclocked athlon ii be to minimal 1600MHz (without changing voltage), and restarts stopped
     
    Last edited: Nov 5, 2011
  2. tomba2k New Member

    Joined:
    May 11, 2009
    Messages:
    33 (0.02/day)
    Thanks Received:
    2
    Location:
    Croatia
    cmon guys, any ideas?
     
  3. white phantom

    white phantom New Member

    Joined:
    Oct 10, 2011
    Messages:
    209 (0.19/day)
    Thanks Received:
    35
    Location:
    Kilmarnock Scotland
    sounds like a very weird problem just going to take a pot shot but since you have not changed the hdd? could it maybe be an internal problem with the drive or the drives on its way out? or what about shorts? have you made sure nothing in the computer is making contact with the case or anything like a screw has fell into the back of the board shorting it out causing restarts?
     
    tomba2k says thanks.
  4. Thrackan

    Thrackan

    Joined:
    Oct 10, 2008
    Messages:
    3,482 (1.58/day)
    Thanks Received:
    656
    My first (and unfortunately only atm) guess is the PSU('s), but that's a bitch to prove.

    Did you reset the BIOS when you were doing your RAM changes?

    I would now try:
    - The Geil RAM ONLY (as the problem started after adding the OCZ RAM, and OCZ DDR-2... well.. it's famous for a reason.)
    - Use the 450w PSU just in case
    - Put it all together and keep the power unplugged
    - Take out the BIOS battery
    - push and hold the power button to drain any capacitor still doing something you don't want.
    - Put back the BIOS battery
    - Power on.

    If it consistently restarts ONLY in Windows after this, try disabling reboot on bluescreen to check a possible BSOD for error codes that might help you solve the problem.

    You can also try booting from a Linux livedisc without the HDD plugged in to rule out anything OS-related.
     
    tomba2k says thanks.
  5. JrRacinFan

    JrRacinFan Served 5k and counting ...

    Joined:
    Mar 17, 2007
    Messages:
    19,414 (6.98/day)
    Thanks Received:
    4,483
    Location:
    Youngstown, OH
    Ok so, you had this problem on 2 seperate setups with same psu? ^^ I am with thrackan if that's the case. I am just having a hard time understanding all of the backlog information.
     
  6. tomba2k New Member

    Joined:
    May 11, 2009
    Messages:
    33 (0.02/day)
    Thanks Received:
    2
    Location:
    Croatia
    about shorts - i cant be sure that all wiring is ok, especially in this situation, but i took both sides down from chassis and thoroughly examined/cleaned the case - no blind-headed screws were there.
    only cables remained after changing psu were drw PATA (80 conductors, although board recognize it as 40, probably because is drw is only drive - master) and hdd SATA data cables. (no power-cable adapters, only psu direct connections)
    i have 2 ouput fans (120mm arctic cooling f12 pro and xfx psu 120mm), and 2 input fans (one 120mm led fan in front, one 120mm regular on side - blowing air to cpu cooler).
    f12 is board controled, psu fan is psu controled, cpu assist and front input fans are 12v - full speed.

    both ocz & geil rams are CL4, but OCZ is a (very tiny) bit better with CL4 timings.
    before this setup, i had CL5 OCZ Gold XTC DDR2-800 (1x1GB) - never had problems with it.
    no bsod ever showed up. (and i did bsod on purpose 1 time on both boards just to check if bsod could not be showed/be showed to quickly)
    since biostar board is powering up usb all the time while power is on (even when system is off) i use 2.5" external hdd to drain capacitators after pulling the plug out.
    i always do that procedure when removing bios battery.
    i don't have any more following previous components: 4000+ amd cpu, ecs mainboard, 8800gts, because i needed some cash to cover expenses (7750+ amd cpu, biostar mbo, x1950 gpu).
    so, i bought 2nd hand x1950 (tested on another pc - working fine) while i'm awaiting for newly ordered hd6770 (coming in 1 week), just to exclude bad gpu problem (because i wasn't sure any more should i believe that my 8800gts and friends 7300gs are ok).

    no, i did setup changes in this order:
    ecs board->biostar board
    400w frontier psu->450w xfx psu
    nv 8800gts gpu->nv 7300gs gpu
    nv 7300gs gpu->ati x1950 gpu

    problem was consistent in all changes!(

    i will try to boot few live linux's without hdd pluged-in 1st thing in the morning.
    thank you very much guys for your effort!
    do you have any new ides after this few additional info's?

    btw, i tried to set dram mode to gaged (2x64bit instead of 128bit), tried to lower my HT to 1800MHz with 7750+ CPU default clock with default voltage and with board's cpu default+0.10V (because when i lower 7750+ 2.7 GHz cpu to 1.6GHz, HT sets to 1.8GHz), tried reseting PCI-E settings (through board's PCI-E slot reset option), tried lower DDR2 speed to DDR-667 (333MHz ddr option in bios), but i every try i have the problem.
    so, for now, i only get win xp & win7 running on cpu underclocked to boards minimal 1.6GHz.
    i'm really loosing my mind at his moment.
    i'll post here my linu live try in the morning.

    thank you again guys for your help!
     
