Bad Video cards - Reflow?

[cleandeal]

Re: Bad Video cards - Reflow?

Hi, Reflowing can't be a permanent repair as the real problem isn't solder problem but internal chip fault. By reflowing, you think that's because of solder balls melt that card work again but it's not true, the real reason is that the reflow remake connections of the chip material itself. You should desolder the chip and solder a new one but keep in mind that you need buy one new chip that isn't same manufacturer defective chip as you would have same problem some years after soldering the new manufacturer defect chip. Also as i see last post about cracking sound, this the "pop corn" effect as the chip litterally explose because of bad temperature profile. The most important in desoldering/soldering is the preheat phase and choose the right profile temp and time...

[kropotkin]

Re: Bad Video cards - Reflow?

This one was AMD/ATI. I forgot what temp I got it to. All I know is I started hearing a cracking sound from the chip itself and thats when I stopped having the liquid flux under the chip I believe made a lot of difference

Just don't tell me you use solid one, like us

[stj]

Re: Bad Video cards - Reflow?

most datasheets state the minimum cell storage life of flashchips is 10years.
so after 20 they could well be corrupt.

this is not true for high desity memory btw, if you think an SSD or similar is going to even last 5years in storage your pushing your luck!

[Uranium-235]

Re: Bad Video cards - Reflow?

I've had long term success with that method on a lot of laptops. The only difference is I didn't need a IR temp meter. I used a 350*C automotive heatgun (Titan Tools 22400 in the US, mine is Kaufland branded)

Honorable mentions:

-HP DV6000 (PM965 + GF8400GS)
-Acer Aspire 7520 (Geforce 7000M)
-HP DV6-2125so (AMD 785G + HD4530)
-ASUS X59SL (SIS 671DX + HD3470)

...as well as a lot of desktop GPUs as well.

This one was AMD/ATI. I forgot what temp I got it to. All I know is I started hearing a cracking sound from the chip itself and thats when I stopped

having the liquid flux under the chip I believe made a lot of difference

[kropotkin]

Re: Bad Video cards - Reflow?

I mean corrupted VGA BIOS, which is a separate flash chip located on card. This can happen with card just laying packed in a box. E-caps also may degrade when not working (strongly depends on their quality). But all BGA story may be excluded if card was no running for long period suffering heating-cooling cycles. It is also might be a case that it has factory defects of soldering, or was deformated significantly at some days.

Dan81, "green" solder melts at ~230C, Pb-containing ~183C. Numbers given here like 200,350 or 500C not a temperature measurements, it is just "parrots". I work professionally with BGA every day, this is rather complicated kitchen, espetially last years. With hotgun one can make only local heating, causing temperature expanding of the desk with forming "dish". In case you will rich melting point of solder balls under chip, it will fall to this "dish" with most distant balls. The same time temperature of a crystall will have 100C higher, and that will the end of technique. Handling FCBGA with right temperature control guarantees at least that it will function the way ingeneers planned them to work.

Than, when I`m saying that these Dv6000 boards "boil" easily, that means heating them over 235C is dangerous for desk, but with any ASUS/Pegatron motherboard one can be quiet up to 250-260C. What is "reflowing" without temperature control - it is an occultism practice

[bigbeark]

Re: Bad Video cards - Reflow?

Well here is the answer: The original DOA one is still dead, the other two work. I was using my new BIOS code reader in the first PCI slot and either the motherboard was thinking it was a PCI video card or more likely the default setting on the BIOS is PCI video and the motherboard does not detect that a PCI-E video card is installed even when there is nothing in the PCI slot. Imagine my panic when none of my PCI-E cards appeared to work. Fortunatly I had an old PCI video card, and was able to boot and reset the BIOS setting to PCI-E. After I shut down, removed the PCI video and inserted the PCI-E card and was able to determine that all was well.

So is the BIOS that crazy or is my PCI code reader a lemon? Or am I just having a bad week?

BTW the keyboard flashes 3 lights, then one. So I believe the board is booting OK, the bad card just leaves the screen black, no artifacts. The board is an ASUS A8N-E.

