borked 480

i really messed up on this

I installed a cap on c1045 + vmem/vgpu read point but i had instability problems after, so i powered off the system and twisted the cap a bit until the legs broke off to remove it. But i think it made things worst, after that the few times i made it to windows i got a green(yellow,pick your color) after a few minutes.

So i decided to remove everythings i installed before the problems started but here's the second problem i think i heated the pcb too much and the damn thing just fell off in 2 pieces. See pic.

The question is where can i get those things?
i measured resisance across the others and i get 0.4ohm, does that sounds right? seems really low

Even after putting it back i'm pretty sure it's dead(deadish), but i won't gonna trash before i try.

anyone can help me? :(
maybe i can be saved

http://imgur.com/ulAzcl.jpg

i trashed a few vga before but this one is really the icing on the cake, i guess that will teach me to do modding late in the night.

I'm pretty sure that's a cap, not a resistor :shrug:

Unsolder another of these, measure capacitance, buy or find another similar cap, resolder, voila!

burnt out caps...pretty much open as the dielectric is already exposed.

dead ceramic capacitor. You should actually be able to still run without it, as it is to help with the tantalum capacitor beside it there.


Seems like you shorted the pad for the ceramic cap, and the tantalum cap (big black flat one). so use a solder sucker and clean that area up a bit.

Also though what could have happenned is, that depending on the uf rating of the cap you bypassed the ceramic one with, it could have damaged the power delivery circuits from the inrush current.


for sure though go and remove all mods, and use a solder sucker to clean up all excess solder, and look for any that might have dropped onto the board. then when it is going again, start the mods one by one.


you from quebec?

how can one small cap kill a power circuit? its a small cap. its prolly just a coupling cap

btw its the same cap from the other phases...

So he's shorting out that area because he ate away the dielectric surface?

he ate away the dielectric surface....what are you talking about?

it looks like a solder bridge from C576 to the pads of the tantalum cap to the left. hard to tell though, because the red line is right over the area.

And at honda.

It wouldnt be the lack of the cap, it would be the cap he put on there, those ceramic caps are probably .1uf if that, he probably put in an electrolytic cap of 680-1200uf. that would cause the inrush current to change drastically because now there is ton of extra capacitance that needs to charge up, which could burn something out.

Ive seen it happen in audio amplifiers.

oh yeah he did add a cap as mentioned..

but adding a cap to the load side won't really affect the charge/discharge all of those caps are in parallel just before the load side..

as mentioned, he did overheat the solder joint...most smd's can only last 5-10seconds of say 500c of uncontrolled solder tip.

I can't offer anything onto whats already been said, but ouch, that could be some real expensive mistake made there :(

thanks for the inputs guys, it really helps

just to be sure
1:perfectly fine with vmem mod
2:added cap @ c1045 + vgpu/vmem read point
3:instabilty at known stable frequency
4:removing the cap by twsting it, power was off(hit the power button a few times),but the card was still in the mobo.
5:booting but for no more than 5-10min
6:removing everythings i've added @ 2 = no boot+cap fell off
7:in deep sh1t
8:crying alone in the dark

for some reasons i though it was a resistor LOL

just to be sure the cap i've added wasn't there, it's located at C1045 just like here. > http://pepinorang.jr.free.fr/screens...0/DSC02424.JPG.

i tried to run the card yesterday and no go, mobo gives a big beep at "2d" post code and stays there. before i removed everything and the cap fell off i was able to boot 1 in 5 times or so.


another thing, i tried to measure capacitance but my dmm only reads in "f" not in "uf", but i'm surprise it's a nice model, not a top of line fluke but i wasn't expecting that this one woul not do everything i needed. So now i guess i need a new dmm which doesn't suck and some old mobo vga to get a cap of the same value. But i'm not holding my breath for it to be working. By g.foyle post i seem to understand that i can't measure capacitance with cap solderer?

WeStSiDePLaYa
Yes i'm from quebec, you too?
i'm 99.9% positive it's not shorting right now, the pic sucks i know. But i was pretty sure it wasn't after soldering the vgpu read point either, so who knows.

ket don't put me down more than i am :)

2d hmmmm no output device initialize? are you doing subzero?

