Announcement

Collapse
No announcement yet.

Abit KD7-G with Teapos...

Collapse
X
Collapse
 
  • Page of 1
  • Filter
  • Time
  • Show
Clear All
new posts
Previous template Next
    #1

    Abit KD7-G with Teapos...

    I bought this board from NewEgg secondhand sale and used as folding machine for over 1,5 years now. From start, I tried 2500Mhz (166x15), however it did not last long. I have to back down at 2330 (166x14 - nothing between) with 1.725Vcore and AXP-M 2500+. After about year of usage I asked on Extremesystems forums how to improve the Vcore regulation to get 2500Mhz stable again and installed a resistor on the control chip. No improve, tough... And month later after some previous quirks the integrated gigabite NIC give up. I disabled it in windows and added a PCI NIC. Yesterday it crashed and since I know the Teapos inside (one even bulging), I waste no time and recapped it. First pics of what was before:


    This one near ram slots is even bulging - sadly it is not much visible on the photo, but take as reference the second Teapo to the right and you will see it. So much for "not bulging/leaking" Teapos


    Teapos everywhere:


    Even the small ones are Teapos!


    But Vcore caps are Rubycons! Sadly MBZ and I would love to see 40% better MCZ ones there, but... One can't have everything...


    And how much Rubycons help, when all around is the cancer called Teapo?!


    At this point I started to think that maybe is was the Teapo(s), who killed my integrated NIC...


    Or maybe it was this mod? Or the Vcore mosfets hardcore cooling I added? These caps must be pretty happy from it, however they are Nichicons and they made it


    And as usual, someone skiped on ceramics in the CPU socket - fixed with 4 4.7uF X7R ones


    So, the board is very interesting mixture of Teapos (grrr!), Nichicons (whoa!) and Rubycons (yup!) - so, to recap all the Teapos, one need:

    Abit KD7G
    ---------
    3x 2200uF 6.3V d10 - P12344-ND
    5x 1500uF 6.3V d8 - P12343-ND - d10!
    10x 1200uF 6.3V d8 - P12342-ND
    13x 120uF 16V d6 - P12922-ND

    I intended to bump the capacity a little - 100 go to 120uF, 1000 go to 1200 and rest stay. However Digi-key made a mistake and send me other caps (not 1200uF 6.3V d8 ones, but 1000uF 10V - d10 ...!) so I gut the stash ready for Epox 8RDA+ instead and used 15x 1500uF 6.3V d8 Samxons GC and 3x 2200uF 10V Samxons GC together with 13x 120uF Panny FM ones

    Aditionaly, I added the mentioned 4 ceramics into CPU socket. But it did not end there. I also added 100nF ceramics in parallel (bridged the caps) to each of the five Vcore Rubycon caps. Also I added these 100nF 16V ceramics to block the 12V power line - one per wire, so, two - to make the voltage clean. Futhermore I added bellow the main PSU connector on board 2x these 5.6 X5R 6.3V ceramics to block out the 3.3V and 5V line as well, as one 100nF 16V one to block the 12V line

    Some pictures:









    Naturally, LPT and COM ports has to go in favor of great airflow of the hot air out of case


    But the NIC won't kicked back again...


    Not even when I removed the resistor mod...

    At least these Pannyes look neat:


    ...and machine is back again stable - at least it won't crashed yet in these few hours and I hope it never will

    Yet I'm certain that PSU recap will be necessary too - the oldie Enermax 431W is full of, well, bad caps...
    "It is dangerous to be right in matters on which the established authorities are wrong." - Voltaire
    "I believe that all the people who stand to profit by a war and who help provoke it should be shot on the first day it starts..." - Hemingway my config - my caps
    Tags: None
    #2
    Re: Abit KD7-G with Teapos...

    I've got this board and it does not work properly. After clearing CMOS, it comes to POST screen and stalls there (after displaying memory size and speed) - but it does not hang, when I press DEL, it goes into SETUP. But when I save the settings (without changing anything) and reboot, it does not come up again - hangs with POST CODE 26h (and no, it's not caused by the BIOS bug as said on Abit page - the BIOS is the latest on and the CPU is Duron 700).
    I noticed that one of the MOSFETs in 2.5V supply has its GATE shorted to GND - by a blob of solder near the HIP6521 chip. The trace from the chip was cut and soldered to GND instead (and the solder work was really bad). I see that on your board too, although soldered much better:

    I wonder if that's some design fault or something?
    I removed that solder blob and soldered the trace back together - and no change in behavior.

