VOGONS


First post, by fool

User metadata
Rank Member
Rank
Member

I have these two non-working cards: Diamond Viper V330 and Hercules 3D Prophet II GTS
Award BIOS gives me 1 long + 2 short beebs. I have checked caps and missing components but looks fine. First I just tried resoldering with hot air blower without positive results. What I got was short circuit under Geforce 2 GPU.

As those are worthless in the current condition, I decided to reball GPU BGA chips. Never done that before, I have only soldered & repaired thousands of BGA chips in previous life. So how hard it could be? Let see... This is just for hobby, I don't expect great success. 😃

I have a chinese hot air blower and a good Weller solder station to work with. Sadly I coudn't take pictures from this section.

As I don't have base heater, just gave a good hot air warm up on the back side, then turned the board around and heated GPU slowly by increasing temperature until GPU was removeable.
GTS had a LOT of non-wetting (at least they looked like) pads on both sides, Riva 128 had only a dozen of them. I hope those open solders are the root cause.

Next I added some flux and slided soldering iron tip around to get all of the pads wet, same thing on GPU pads. After all pads were coated by shiny tin, removed it with solder wick.

Yet I don't have any solder balls and stencil to perform reballing. After board pads were clean, I could measure pitch which is 1.27mm. Some chart (found on Google) advised to use 0.76mm balls on 1.27 pitch, so that's what I ordered. Also 1.27 pitch reball stencil ordered. Total: 8€.

Now I just have to wait orders to arrive... Then I will update the story and result, with pictures of course.

IMG_20201026_181907.jpg
Filename
IMG_20201026_181907.jpg
File size
1.63 MiB
Views
882 views
File license
Public domain
IMG_20201026_181955.jpg
Filename
IMG_20201026_181955.jpg
File size
1.69 MiB
Views
882 views
File license
Public domain
IMG_20201026_181941.jpg
Filename
IMG_20201026_181941.jpg
File size
1.94 MiB
Views
882 views
File license
Public domain
IMG_20201026_181800.jpg
Filename
IMG_20201026_181800.jpg
File size
1.64 MiB
Views
882 views
File license
Public domain
IMG_20201026_181811.jpg
Filename
IMG_20201026_181811.jpg
File size
1.74 MiB
Views
882 views
File license
Public domain

Toshiba T8500 desktop
SAM/CS9233 Wavetable Synthesizer daughterboard
Coming: 40-pin 8MB SIMM kit, CS4232 ISA wavetable sound card

Reply 1 of 13, by Doornkaat

User metadata
Rank l33t
Rank
l33t

Hey dude, thanks for sharing the process! 👍
What heat are you using on the cards and is there an advantage in reballing instead of reflowing (adding only flux and hoping for solder connections to reform)?

Reply 2 of 13, by fool

User metadata
Rank Member
Rank
Member

I'm no expert on this topic, but I understand that non-wetting is caused from oxidation. If there is oxidation on the pad, solder will not stick on if only heated. Purpose of this reballing is to scap off oxidation layers so that solder will stick on the pad with a much higher probability.

I used something like 400+ °C temperature in hot air blower.

Toshiba T8500 desktop
SAM/CS9233 Wavetable Synthesizer daughterboard
Coming: 40-pin 8MB SIMM kit, CS4232 ISA wavetable sound card

Reply 3 of 13, by bloodem

User metadata
Rank Oldbie
Rank
Oldbie

Unfortunately, there's a good chance that the chips themselves are fried, but I guess it's a nice project anyway. 😀 Looking forward to seeing the cards working again... hopefully 😀

1 x PLCC-68 / 2 x PGA132 / 5 x Skt 3 / 9 x Skt 7 / 12 x SS7 / 1 x Skt 8 / 14 x Slot 1 / 5 x Slot A
5 x Skt 370 / 8 x Skt A / 2 x Skt 478 / 2 x Skt 754 / 3 x Skt 939 / 7 x LGA775 / 1 x LGA1155
Current PC: Ryzen 7 5800X3D
Backup PC: Core i7 7700k

Reply 4 of 13, by Doornkaat

User metadata
Rank l33t
Rank
l33t
fool wrote on 2020-10-26, 17:27:

I'm no expert on this topic, but I understand that non-wetting is caused from oxidation. If there is oxidation on the pad, solder will not stick on if only heated. Purpose of this reballing is to scap off oxidation layers so that solder will stick on the pad with a much higher probability.