[bigbeark]

Re: Bad Video cards - Reflow?

piernov,
bigbeark, where did you by it and when? ATI chips of those period at my feelings falled much more rare than nVidia, but it can be also corrupted BIOS, and many other things, so some movements towards any measurements needed. For example POST-code on MB when you have a black screent, or any description of behaviour. Look carefully for damaged or missing elements, clean the contacts and so on...

Bought 5+ years ago from a computer recycler, new in box. Bios code is '00'. No damage or missing bits, cleaned the contacts. These died when in storage. Outside chance a fuse inside is blown, but I am testing with a high-end 800W Power Supply, 1000 watt BFG. I will dismantle one and check the fuses.

[Dan81]

Re: Bad Video cards - Reflow?

None of the chipsets I used 350*C on had issues after reflowing at 350*C. 200*c seems like a temporary fix, as I feel the chip doesn't get hot enough. 350*C is the default RoHS lead free solder melting point, and thus makes much better contact.

Upon soldering these early Pb-free Quanta motherboards one should be espetially carefull, because they tempt to absorb humidity well and can "boil" during chip desoldering. Recommende to dry it for 24h at 120C.

I had chipsets boil only if I used 550*C (I had this happen once on two Acer mobos, one was a MXM slot Aspire 5520 mobo and the other was a free Aspire 5738 board that came with a T4300). 350*C did the trick for the 8400GS (it was the PM965 version, not the older i945 variant w/ Go7xxx series chipset) and it was alive again. I even played a few hours of Ford Racing 2 and NFS Underground 2 to stress the GPU a bit. Not even an artifact or anything. Used MX4 on CPU and GPU, so no more big issues with thermals.

[kropotkin]

Re: Bad Video cards - Reflow?

Honorable mentions:

-HP DV6000 (PM965 + GF8400GS)

Upon soldering these early Pb-free Quanta motherboards one should be espetially carefull, because they tempt to absorb humidity well and can "boil" during chip desoldering. Recommende to dry it for 24h at 120C.

The only difference is I didn't need a IR temp meter. I used a 350*C automotive heatgun

"After you saying things like this I just can not be sure that you are not killing people on streets at the night" (c)

[kropotkin]

Re: Bad Video cards - Reflow?

piernov, I wrote about microballs under crystall of FCBGA packages, surely they heve same melting temperature as for chip.
The physics of FCBGA degradation is very simple - a plenty of repetitive temperature microdeformations upon turning power on/off cause mechanical brokage of crystall soldering to chip base. By heating it little over soldering temperature will restore those contacts. Not for ever, of course. Heating 200C for one minute does not sure, that soldering of crystall will be nicely restored, although can be nicely used for diagnostics. Doing it wright way, till soldering pins - that do has sense, that is only wtat I mean. Some of Acer 5552G with "cooked" RS880 do lieved up to 3 years more, but surelly not less as 1 year. What than the reason to byu "new" AMD chips which serve same time long. They are just assembled that way.
Of course, it can be dangerous and may be performed by trained people only on approbed IR-stations, do not recomend to anyone to do it any other way.

bigbeark, where did you by it and when? ATI chips of those period at my feelings falled much more rare than nVidia, but it can be also corrupted BIOS, and many other things, so some movements towards any measurements needed. For example POST-code on MB when you have a black screent, or any description of behaviour. Look carefully for damaged or missing elements, clean the contacts and so on...

[Dan81]

Re: Bad Video cards - Reflow?

I actually did have long term success with a reflow of a laptop northbridge/vga but it was risky. Liquid flux under it, foil around it, and a hot air gun with an ir temp meter.

I've had long term success with that method on a lot of laptops. The only difference is I didn't need a IR temp meter. I used a 350*C automotive heatgun (Titan Tools 22400 in the US, mine is Kaufland branded)

Honorable mentions:

-HP DV6000 (PM965 + GF8400GS)
-Acer Aspire 7520 (Geforce 7000M)
-HP DV6-2125so (AMD 785G + HD4530)
-ASUS X59SL (SIS 671DX + HD3470)

...as well as a lot of desktop GPUs as well.

[piernov]

Re: Bad Video cards - Reflow?