    Last edited: Nov 4, 2011
  7. micropage7

    micropage7

    Joined:
    Mar 26, 2010
    Messages:
    5,949 (3.55/day)
    Thanks Received:
    1,380
    Location:
    Jakarta, Indonesia
    shortcircuit usually make your cpu shutdown in couple of seconds, i agree to check your psu or try to turn it on in minimal condition (mb, 1ram, 1odd, psu) and take the board off from case
     
    tomba2k says thanks.
  8. JrRacinFan

    JrRacinFan Served 5k and counting ...

    Joined:
    Mar 17, 2007
    Messages:
    19,414 (6.98/day)
    Thanks Received:
    4,483
    Location:
    Youngstown, OH
    Answered your own question then tomba with what you said then it sounds like CPU related more specifically IMC

    How old is your hdd and is this a new clean install of OS?
     
  9. tomba2k New Member

    Joined:
    May 11, 2009
    Messages:
    33 (0.02/day)
    Thanks Received:
    2
    Location:
    Croatia
    i was thinking about taking mbo out from case, but i have two issues with this:
    1) how to make sure that grounding is ok
    2) should i take pw swith from my case or build/buy some new power switch?

    hdd is about 5 years old now.
    as Thrackan said problem started with adding ocz ram, but only AFTER overclocking cpu to 2.4 instead of 2.3GHz (that setup was overclocked from 2.1->2.3GHz for about 5 years)
    fsb: original 200MHz, overclocked for 5 years to 220MHz, problem started when i overclocked it to 228MHz. (to get CPU/6->399MHz on 400MHz ddr2)
    reason why i was troubled with idea that gpu's might be damaged was that i wasn't sure was PCI-E speed unaffected, but x1950 gpu (according to previous owner) was never overclocked, 7750+ was never overclocked, and biostar board was never overclocked.
    strange about overclocking is that when i just got that setup, i wasn't able to clock it over 220 fsb. about a week ago, when i clocked fsb to 228MHz, i needed to set voltage to 1.4V (instead of default 1.3V on ecs board).
    setup worked for a few hours (while playing counter-strike 1.6 mix with my friends), than freezed.
    i manually restarted the system, and after the freeze and restart, i was able to clock fsb to 228MHz without changing the default 1.3V on cpu.
    that lasted about on hour.
    after that, problem appeared.
    now, overclocked board and cpu are gone, gpu on overclocked system is gone, but problem remains.
    it really sounds like electricity installation problem, but listen this:
    when i first time tried to overclock my pc, i tried it on 3 different locations (didn't planned that, it was just the coincidence), i had never got above 220MHz (didn't change the voltage then).
    now, on same place, i 1st could be able to clock in to 228MHz without voltage change, then i was able, than i could do anything without seriously underclocking my cpu speed.

    P.S. about OS "ageness" - problem is consistent on my old win7 installation, new xp installation (on 2nd partition) and new win7 installation.
    on installed system, restart happens after windows logo animation (just when login screen should appear) and during windows logo animation when installing new os instance (after text mode drivers load)
     
    Last edited: Nov 5, 2011
  10. tomba2k New Member

    Joined:
    May 11, 2009
    Messages:
    33 (0.02/day)
    Thanks Received:
    2
    Location:
    Croatia
    it seems that live linux (xubuntu) i tried last time freezed by accidance, because this time it worked both with hard disk plugged in or out.
    when i was testing system without hdd, i plugged out sata data cable from board and sata power cable from disc, put bios battery off and have done mentioned procedure about draining power off.
    anyway, xubuntu 8.04 loads without a problem, knoppix 5.1.1 freezes on boot (needs manual restart) and ubuntu 8.10 restarts on boot (same as windows).
    but, xubuntu doesn't shut down cleanly, it gives messed up screen (link)
    anyway, i booted xubuntu 8.04 from usb stick, knoppix and ubuntu 8.10 from cd.
    i tried to boot installed win xp with drw plugged off too.
    it doesn't seem that ata/sata drives cause the problem.

    and i still can't understand what exactly underclocking solves to prevent restarting:S
     
  11. Jetster

    Jetster

    Joined:
    Jan 17, 2010
    Messages:
    5,118 (2.94/day)
    Thanks Received:
    2,094
    Location:
    Oregon
    DDR2 does not like 4 Sticks. Go back to two and stock clocks. If that doesnt fix the problem then memtest86

    Classic case of changing two many things at one time. And why would you try to overclock onboard video?
     