I used something like 400+ °C temperature in hot air blower.

Ok, thanks for explaining!
I have two cards that I hope I can fix by reflowing so fingers crossed there is no corrosion on their pads!

Reply 5 of 13, by SSTV2

User metadata
Rank Oldbie
Rank
Oldbie

I don't want to be a party pooper, but if you had heated those BGA chips above 250*C for a prolonged period of time (let alone 400*C !), it can be safely assumed that they are both dead 🙁

Read this doc., melting temperature for leaded alloy is 183*C and for lead free it's 222*C. In general, you should avoid going above 235*C when soldering or reflowing any chip.

Reply 7 of 13, by Hanamichi

User metadata
Rank Member
Rank
Member

Yep, I think he meant 400°C for the heatgun/pen temp not the board or chip temp. By the time the air reaches the chip it's quite a bit cooler.

Preheating to say 100°C+ on the board does help then you can do a quick blast at close range for the chip.
My crude attempt fixed a GF4

Re: Fixing geforce cards

Hoping for successful resurrections 👍

Reply 8 of 13, by adalbert

User metadata
Rank Oldbie
Rank
Oldbie

Preheating is important with such repairs. You could try using electric cooking stove (don't put the card directly on top of it, raise it few cm) and a temperature sensor on the board... BGA replacement is kind-of doable with primitive tools when you are dealing with leaded solder (I only reballed memory chip, not GPU), but with lead-free it's very difficult and I failed miserably with that (ended up popcorning the GPU, even though I preheated it).

Repair/electronic stuff videos: https://www.youtube.com/c/adalbertfix
ISA Wi-fi + USB in T3200SXC: https://www.youtube.com/watch?v=WX30t3lYezs
GUI programming for Windows 3.11 (the easy way): https://www.youtube.com/watch?v=d6L272OApVg

Reply 11 of 13, by quicknick

User metadata
Rank Oldbie
Rank
Oldbie

Very nice and clean work so far, nearby SMDs look untouched, hopefully you'll have a good result! And after seeing this I feel encouraged to try my luck with some similar operations 😉

Reply 12 of 13, by fool

User metadata
Rank Member
Rank
Member

In general, you should avoid going above 235*C when soldering or reflowing any chip.

With that temperature you won't get any BGA's removed, especially aged lead-free things. PCB this size will absorb a lot of heat energy. Actually I can't remember causing IC-damage with excessive temperature, more likely reasons are moisture in PCB->popcorning, and too fast temperature rise. I was a bit amazed there was no popcorning taking into account the time and temperature I had to use. Proper reflow would require right size nozzle and controlled air flow.

Unfortunately, there's a good chance that the chips themselves are fried, but I guess it's a nice project anyway.

Yes, that is very likely as I remelted these once before removing. This is more like a practise event that I have wanted to attempt for a long time.

You could try using electric cooking stove (don't put the card directly on top of it, raise it few cm) and a temperature sensor on the board...

Actually a very good suggestion. I have planned to build one from high power wirewound or PTC resistors.

how do you know its not the memory the issue?

- I dont, I just hope it's not. These memory chips are not BGA chips, so solder faults are easy to find - also not so probable. Finding electrical fault + replacing memory chips does sound very time consuming and frustrating process what I'n not going to perform.

Toshiba T8500 desktop
SAM/CS9233 Wavetable Synthesizer daughterboard
Coming: 40-pin 8MB SIMM kit, CS4232 ISA wavetable sound card

Reply 13 of 13, by andrea

User metadata
Rank Member
Rank
Member
fool wrote on 2020-10-27, 16:33:

how do you know its not the memory the issue?

- I dont, I just hope it's not. These memory chips are not BGA chips, so solder faults are easy to find - also not so probable. Finding electrical fault + replacing memory chips does sound very time consuming and frustrating process what I'n not going to perform.

In my experience a card with bad memory will still "work". Sure, the screen will be full of rubbish and will crash on driver load, but the computer will at least attempt to boot and not beep 2long 3short