Page 1 of 2 12 LastLast
Results 1 to 25 of 47

Thread: 8RDA and GF4 dying solution within...

  1. #1
    Technician
    Join Date
    Jun 2002
    Location
    Merseyside, UK
    Posts
    2,661

    8RDA and GF4 dying solution within...

    A little awareness guys. Thats all.

    Take a look at your board. Now take a look at the clearance between dimm slot 1's clip and the card. Thats right, about 0.2mms. Now look at whats missing off your GF4.....

    Because of the length of the GF4 and the proximity of the dimm socket clip theres a tiny transistor that gets knocked off if you install the card straight. What you have to do is bend the back of the card slightly away from the dimm slot when you put it in so the edge of the dimm clip clears the transistor on the GF4.

    The 8RDA isnt killing boards - but the design leaves a lot to be desired. Its kind of hard to explain without my camera handy to take a picture and sho you all, but I had the same thing happen to my friend yesterday. Perfectly working card, installed in the 8RDA+ and it was dead. Tried another board - dead. Examined the card against a known working Ti4600 after I noticed how close it was to the dimm clips and sure enough the transistor had been knocked off it.

    Spread this knowledge around please lads or a lot of people are going to be VERY upset over christmas.

    Ill say again, all you HAVE to do when installing a GF4 is to bend the back part of the GF4 slightly away from the dimm clips (about 1mm if that) when you plug it in. When its in the transistor is below the dimm clip and not being touched at all. Remember to do the same if you take the card out too btw .

    Take care all and have a good christmas.

    PiLsY.

  2. #2
    Xtreme Addict
    Join Date
    Nov 2002
    Location
    MA.
    Posts
    1,783
    Thank You VERY much Pilsy. I don't have a geforce4, but have read alot of this killing crap........
    I'll spread the word.

    Zen2 Has brought AMD back!

  3. #3
    Overclocking Guru
    Join Date
    May 2002
    Location
    SoCal, USA
    Posts
    5,128
    Excellent observation PiLsY

    So much for the mystery of blowing out AGP cards

    OPP

  4. #4
    Xtreme Addict
    Join Date
    Nov 2002
    Location
    MA.
    Posts
    1,783
    Too bad another gf4 had to die because of this oversight on Epox's part though. I am wondering if they changed the design of the dimm clip or moved something ( they may have seen this AFTER a ton of boards were made, and revised it quickly ). hence the 1.1 revision??
    Zen2 Has brought AMD back!

  5. #5
    Xtreme Mentor
    Join Date
    Jun 2002
    Location
    PA, USA
    Posts
    2,662
    Interesting. Hopefully those who've lost their cards can verify this. Thanks for the heads-up PiLsY
    E8600. EP45-UD3P. 500FSB 24/7. Still WinXP. Heatware

  6. #6
    Xtreme Member
    Join Date
    Nov 2002
    Location
    Berkeley, CA - USA
    Posts
    115
    That would also somehow explain why none of the ATI cards have been dying in the 8rda+

    --fibo
    Epox 8RDA3G+ v2.0
    AXP 2100 "B" (11.5 x 200fsb) - 2.3GHZ
    ATI Radeon XT800 XT (Omega Cat 5.8)
    3x512MB Corsair XMS 3200 3-2-2-5 @ 2.77v
    SLK-800 w/ SmartFan II
    Windows XP Pro w/ SP1
    2 Seagate Barracuda 120GB 7200rmp 8MB
    1 LiteOn CD/RW - 1 Pioneer DVD
    Anix Platinum XP Case modded

  7. #7
    Xtremely Pale
    Join Date
    Dec 2002
    Location
    new jersey
    Posts
    493
    i have a question dont the 4200 have this problem too the reason that i say this is that the 4200 is a shorter card than the 4400 and the 4600 i have a 4200 and my 8dra is getting opened up tomaro night and i hope your right about this but how can this be the awnser for two different lenghts of cards?

  8. #8
    hwbot crew
    Join Date
    Jun 2002
    Location
    Belgium!
    Posts
    880
    Thank you for clearing that up, pilsy!


    @doyt: my suspicion is that all who have a dead GF4 ti4200 due to an epox 8rda+, have an enhanced version of a Ti4200. A 10 layer pcb design with a cheap ti4200 on it (eg. Albatron Ti4200P?) would have the same problem as a ti4400/4600.
    HTPC (win xp): Turion MT-30 @ 2Ghz | NF4 | XFX 7900GT | 26" TFT
    Development (mac osx): Macbook Pro | Core Duo 1.86Ghz | 1.5GB DDR2


  9. #9
    Xtreme Member
    Join Date
    Nov 2002
    Location
    Berkeley, CA - USA
    Posts
    115
    Originally posted by doyt
    i have a question dont the 4200 have this problem too the reason that i say this is that the 4200 is a shorter card than the 4400 and the 4600 i have a 4200 and my 8dra is getting opened up tomaro night and i hope your right about this but how can this be the awnser for two different lenghts of cards?
    There are a few TI4200's out there that have the same size characteristics than a TI4400 or TI4600 (i.e. Albatron TI4200 Turbo)

    --fibo
    Epox 8RDA3G+ v2.0
    AXP 2100 "B" (11.5 x 200fsb) - 2.3GHZ
    ATI Radeon XT800 XT (Omega Cat 5.8)
    3x512MB Corsair XMS 3200 3-2-2-5 @ 2.77v
    SLK-800 w/ SmartFan II
    Windows XP Pro w/ SP1
    2 Seagate Barracuda 120GB 7200rmp 8MB
    1 LiteOn CD/RW - 1 Pioneer DVD
    Anix Platinum XP Case modded

  10. #10
    Xtreme Enthusiast
    Join Date
    Nov 2002
    Location
    San Diego
    Posts
    796
    i noticed this when i installed my 4600, i also was careful to avoid hitting that transistor...Just bent it back a tid and pushed it in...not problems so far except that this Mb can not do above 190 FSB stable...