  12. tomba2k New Member

    Joined:
    May 11, 2009
    Messages:
    33 (0.02/day)
    Thanks Received:
    2
    Location:
    Croatia
    to see how far does it go on ecs, cause it goes up to 750MHz on Asus's 780G^.^
    it seems you didn't read everything i wrote.
    i've changed components one by one, and i currently don't have any component that was overclocked in my current setup, yet i still have the same problem.
    and i have tried all combinations (1,2,3,4 mem sticks, in single and dual channel combos)
    anyway, i'm reluctant to believe that the best geil's and ocz's DDR2-800 (there's no better DDR2-800 than CL-4) rams should "dislike" being in 4x1GB dual channel combo:S

    this case is everything BUT classic, please read my upper posts if you have that much time.

    what i have noticed is that, when restart happens, although my keyboard always shows caps/scroll/num-lock leds correctly, most times does not respond to keypresses at all until i manually restart it again. and it happens only after problematic restarts (i never head any problem with my saitek keyboard before, i rather believe its up to board's "post-problematic restart" state)! strange:S

    P.S. by underclocking i mean putting clocks bellow default values.
    btw, before clocking my ol athlon from 220 to 228MHz, i tried to set manual timings for dr because ecs board kept setting it to CL5, and i've successfuly set all 4 rams to work in dual channel on CAS: 4, TRCD: 4, TRP: 4, tRTP: "3-5" (2nd option is "2-4" on ecs), TRAS: 12 and TRC: 23, in 2T mode (understandably, i couldn't get 1T for 4 sticks), and i had got 228MHz, but i had kept getting HT sync error if i tried to set mentioned CL4 timing and 14% fsb clock at the same time (although 14% overclock meant for DDR freq to be 2391/6=399MHz)

    and i repeat: my current 7750+ setup isn't and wasn't overclocked (including ram - they have never run above 400MHz), it was only underclocked to avoid system restarts!

    i've put all 4 modules back now, and underclocked again my athlon 7750+ to 1.6GHz, to reinstall once again windows 7 x64 pro.
    when installation finishes, i will run memtest while all 4 modules are in.
     
    Last edited: Nov 5, 2011
  13. micropage7

    micropage7

    Joined:
    Mar 26, 2010
    Messages:
    5,949 (3.55/day)
    Thanks Received:
    1,380
    Location:
    Jakarta, Indonesia
    should i take pw swith from
    my case or build/buy some new
    power switch?
    > not, you can turn it on by using pen and short pwr pins
     
    tomba2k says thanks.
  14. TC-man

    TC-man

    Joined:
    May 25, 2011
    Messages:
    348 (0.28/day)
    Thanks Received:
    117
    Location:
    Netherlands
    Hi,

    I think the first case with your old PC the restarts got to do with a weak/no name power supply unit, especially when you also tried to overclock it. Perhaps also try to lock the pci-e clock to 100 mhz (if you didn't yet), since overclocking the motherboard/cpu will affect the pci-e clock; a too high pci-e clock can cause anomalies and instability in all situations.

    Then the second case with the new Biostar motherboard, it may have something to do with memory modules compatibility: perhaps the Biostar motherboard does not like Geil and OCZ branded memory modules. Although you can try to flash a more up-to-date bios (if there's one) that may improve ram compatibility. Also perhaps try out other branded memory modules, e.g. Corsair, Kingston or G.Skill.

    Furthermore, how's the temperature of the processor? Make sure you apply some fresh thermal paste if you didn't. Dried-out thermal paste can cause the symptoms (e.g. when not re-applying thermalpaste when replacing/switching processors/cpu-coolers) you are having now.
     
    tomba2k says thanks.
  15. tomba2k New Member

    Joined:
    May 11, 2009
    Messages:
    33 (0.02/day)
    Thanks Received:
    2
    Location:
    Croatia
    pci-e is locked.
    i have tested my mbo, cpu, gpu and psu with friends ram, keyboard, mouse and monitor, without putting it in chassis (tested on table screwdriver as power-switch), so i determined that mbo is the faulty component.
    i bought 2nd hand MSI K9A2 Platinum, and everything is OK now.
    it seems that old psu damaged both ecs and biostar mainboards.
    on every cpu removal from the system i cleaned the paste and used again (as i always do) arctic cooling MX-2 or arctic silver 5 TIM's, although AC5 gives me only 2-3°C lower temp than without any (tried once - once both cpu and cooler surfaces are almost polished).

    basically, i use this ram setup because i am more confident with geil, corsair and ocz than with kingston or g.skill.

    anyway, my current setup is:

    my current setup is:
    MSI K9A2 Platinum
    Athlon II x2 7750+ be with phenom stock cooler
    2gb ocz platinum +2gb Geil BlackDragon ddr2-800
    ATi Radeon x1950 Pro 256mb ddr3
    160gb 3.5" hdd sata2 + 250gb 2.5" external sata2
    LG drw
    Saitek Cyborg Keyboard
    Canyon RT61 PCI & D-Link RT73 wlan
    Frontier Achilles with XFX Pro 450w psu

    and system is working fine.

    now, there is a problem that PCI wlan card cannot provide more than 512kbps download speed, but USB wlan adaptor works normall :S
    PCI wlan card is in the only PCI slot, gpu is in the first PCI-E slot (slot0).

    thank you again guys very much for confirming my thoughts and giving adeas, it helped me a lot to solve this situation.

    @admins: you can lock this topic now if needed.
     

Currently Active Users Viewing This Thread: 1 (0 members and 1 guest)

Share This Page