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

Screen Flickering

Discussion in 'AMD / ATI' started by Tintai, Jun 25, 2013.

  1. itsakjt

    itsakjt

    Joined:
    Oct 9, 2010
    Messages:
    1,173 (0.83/day)
    Thanks Received:
    381
    Location:
    Kolkata, India
    Well its definitely not a compatibility issue since it never occurred previously. So either 1, 2 or all 3 things can be at fault. 1.Monitor , 2. GPU , 3. Cable.

    I suggest you first try a different cable and then check the others. Best of luck. :)
  2. Tintai

    Tintai

    Joined:
    Apr 29, 2013
    Messages:
    380 (0.78/day)
    Thanks Received:
    55
    Location:
    Poland
    I don't think that problem is mobo, PSU or something else. That card is too hot and I have too many problems. On new mobo and my old mobo screen is flickering.
    Monitor, cable? I running 120Hz on Intel HD 3000 over then 10 hours and there is no problem on the same cable and monitor.

    I try later with DVI-SL.

    I think, yes.

    No. But this issue began after few months. And on the DVI-DL and 120Hz I always have a problems.

    And if the problem is in the heating VRM and when I change thermal pads problem will be fixed? Or in this case my card is permanently damaged?
    Last edited: Jun 27, 2013
  3. Law-II

    Law-II

    Joined:
    May 9, 2011
    Messages:
    1,566 (1.30/day)
    Thanks Received:
    641
    Location:
    Mainland Britain
    Hi

    This is possible the card is damaged; temp at 58 is no where near thermal max for the VRM; but will not know untill thermal pads are replaced [Tip: when replacing the thermal pads for the VRM's in both positions; change the pads on the memory at the same time]*Welcome to fault finding

    VRM General rule of thumb
    Keep these around 70~74 for everyday use and gaming
    Keep these under 70~74 for everyday use using H2o water block, third partie cooling
    Between 74~80 for Timed Benchmark; returning core and mem to lower clocks once finished
    Between 80~90 Short Benchmark with high core and mem clocks using third partie cooling
    Danger 90~100 Not for the faint harted and should be ringing alarm bells - will shorten the expected life span of the product, if used for everyday use
    Possible permanent damage/degrading of the VRM or failure 100~120 [this is why thermal throttling was bought in to protect the end user from themselves]

    I understand your frustration; however the reason I have asked to check the motherboard slot and the vga card pcie blades is to rule *every possibility out. If you do not follow this format to the letter and keep deciding what is best way, it is pointless trying to assist further.
    Example: This is how to trace a fault by the book

    When *testing the drivers use a driver from a few months back when everything was working well.

    Dose the vga card have a dual bios function? [if yes] move switch to position #2 while testing

    atb

    Law-II
    Last edited: Jun 27, 2013
    d1nky and Tintai say thanks.
  4. Tintai

    Tintai

    Joined:
    Apr 29, 2013
    Messages:
    380 (0.78/day)
    Thanks Received:
    55
    Location:
    Poland
    Yes dual bios. I flash second.
    Ok check PCIe slot when I back home.

    And my max VRM temp was 109C.
    Normally in gaming ~100C.
  5. Law-II

    Law-II

    Joined:
    May 9, 2011
    Messages:
    1,566 (1.30/day)
    Thanks Received:
    641
    Location:
    Mainland Britain
    Note: DO NOT Flash the vga card while in switch position #2 [*this is the reserve default bios]

    Thank you

    Edit: What application and version number; was used to monitor temps, GPU-Z?

    how long for ? any idea!

    That is to high; replace the Thermal Pads as soon as you have time
    Tripple check the vga Fansink mount If it dose this it needs some love and will require attention

    atb

    Law-II
    Last edited: Jun 27, 2013
  6. d1nky

    d1nky

    Joined:
    Jan 21, 2013
    Messages:
    3,627 (6.22/day)
    Thanks Received:
    1,248
    tintai i dont mean to be harsh but from what it seems the cards has defected from user error.

    your vrm temps have been a problem since you had the card and you also flashed the matrix bios. did you buy the card second hand? if so, the person probably knew about this and wanted rid. - my assumption.

    and everything i can think of has been said a few times.
  7. Tintai

    Tintai

    Joined:
    Apr 29, 2013
    Messages:
    380 (0.78/day)
    Thanks Received:
    55
    Location:
    Poland
    I mean "I flashed" but I have a backup and I don't use bios #2.
    Bios #1: 6798.15.12.0.4.AS05.U3
    Bios #2: 6798.15.12.0.4.AS05.U2
    Both bios are originally.

    I checked and carefully cleaned PCIe slot but this didn't help.

    GPU-Z, AIDA64, HWiNFO64.
    Newest.

    How long I have too height temp?
    I saw this first time when I tried oc my card and I saw that value of VRM is too hot. 105C when benchmark. It's was ~year after purchase.
    But before I play in many games(e.g BF3 - 109C max). But in this time I didn't know about this issue.

    I know. I must buy this 1mm and 0.5mm pads and install them.

    I never flashed to Matrix bios or I don't remember. Also I didn't use second bios. I use only first, original - never flashed.