  11. #11
    Xtreme Member
    Join Date
    Nov 2002
    Location
    Sweden
    Posts
    141
    I´v had one card dying on me.
    Visiontek GF4 Ti4600.

    RMA`d it already but have dropped a mail to the shop where I bought it.
    Hopefully they can check out if the transistor is missing.

    Happy Holiday´s!
    Caper

  12. #12
    Xtreme Enthusiast
    Join Date
    Nov 2002
    Location
    San Diego
    Posts
    796
    oh yea, and its not enough to make me return it for a rev1.1 board....

    I think they difference ont eh rev1.1 board is that they moved the Memory slots up a little or lowered something else...Not positive though but acording to aoa nazis the difference was only cosmetic.

  13. #13
    Xtreme Member
    Join Date
    Dec 2002
    Posts
    345
    very good observation there...i was a bit hesitant buying this board (even though i have the 9700 pro), but may reconsider it...

  14. #14
    Registered User
    Join Date
    Jul 2002
    Posts
    2
    I can't acknowledge this theory ... I'm have somewhat older pics of my Leadtek A250TD (GF4Ti4400) ... and some pics which were made directly before dispatch to Leadtek ... I compared this pics ... I can't see missing components within the questionable area ... and if components had so obviously been missing, Leadtek wouldn't have exchanged the card within 48 hours!

    @PiLsY:

    IMHO a nice attempt of explanation, but doesn't look like this is the true reason for all the smoked GF4!
    Last edited by Stanley Jobson; 12-23-2002 at 08:16 PM.

  15. #15
    Registered User
    Join Date
    Nov 2002
    Location
    Texas
    Posts
    88
    There dosen't seem to be any difference in the spacing between the v1.0 and v1.1 boards. I have both, dimm seems to be in the same place on both boards.
    Abandoned [H] 11/14/02

  16. #16
    Registered User
    Join Date
    Dec 2002
    Location
    Australia
    Posts
    2
    ahh... thanx *very* much...
    I just got my Albatron GF4 Ti4200P Turbo & am waiting for my 8RDA+...

    I'm not really sure about "bending" the geforce4.. as maybe if i move my computer around (as i do), the card wi ll bend back and hit the clip?

    Anyways -- i know this sound stupid, but can i possibly get a hacksaw and saw off the protruding bit off the DIMM clip?
    sounds a bit idiotic , but it may do the job... any suggestions?

    thanx!

  17. #17
    Technician
    Join Date
    Jun 2002
    Location
    Merseyside, UK
    Posts
    2,661
    Stanley - the transistor in question is about 1mm x 2mm square. I very much doubt youd even notice if it was missing unless you were looking directly at the spot where it was supposed to be.

    SYNer - just bend the card m8. It wont hurt it, infact you probably wouldnt even notice but they bend anyway - designed to.

    The other alterantive is pulling the dimm clip out of its mounting. Easy enough to do, but means you wont be able to use slot 1. I certainly wouldnt take a hacksaw to your motherboard - warranty void, and one slip and its all over.

    PiLsY.

  18. #18
    Technician
    Join Date
    Jun 2002
    Location
    Merseyside, UK
    Posts
    2,661
    Found a picture...

  19. #19
    Registered User
    Join Date
    Dec 2002
    Posts
    2
    Good find PiLsY

    I was one of the few who lost a g4 4600.
    My transistor is missing.

  20. #20
    Registered User
    Join Date
    Dec 2002
    Location
    Australia
    Posts
    2
    ok thanx a heap PiLsY

  21. #21
    Technician
    Join Date
    Jun 2002
    Location
    Merseyside, UK
    Posts
    2,661
    Looks like thats confirmation then...

    kuzma - I'll try and find out exactly what that transistor is and get a part number. The card is not damaged in the least apart from that. Solder a new one on and itll be as good as new .

    PiLsY.

  22. #22
    Registered User
    Join Date
    Dec 2002
    Posts
    2
    Thanks PiLsy, if there is some way of fixing the card, that would be great. It was a Visiontek card, so rma was very unlikely.

  23. #23
    Xtreme Member
    Join Date
    Jun 2002
    Location
    Sweden
    Posts
    182
    Kuzma : Nah, Visiontek is back, and they are honouring their RMA's, there are confirmed reports of this. So just RMA it, you will get a new card.

    /Damir

  24. #24
    Xtremely Pale
    Join Date
    Dec 2002
    Location
    new jersey
    Posts
    493
    well i have just hoked up my 8rda+ rev 1.1 and 4200 not turbo
    and no problems here the card dosent even come close to the dimms so that helps to further prove the transistor theory

  25. #25
    Technician
    Join Date
    Jun 2002
    Location
    Merseyside, UK
    Posts
    2,661
    Just checked against the lists over at AOA and OCUK, and all the cards reported dead are the full length PCB versions.

    Although weve had no confirmation from Epox, Kuzma's report seems to acknowledge that there is at least a problem here. Combined with only full length PCB cards having been reported dead so far id stake my reputation on it (although having said that I bet I come to regret it ).

    PiLsY.

Page 1 of 2 12 LastLast

Bookmarks

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •