x-box cpu

[pimpneightez]

Tried doing a search but couldn't find anything. Anything in the 2 x-box CPU's that are soldered on the board? Instead of the little gold triangle they have a silver triangle like the north/south bridges.

 

[necromancer]

those are GPU
graphics processor units

the cpu is in the upper left corner of this pic
http://techon.nikkeibp.co.jp/english/NEWS_EN/20051129/111162/xbox360main_omote.jpg

 

[joubjonn]

if it's got a black top (two pieces) then it's almost for sure got some gold wires inside the black piece

if it's one solid green fiber CPU then it's not worth processing, in my opinion. I save them but I don't know why. when you burn them they have a few layers of fiber glass webbing and copper foils. might be some action on those copper foils but I'm not sure it's worth the hassle for a very small amount I imagine you would need many pounds to make a gram.

 

[pimpneightez]

Gotcha. GPU it is. What's the best way to get the CPU off? In the past I tried a heat gun but half the gold foil underneath always peels off and is stuck to the board.

 

[g_axelsson]

those are GPU
graphics processor units

the cpu is in the upper left corner of this pic
http://techon.nikkeibp.co.jp/english/NEWS_EN/20051129/111162/xbox360main_omote.jpg

The CPU in the upper left corner is a epoxy covered BGA capsule. Inside the epoxy, the chip is connected to the fiber board with gold bonding wires. This chip also have a heat spreader on top, it is easy to remove before incineration and further processing. Very good material to work with.
The two GPU chips in the center are flip chip BGA capsules. They have solder bumps between the fiber base and the chip. The only hope for any precious metals are silver in the tin solder (a few percents) and possible palladium in the ceramic capacitors on top of the fiber board. My advise is to leave them on the board.

Göran

 

[solar_plasma]

Try to find "testing black chips and SMD packages on their raw gold content.PDF"

FCBGA's run at 0,279% in this test.

Edit... I thought of a
PlayStation, sorry. This one though is also described in the test.

 

[g_axelsson]

Found this on the forum, 5 pounds of FC BGA no gold.
http://goldrefiningforum.com/~goldrefi/phpBB3/viewtopic.php?f=34&t=18106#p182618

And my old post on the subject.
http://goldrefiningforum.com/~goldrefi/phpBB3/viewtopic.php?f=52&t=11827&start=280#p187975

Try to find "testing black chips and SMD packages on their raw gold content.PDF"

FCBGA's run at 0,279% in this test.

Edit... I thought of a
PlayStation, sorry. This one though is also described in the test.

Is that 0.279% number for the FCBGA or the whole playstation? What type of chips are used in the playstation? Do you have a link?

I couldn't find that pdf, but I found the "Flip Chip Ball Grid Array Package Reference Guide" on Texas Instruments. in it they talk about different surface treatments and in the case ENIG (Electroless Nickel Immersion Gold) is used on a FC BGA this is what they say about it.

The ENIG finish consists of plating electroless nickel over the copper pad,
followed by a thin layer of immersion gold. The allowable stresses and
temperature excursions the PCB is subjected to throughout its lifetime,
determine the thickness of the electroless nickel layer. This thickness is
typically 5 μm nickel and about 0.05 μm for gold, to prevent brittle solder joints.

On page 8 there is a nice cross section of a flip chip BGA for anyone interested.

ENIG isn't used on every FC BGA board but if it is used you usually see it as any surface without solder on top of it is gold plated. But remember, any non-visual gold is dissolved in the solder balls and in the beginning the exposed surfaces of the fiber board were only covered by about 0.05 μm gold. That was too little to register on Geo's test.
On the x-box we can spot a couple of small dots, "landing marks" used for orientation, that isn't soldered. On the left BGA we see two small squares and a triangle and on the right BGA we see three small dots. None are yellow, they are all silver colored so my guess is that these BGA:s were protected by silver during assembly. No gold there.
http://techon.nikkeibp.co.jp/english/NEWS_EN/20051129/111162/xbox360main_omote.jpg

Göran

 

[solar_plasma]

The PDF is obviously gone. Nowhere to find anymore. It was written by Roy Eugster, Switzerland.

As far as I understand, the test was made on the whole FCBGA inclusively the heatsink. The test was made on several kinds of IC's, - allmost all kinds I find in PC-boards: DIL8-20, FCBGA(A),(B),(C),(D),(E),PQFP100, QFP20-36, QFP68-200,SOIC20/SO8-20/SOP14-20,SOJ-W16-28,SOJ20-28,SOJ44,TQFP <=64, TQFP68-200,TSOP44-54

The GPU above seems to be a FCBGA(B), doesn't it? 30g FCBGA(B) gave a yield of 0,1328g in that test, 0,443%.

It would be fine, if anyone could find the link. I can't.

