RAM chips

[Anonymous]

Ok, some of the RAM chips on the first picture are BGA chips. There does seem to a be a little bit of confusion on the thread so far concerning these, as a lot of people do assume that a BGA is something like a northbridge chip. You can have BGA chips that are on RAM dimms, they are just a different construct to those that are the "high yield" BGA.

For a RAM chips they are similar chips to the others on RAM just connected in a different way.

Dave (Frugal) thanks for pointing out to me the lack of detail in my previous post - I didn't explain my comment nearly clear enough.

 

[g_axelsson]

I found that brown bottomed chips have much more gold in them then green chips.

Are you talking about the BGA memory chips? (BGA = Ball Grid Array)

Do you have run some tests on yields? How does you know that there is more gold in the chips with brown bottom rather han the green? If you only goes by visual clues I would say that you are making a big mistake, most of the gold is in the bond wires and you need to know what to look for if you only have a hand lens, if you have a microscope the wires are easily spotted.
The gold that you can see with your eyes are only thinly plated and only a minor part of the gold in an IC.

If you have any yield numbers, please share it with us. I'm specially interested in the BGA memory chips. What I have seen in my microscope is that the chip is mounted upside down and the connections are along the center line of the chip. The solder balls are mounted on two small fiber boards on each side of the contact line and then bonded. The bond wires are covered in epoxy.

Göran

 

[solar_plasma]

.............. You can have BGA chips that are on RAM dimms, they are just a different construct to those that are the "high yield" BGA. For a RAM chips they are similar chips to the others on RAM just connected in a different way............

I think they have less epoxy/silica material and no kovar legs, but the same technical values, which makes me expect more g/kg yield.

Those bga's in this link beneath are not the same, since they have only 64 contact points, but the construction might be pretty the same:

http://goldrefiningforum.com/~goldrefi/phpBB3/viewtopic.php?f=34&t=10973

So, as far as nobody proves the opposite, I count them as high yield (>1g/kg). If you have driven cars with 200 horses and you see another with 200 horses, you would expect it to be as fast as the others, if it seems to have the same shape as the others. Of course you could be wrong, maybe it does not drive at all, because it's made of cheap parts.

 

[Geo]

Solar, that has been awhile back. I may be a bit off (I am getting old) but from the bottom, they look like VGA chips (north/south bridge). It may not be, I can't see the tops. If I had to give an opinion (I don't but I will anyway), the unusually high gold reported was because it is a VGA chip and not RAM chips.

 

[solar_plasma]

You are right and I could be completely wrong, but the functions are lying within the PM-free die. What counts, is number, thickness and length of the bond wires per weight of garbage.

 

[g_axelsson]

I agree with Solar, it is the type of package that decides if there is a lot of gold or not, not the function of the device. We should concentrate on the package, not the function.

Some earlier postings I've done on the same topic.
http://goldrefiningforum.com/~goldrefi/phpBB3/viewtopic.php?f=34&t=20267#p207312
http://goldrefiningforum.com/~goldrefi/phpBB3/viewtopic.php?f=52&t=11827&start=280#p187975

Göran

 

[Geo]

Actually, comparing VGA chips to ram chips is like comparing a Geo Metro to a Cadillac. Ram chips have iron that must be dealt with and VGA chips (black epoxy) contains only precious metal.

 

[g_axelsson]

Geo, have you looked at the pictures of the first post? The first picture shows BGA capsules, the second TSOP and the third one SOJ. TSOP and SOJ uses lead frame technology while the BGA uses CSP (Chip Scale Package) and lead bond or wire bond.
Looking at page 21 and 27 in http://www.invensas.com/Company/Documents/Invensas_ASQED2011_MemoryPackagingChallenges.pdf shows the different types of connections.

One thing that Romix actually got right was that there seems to be two different types of chips, one brown and one green. Check page 21 and you see two different types of substrate, brown polyimide film used for lead-bonding and green covered polyimide film used for wire bonding. The lead-bonding is made by gold plated copper ribbon so it takes a lot less gold per package.
So the brown bottom BGA memory chips actually contains LESS gold than the green bottomed chips.

Göran

 

[Claudie]

Very informative PDF Göran.
Thank you for sharing.

 

[solar_plasma]

As a side note: Since the die in SOJs, TSOPs and many other packages is placed in the center, I always wondered, if it would make sense to cut off the outer areas or to punch out the center with a pressing. I tried with a pressing that had a round bit with a spike in the middle, but unfortunately it smashed the whole of the chip, not much better than a hammer. I think the spike is the problem.

But basically cutting off the 2/3, that are garbage, could be a fine thing for those who only process small amounts. It would triple the yield per kg of prepared stuff.

 

[rrantique]

Here is a picture of the gold bonding wire still attached to the end of a leg. This will give you some idea how small they are. I'm still not sure about how to recovery these little wires or the best way to crush the chips once burnt into ash. After screening off and panning I still find small parts with the gold wire in little parts of the chips and attached to the legs when viewed under a microscope. This make me think that a large amount of gold is going to get wasted.

 

[richard2013]

Hello All!

For the kovar pins or legs with still gold bonding wire still attach how do we best to remove them?

 

[solar_plasma]

some useful answers:
http://goldrefiningforum.com/~goldrefi/phpBB3/viewtopic.php?f=37&t=13418

Once I did a pound of legs in old Ap, I can not remember to have gotten any gold from this, maybe only enough to darken a filter. The solution tested negative, - still acidic and copper laden it went to my stockpot.

 

[kurtak]

Hello All!

For the kovar pins or legs with still gold bonding wire still attach how do we best to remove them?

Here is what I do

(1) remove magnetic legs (kovar) from chip ash with magnet (this also removes any magnetic heat spreaders which are silver plated)

(2) put in 4 liter beaker

(3) cover with HCL & add some D-water

(4) put beaker on hot plate in fume hood with watch glass on top (watch glass used to reduce acid/water lost to evaporation)

(5) bring acid/water up to simmer type boil

(6) let simmer till base metals are dissolved (about 24 hours depending on batch size)(add D-water as needed due to evaporation)

(7) filter/wash remaining solids

(8) dry & incinerate (incinerate to rid trace acid & oxidize trace BMs)

(9) add to other low grade stuff like incinerated filters, incinerated ink cartridge mylar, etc.

(10) smelt with silver as collector metal

(11) re-melt with silver cement when making anode bars

(12) run in silver cell

Any gold that was dragged out of the ash when removing magnetic material from the ash (which includes heat spreaders &/or legs) is then recovered from the anode slims --- this also recovers the silver plating from the heat spreaders - so no values are lost

Kurt

Edited - to correct messed up step numbering

 

[shaftsinkerawc]

Is step 3 proprietary?

 

[Barren Realms 007]

Is step 3 proprietary?

It's a secret. If he tells you he will have to send a hit team after you. :twisted:

 

[shaftsinkerawc]

Thanks, Tell him to send my reprief as it's getting colder here by the day.

 

[kurtak]

opps my bad - step 4 is actually step 3 step 5 is actually step 4 & so on - will go edit it now

Kurt

 

[richard2013]

Thank you Kurt

 

[Romix]

What about smallest chip on RAM boards.
Any gold inside?