    No. I buy it from shop but in the OEM version. Non BOX.
    Last edited: Jun 27, 2013
    Law-II says thanks.
  8. Law-II

    Law-II

    Joined:
    May 9, 2011
    Messages:
    1,566 (1.30/day)
    Thanks Received:
    641
    Location:
    Mainland Britain
    Hi

    thank you very much for all the info

    Well done; on keeping backups

    Thanks, I will make a note and tick this off the list of things to be done

    Benchmarks tend to push the VRM's to there limit and a temp as high as that in a benchmark would not be an issue as the benchmark is not normaly run over a prolong duration. An ingame benchmark would *under normal conditions not push the card to it's limit; saying that 109 in BF3 is I hope an exception to the rule.
    ;This said;
    I am starting to think that the GPU Core may have been *badly binned wafer and just pumped with extra voltage to get the card stable at 1000Mhz [may have been only 50Mhz away from it true limit]

    Question: How high did the core go when overclocking?

    At this stage my gut feeling, is the original thermal pads on the vga card *may have been poorly fitted, dryed out and or fansink has an issue with PSB contact for the VRM to run that high in normal use.

    Wayforward:
    At this time is to get the thermal pads replaced, reply to this post once fitted.

    atb

    Law-II
    cadaveca and Tintai say thanks.
  9. Tintai

    Tintai

    Joined:
    Apr 29, 2013
    Messages:
    380 (0.78/day)
    Thanks Received:
    55
    Location:
    Poland
    Yes but on benchmark max temp is 105C on 1080MHz core(fans ~70%) and on the Battlefield 3 is ~109C on the stock core(fans auto). Idle(fans auto) ~60C.

    This pads will be ok?
    Memory
    VRM
    It's all what I need?

    And thanks for your great help. I'm so glad :)
  10. D007

    D007

    Joined:
    Mar 7, 2007
    Messages:
    3,068 (1.12/day)
    Thanks Received:
    377
    Location:
    Pompano beach, Florida
    Native resolution of the monitor is progressive scan yes?
    Maybe you have it set to something like 1080i instead of 1080p.
    I know that can cause issues.

    Try running @ 59 & 60 hz refresh rate.

    PS: 105 c is not good.
    No matter the system or duration. If you hit 105c something is wrong imo.

    gl ^^
  11. Law-II

    Law-II

    Joined:
    May 9, 2011
    Messages:
    1,566 (1.30/day)
    Thanks Received:
    641
    Location:
    Mainland Britain
    Hi

    For thermal pad thikness refer here

    Advice
    two strips of 120 x 20 x 0.5mm Phobya; two strips of 120 x 20 x 1.0mm Phobya & cut to size
    [Note: may not all be required; however if the Fansink dos not make good contact with the VRM's and MEM the 0.5 will be invaluable]

    Edit: MX4 thermal paste and or other non-conductive thermal compound for GPU; [Note: non lint based pad or coffee filter to remove old thermal past; work from the outer edge to the middle of the GPU to aviod spreading thermal compound] use pea; or line method when placing MX4
    Example:

    Make sure you have the best tools for the job, ensure you have screwdrivers that fit perfectly in the *screwhead [*stripping would be a disaster]; an antistatic bag to place the vga card on with a box underneath; good light; scissors and or a sharp blade to cut the thermal pad to size; ruler or method to size the pads.

    [Note: removing the old pads do these one at a time so you keep track of which is which]

    When removing old pads Use a lint free pad or coffee filter to clean any excess residue from the VRM's and MEM modules.

    When fitting the new pads; once they are in the correct place fit the Fansink to the card and squeeze it together [Note: place the card between an anit static bag; aviod direct contact with the PSB]
    Next: remove the fansink and check the thermal pads for impression/indentations of the VRM's and MEM modules. [if you are happy that the fansink has made good contact; adjust/realine the thermal pads and proceed to fit the fansink] Note: do not over tighten the screws; *do not want to strip a thread.

    atb

    Law-II
    Last edited: Jun 28, 2013
    Tintai says thanks.
  12. Deadlyraver

    Joined:
    Nov 17, 2011
    Messages:
    270 (0.27/day)
    Thanks Received:
    23
    I had a flickering issue that was minor compared to yours. Found that the ATX cable I used was hindered by an older extension. Switched to a whole cable and it worked.
  13. Tintai

    Tintai

    Joined:
    Apr 29, 2013
    Messages:
    380 (0.78/day)
    Thanks Received:
    55
    Location:
    Poland
    I try everything with this and nothing worked

    Ah ok. I need all needed tools. Thank you.

    I don't understand. What ATX cable has to this?
  14. Tintai

    Tintai

    Joined:
    Apr 29, 2013
    Messages:
    380 (0.78/day)
    Thanks Received:
    55
    Location:
    Poland
    Ok, so I change all thermal pads and past on core(originally were all pads).
    I carefully clean all PCB and nothing. VRM1 temp is still to height. Now I waiting for flickering ;p

    So this problem is not issue with thermal pads. Now everything should be better.

    Here is the photo:
    [​IMG]
    What is this marked on red? I should put pads on this chips? @edit: I put pads on this marked chips. I looked bad.

    Now the temps are: (idle)
    Diode: 37C (I think is same or similar)
    VRM1: 63C (same)
    VRM2: 26C (~6C better)

    And I check temps in BF3.

    PS. 27 minutes and I still can't see any flickering. Hm.. I'm still waiting.
    Last edited: Jul 6, 2013
    Law-II says thanks.
  15. d1nky

    d1nky

    Joined:
    Jan 21, 2013
    Messages:
    3,627 (6.22/day)
    Thanks Received:
    1,248
    nooooo thats vram = video memory


    see the line of capicitors and chokes with print on them.

    what you have covered in a grey pad, thats vrms! they need thermal pads!


    research mosfets on gpu pcb! there are also two more near to where you have highlighted in red, next to the chokes!



    z8n8.jpg

    that grey pad doesnt seem to be the right one if it was the asus one, hence high vrm temps, most ive seen are white or pink
    Last edited: Jul 6, 2013
  16. Tintai

    Tintai

    Joined:
    Apr 29, 2013
    Messages:
    380 (0.78/day)
    Thanks Received:
    55
    Location:
    Poland
    I replace all pads.
    This big gray I change to 1mm. Others pink I change to 0.5mm.

    Ahh sorry. I looked bad. So I change this what I marked too.

    And something must wrong with this VRM1. I replace this and VRM2 have better temp then had before.
  17. d1nky

    d1nky

    Joined:
    Jan 21, 2013
    Messages:
    3,627 (6.22/day)
    Thanks Received:
    1,248
    change them all if you need/want too.

    i thought ya was just changing the vram, well change them all. make sure the heatsink and brace is down tight let it burn in for a bit.


    see in my pic you want to put pads on the 2 far away from the rest next to the chokes, i believe thats vrm1.

    i havent got gpuz installed so cant tell ya and ive forgotten.
  18. Tintai

    Tintai

    Joined:
    Apr 29, 2013
    Messages:
    380 (0.78/day)
    Thanks Received:
    55
    Location:
    Poland
    I know. I replace all pads and everything is fine.
    This gray pad was not good.

    VRM1 is next to the VRM2, yes? So now VRM1 is 63C and VRM2 is 26C. And 26C is much better then before.
  19. d1nky

    d1nky

    Joined:
    Jan 21, 2013
    Messages:
    3,627 (6.22/day)
    Thanks Received:
    1,248
    no they are seperate.

    where i marked on the pics.
  20. Tintai

    Tintai

    Joined:
    Apr 29, 2013
    Messages:
    380 (0.78/day)
    Thanks Received:
    55
    Location:
    Poland
    So there are pads too.
  21. d1nky

    d1nky

    Joined:
    Jan 21, 2013
    Messages:
    3,627 (6.22/day)
    Thanks Received:
    1,248
    yea there should be!

    any driver mosfet should have a pad on.
    Tintai says thanks.
  22. Tintai

    Tintai

    Joined:
    Apr 29, 2013
    Messages:
    380 (0.78/day)
    Thanks Received:
    55
    Location:
    Poland
    Ok, when I back home I take off heatstick(It's relay easy) and I see again.
    Thank you :)

    @edit:
    I opened it and I check all pads. I added one pad but this change nothing. I can't add more pads.
    VRM1 temperature is still the same.

    Please tell me how long my card will be live? I must prepare to buy another.

    @edit2:
    And I have a good news(I thing). Up time 1h, 3m and still no flickering.
    Last edited: Jul 6, 2013
  23. Law-II

    Law-II

    Joined:
    May 9, 2011
    Messages:
    1,566 (1.30/day)
    Thanks Received:
    641
    Location:
    Mainland Britain
    Hi

    Flicker Fixed, well done

    Question: What temps are you getting when running BF3? [log temps with GPU-Z] thanks.

    nb: had a great time at the BBQ, to much wine to answer all qestions at the moa; will be back online on Sunday

    atb

    Law-II
  24. d1nky

    d1nky

    Joined:
    Jan 21, 2013
    Messages:
    3,627 (6.22/day)
    Thanks Received:
    1,248
    vrm1 sounds fishy!

    but if the flickering has gone then its good, just up the fan profile/dont overclock to keep that vrm temp down.

    maybe the baseplate isnt connecting properly to the vrm to make good contact for heat exchange, who knows!

    come to think of it, the flickering could of been something totally unrelated, bad gold finger contact, dirt etc etc

    and LAWII is too drunk?! :toast:
  25. Tintai

    Tintai

    Joined:
    Apr 29, 2013
    Messages:
    380 (0.78/day)
    Thanks Received:
    55
    Location:
    Poland
    For now. Will see. Thanks.

    I check in AIDA64 and on my G19 monitor.
    GPU Diode: 72C max (same or similar)
    VRM1: 102C max (better. before was 109C max)
    VRM2: 26C :)eek:)

    Perhaps. I saw this chip from VRM1 don't touch any radiator. Pad on this chip touch only piece of baseplate. And this second chip(to the right on d1nky screen) don't touch any radiator.

    :toast:

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

Share This Page