    Comment

      #3
      Re: Abit KD7-G with Teapos...

      Well, there is some things at least weird on the mobo, sure thing. However if you can get to the bios, then setup into the first O/C menu CPU command decode from fast (default) to normal - or your machine won't post again on next reboot, as you are experiencing

      Hope this help you.

      Dunno why, but it does work like that for me pretty stable - in fact, it crashed only once, and that was thx to Teapos...
      "It is dangerous to be right in matters on which the established authorities are wrong." - Voltaire
      "I believe that all the people who stand to profit by a war and who help provoke it should be shot on the first day it starts..." - Hemingway my config - my caps

      Comment

        #4
        Re: Abit KD7-G with Teapos...

        I found that the "hang" during POST can be solved by setting DRAM drive strength to High.
        Setting "Halt On Error" to "No error" and rebooting with INS pressed allowed me to boot.
        Are the Abit BIOS programmers so stupid that the default settings don't work?
        I'll look at the CPU command decode - haven't touched that yet...

        Comment

          #5
          Re: Abit KD7-G with Teapos...

          Interesting. I must admit that I did not touch the drive strenght, yet anything under 2.85V to my rams cause reboot problems... (as well, as Prime95 problems...) At 2.65 it mostly refuse to post, hehe

          Could be related? But once I get my KD7-G stable, I won't toughing it again. And maybe you not need to touch the CPU command decode, because you using Duron - I using AXP-M CPU... (this is, according to the bios a "unknown CPU", hehe But working pretty well still )
          "It is dangerous to be right in matters on which the established authorities are wrong." - Voltaire
          "I believe that all the people who stand to profit by a war and who help provoke it should be shot on the first day it starts..." - Hemingway my config - my caps

          Comment

            #6
            Re: Abit KD7-G with Teapos...

            Trodas, how do you remove the com and lpt ports so well?? Do you unsolder them, or what? Nice work. BTW, who do you fold for?? Sorry for the OT guys.

            Comment

              #7
              Re: Abit KD7-G with Teapos...

              Default value of CPU command decode is Normal... Changing it to fast does not help.

              Comment

                #8
                Re: Abit KD7-G with Teapos...

                Just FYI post code 26h reserved...

                The next one is:

                KBC final Init; Final Initial KBC and setup BIOS data area
                Last edited by Per Hansson; 07-17-2006, 04:39 AM.
                "The one who says it cannot be done should never interrupt the one who is doing it."

                Comment

                  #9
                  Re: Abit KD7-G with Teapos...

                  moddolicous - well, it is not that hard when you know, how to do it. If there is interest, I might agree to recap one older Socket A bad board and do a little photographic guide how to remove these - or at least how do I do it
                  To put it short, first you have to decapitate and destroy what is holding it all together - and then you just unsolder the pins and support plates and that it is. When you do it seven times like I did already, you feel like a pro doing it - easy - just the desoldering part of these grounded LPT ports is quite booring one... (these has to be heated up for about 12-16sec by 75W soldering iron - no fun...)
                  And I do fold for SPCR - Silent computers, hehe. Team number: 31574. They got me first and convince to fold + I abobe all respect silence and using mostly fanless designs to archieve that, so... Silent computers are thing for me

                  http://vspx27.stanford.edu/cgi-bin/m...sername=trodas
                  http://folding.extremeoverclocking.c...php?s=&u=78645


                  Rainbow - well, it act then weird with AXP or AXP-M only? I don't know. But for me, the fast is default there and leaving it at fast result in non-booting machine I bugged Abit about it, however as usual - no response
                  "It is dangerous to be right in matters on which the established authorities are wrong." - Voltaire
                  "I believe that all the people who stand to profit by a war and who help provoke it should be shot on the first day it starts..." - Hemingway my config - my caps

                  Comment

                  Previous template Next
                  Working...
                  X