In a PS2 there are 2 big pieces of FCBGA(E).

edit: I do not talk about that flip chip garbage. I do not know if the names for the IC's in the test are correct.

 

[solar_plasma]

Now you say it, I think the names for the BGA's are incorrect.

 

[solar_plasma]

Found something interesting on the xbox board when I cracked this little blackbox with four goldplated legs open: a goldplated metal ball. Holy... at first sight I believed it to be mercury :shock: and put it very cautious into a closed vessel for further examination. ThenI saw it was hard and golden :lol:

The other pictures are from a weird flipchip with two dies, which has goldplating beneath the dies. I do not know what it is from. Probably not much gold, but first time I found gold in those nasty flipchips.

 

[solar_plasma]

solar_plasma wrote:

Try to find "testing black chips and SMD packages on their raw gold content.PDF"

...it is written by an author from Switzerland, maybe it is from freechemist?

 

[Richard NL]

solar_plasma wrote:

Try to find "testing black chips and SMD packages on their raw gold content.PDF"

...it is written by an author from Switzerland, maybe it is from freechemist?

I don't know if it's from Freechemist.
But did you mean this one: http://tinyurl.com/p8b2tn4

Best Regards.

 

[solar_plasma]

Yes! Thank you!

 

[macfixer01]

solar_plasma wrote:

Try to find "testing black chips and SMD packages on their raw gold content.PDF"

...it is written by an author from Switzerland, maybe it is from freechemist?

I don't know if it's from Freechemist.
But did you mean this one: http://tinyurl.com/p8b2tn4

Best Regards.

Well that was a waste of frickin' time! After they suckered me in to signing in just so I could download this document, only THEN did it tell me the owner marked it private so it can't be downloaded anyway!

 

[solar_plasma]

At least the whole document is readable online. I just wonder how somebody who made such a good work came to call high-yield BGAs for FC(flip-chip)BGAs.

 

[eastky]

I have that PDF file saved. Tried to upload to the forum but couldn't. I will email it to somebody if they want to post it on the forum.

 

[Richard NL]

Well that was a waste of frickin' time! After they suckered me in to signing in just so I could download this document, only THEN did it tell me the owner marked it private so it can't be downloaded anyway!

I'm sorry for your time,and your trouble.
I have set it to public.
The last documents i uploaded were private and it remembers the last settings.


Best regards,

 

[g_axelsson]

At least the whole document is readable online. I just wonder how somebody who made such a good work came to call high-yield BGAs for FC(flip-chip)BGAs.

You are correct, FCBGA contains almost none gold since there are no bond wires in them.
More about FCBGA http://www.amkor.com/go/packaging/all-packages/superfc/superfc-super-flip-chip

The correct names should be :
12 : PBGA (Plastic Ball Grid Array)
13 : PBGA (Plastic Ball Grid Array) with embedded Cu heat spreader
14 : PBGA (Plastic Ball Grid Array)
15 : PBGA (Plastic Ball Grid Array)
16 : TBGA (Tape Ball Grid Array)

More about PBGA and TBGA http://extra.ivf.se/ngl/e-bga/ChapterE1.htm

One interesting result in the pdf though (thanks to the pictures) is the number 16. I got a bunch of them and I haven't refined any of them. Nice to see such high yield from them, especially when most of the mass is made up of the copper heat spreader.

But a bit of advice is important here. The numbers are based on only a few samples and the statistical errors are huge, just compare number 12 to 15. Most of the gold sits in the bond wires and the number, length and diameter vary among different chips and even different dates as manufacturing processes are improved.

Göran

 

[solar_plasma]

But a bit of advice is important here. The numbers are based on only a few samples and the statistical errors are huge, just compare number 12 to 15. Most of the gold sits in the bond wires and the number, length and diameter vary among different chips and even different dates as manufacturing processes are improved.

Göran

Absolutely. Though, one of the most interesting thing the results are telling in my opinion is, that there is a significant relation between yield and number of legs per total weight. I believe that makes it possible to make a qualified estimation if it is very low (far under 0,1g/kg, medium 0,5-1g/kg or high grade 1,5-4g/kg.

 

[solar_plasma]

Does anybody know the weight of the TBGA from the xbox? Unfortunately I already disassembled those I had.

edit: Göran, reading your document, I think it is rather a SPGA than a TPGA, since the TPGA seems not to be gold bonded and I clearly saw a lot of gold wires within the xbox IC. That would be another important note, if not all ICs with this kind of copper heat spreader would contain gold at all.

further, I can't believe that the yield is calculated inclusively the heat spreader. If the IC weighs 20g with heat spreader, this would mean, that one of those ICs yields more than half a gram. Even if it weighs 10g this would be equal to the best ceramic CPUs, calculated per piece.