F is much bigger than uF...it should be able to read it capacitance on the other phase...

when measuring capacitance..you have to wait 3-60 seconds depending on capacitance...

Quote:

---

Originally Posted by mk-ultra

ket don't put me down more than i am :)

---

My bad. Lets all have our fingers crossed a operation will fully cure her :up:

http://2.bp.blogspot.com/_uislBepRAK...hChefSmall.jpg

well guess what, I'm posting that on a gtx480 powered system :D

more news in a few minutes
EDIT:
only one boot and that's it, second boot nearly went to windows, blank screen beforegetting the welcome screen
third+ boot = 2d with short beep, long beep, no beep at all, pick one.

uhmmmm, i'm gonna try it in another PC
edit wait second pc does not have 8+6pin

Let me join your "borked 480" club.

Three caps, almost the same place as OP's card i don't know how it happened. Won't boot at my 24/7 (4.5Ghz) clocks, it'll only boot at max 3.6Ghz, card works fine without drivers installed or in safe mode, with drivers installed it crashes right after the welcome screen. I've resoldered two of them already, the biggest one and the one that's broken off on one side but i'm having such a hard time with the last one. The UD5 PCI-E unlock mod is a a piece of cake compared to this one...

http://img823.imageshack.us/img823/2913/borked480.jpg

Remove all the solder from the modded area, ensure no pads/components are bridged, and see if you can find a capacitor the same rating.

I'd strongly advise you to stop trying to boot it up as it is, it's not going to magically fix itself.

Also, a couple of pointers: Never 'twist a cap to break it off', always unsolder it carefully! You need to be gentle and use common sense when modding electronics. Also don't modify one power phase and leave the others untouched - you don't want asymmetry in a power circuit like that. Adding one or 2 caps to random phases isn't going to do anything positive, they must be added carefully and be a low ESR/correct capacitance cap for that particular part of the circuit.

Best of luck, I hope you fix it.

killed a few gpu's in my time... definatly feel your pain... good luck with the restoration dude...

thanks for the tip HC, didn't know i had to wait to get a reading..

one thing i find strange tho, i have many smd or small caps from old boards and they all get a reading. But all the caps i took from my old asus p3 mobo my dmm reads "OL" just like when you measure resistance and the probes doesn't touch anything. Same thing for the similar caps i've burned on the 480.

Cryptik,
i know i should have been more careful and take my time, live and learn i guess.
i've learned from my mistakes this time and won't do it again.

btw my mobo does the same thing with other video card, sucessful boot aftera few times, but i get a blank screen after 10-20minutes, she started acting weird after my last cold session but right now it loks like a rma case.

purecain,
i already have 3 murder under my belt, this one will be the forth if i can't save it

what capacitance capacitors did you guys add? it seems you guys killed your power circuitry from too much capacitance.

mk-ultra
if you need caps or parts just pm me n i'll see if i got it
if you from montreal just come n take em for free
ket brought U down
i'll bring ya back up

I'm doubtful that the capacitor mod was directly responsible for the problem - The ESR and ESL of typical electrolytic capacitors should be high enough to prevent most problems. It might be possible (although unlikely) that the added capacitance effected the regulator stability and sent it into oscillation. I would suggest cleaning the damaged area and PCI-e edge connector with IPA and blasting the card with compressed air incase there is any loose solder lodged somewhere.

I would advise against blindly modding hardware like this, hoping that adding capacitors will help, unless you understand EXACTLY what you're doing - Adding capacitance to the output of a switching regulator will effect its frequency response, so ideally you should also be adjusting feedback compensation networks to account for the lower bandwidth. If this isn't done, at best you won't see the full benefit of the extra capacitance, and at worst your regulated DC supply will turn into something approximating a sine wave. I would expect the most benefit to come from adding small value (<<1uf) ceramic capacitors close to areas of high current demand, which should be much more effective at removing high frequency noise due to superior high frequency characteristics.

Edit: Could you also check the voltage across the black capacitors in your photo - I'm wondering if they're on the input supply side and you've added too low a voltage capacitor which subsequently failed closed.

its the load side caps...

tantalums haven't been used in 12v input side of voltage regs...

caps close?

btw just an update on this i've replace the missing cap and still no go, it booted one time, ended up with a all green screen after like 10mins

found a good deal on a canadian forum and brought another one