In most situations (and almost always the case for GPU, AMD northbridges, NVidia MCP… starting to happen more and more lately with Intel too) the chip itself fails, not the balls under it, so reflowing it pointless. Giving them a little heat can resurrect them for days/weeks/months, but they'll fail again. Actually trying to reflow them is riskier than just giving heat below reflow temps. You have way more chances of cooking the chip, destroying the balls, bending the PCB… especially without proper equipment. Just give the chip 200°C for a minute and a half and you're good to go, but for a limited amount of time.

Heat and power dissipation can indeed lower the life span of a chip, but it doesn't mean everything. There are known defective series like the NVidia chips from 2006 to 2008. And some have a failure rate way above average like the HD4000-HD7000 from AMD, and especially the HD6000M series. Compared to the same era NVidia chips, those AMD ones have an insanely higher failure rate.

There is no way to actually repair those devices except by replacing the chip, which is often not available and may fail again anyway, and it also requires expensive equipment and experience.

Anyway, don't jump right away on the conclusion that the GPU failed. For older cards it'll often be a bad contact in the slot, so a good cleaning with contact cleaner can help. As it's been said already, electrolytics capacitors can be the cause too. And sometimes you only have one video output dying (shorted ESD protection diode or other random issue like damaged connector) and the other still works.

[kropotkin]

Re: Bad Video cards - Reflow?

Success of "reflowing" depend on many factors, one of wich is TDP of chip. Lower heat emission = higher life time. Quality of crystall cooling has the same relation to it. As a result - in some laptop models with nice cooling systems can reanimate (not heatest) AMD northbridges for up to 2.5-3 years. AND it do has sence only if chip temperature overcomes recommended for soldering, to sure microballs under crystall are melt.
For diagnostic purposes it works, sure. If religion do not forbids, can also be used as cheap repair solution, but than have to be performed right way, on IR-station

[Uranium-235]

Re: Bad Video cards - Reflow?

I actually did have long term success with a reflow of a laptop northbridge/vga but it was risky. Liquid flux under it, foil around it, and a hot air gun with an ir temp meter.

[ChaosLegionnaire]

Re: Bad Video cards - Reflow?

he said one is dead right out of the box!
no overheating there!!

i never was referring to the one that was doa. how can the fan run if it is never used? the failure then could be due to mishandling during shipping so stop trolling my posts.

Worked when put away last year.

umm... thats exactly how a video card fails. works when put away and then when u try to use it, just doesnt work. either artifacts, not detected or no display. i got tons of video cards failed like that.

[stj]

Re: Bad Video cards - Reflow?

Caps - are the most misterious components ever - a carton rolled with foil and some leggs added - and it works

almost,
you forgot that it has to be filled with a precise blend of acid and cat-piss

[kropotkin]

Re: Bad Video cards - Reflow?

Caps - are the most misterious components ever - a carton rolled with foil and some leggs added - and it works
If no professional reflow equipment available, you may try another way - heat soldering iron maximally, take big drop of solder, and move it carefully over a crystall for a minute.
Normal way to go = are measurements, of course

[Topcat]

Re: Bad Video cards - Reflow?

I've never had any real long term success reflowing anything with what would be considered 'standard equipment' (SMT hot air guns, heat guns, or ovens... I've had some short-term success, but it usually came back to haunt me......and I didn't see enough business in it to invest in a good & proper reflow/reballing station.

[bigbeark]

Re: Bad Video cards - Reflow?

it depends on the video card. so for those x1650 pro video cards, i know they tend to overheat because of the crap coolers put on them. due to the age of these video cards, sometimes the fan also tends to fail silently and without any warning, making them fail fast from overheating due to a failed fan.

exactly how have the two cards u have "died"? artifacts? not detected at all? hangs/freeze when loading the driver in windows? or 2d displays fine but running anything 3d hangs/freezes/artifacts the card? please explain how they "failed" precisly and clearly.

Worked when put away last year. Building up some new machines, installed on known good boards this week - no video at all. Fans spin up, that's it. Never overheated, no artifacts, just suddenly they don't work. Just like the one that was DOA. Seem to have gone bad in storage.

[stj]

Re: Bad Video cards - Reflow?

he said one is dead right out of the box!
no overheating there!!