What is the exact composition of Ram BGA?

Gold Refining Forum

Help Support Gold Refining Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

archeonist

Well-known member
Joined
Jun 27, 2012
Messages
286
Can somebody tell me what the exact composition is of BGA from ram? I can't find it in it's exact form on the forum. I ask this question because I want to know what oxides there will be formed after incineration.
I see to many times that people don't take proper precautions, they create clouds of dust in their gardens or they drain waste water away.
As far as I know the dust of these BGA chips consists of the following (in various oxide states):
ironoxide
copperoxide
tinoxide
zinkoxide
leadoxide
(cadmiumoxide??)
Maybe also some metals in their pure state.
Some sillica compunds
Some carbon compounds will be left, depending on the succes of the incineration step.

What do you guys do with the water after bleu bowling? Just sette out the oxides and decant the water??
 
Thank you Tzoax! Very clear data. So that makes that one of the main concerns after incineration is leadoxide.
As most oxides are almost insoluble in water I guess it would be ok to settle out the dust and pour off the bleu bowl water (if used) down the drain?
 
Above data is from pre RoHS. Lead is probably only in solder balls. If you have newer RAMs there should be no lead.
 
I dont mean to skew the thread (but here I go..)

So, I am curious what that says about the 8 pin plastic DIP. I mean, there is just NO way to have accurate data on that for the composition of its contents.

let me elaborate

There are just far too many components that can and do come in an 8 pin chip. Op Amps, 555 timers, constant current drivers, schmitt triggers, buck converters, boost converters (of course peripheral components are needed to complete those circuits), ....this list goes on an on... even resister arrays are in 8pin plastic DIP chips, and they have literally no value internally what so ever.. in fact, most all of those chips I listed have next to no value on their innards. ...maybe if we were talking about nothing but PIC chips or ATTiny chips, then it would be feasible to have an accurate composition of them entirely- but then it would get drawn into the discussion of. - what plant were they made at, by which company, etc... etc...

Im not trying to be overly critical here, (it just seems like it), but I believe FPGA and BGA chips, will be in the same boat too... like the N/S bridge bga chips are wayy different than the graphics BGA chips.
 
archeonist said:
Thank you Tzoax! Very clear data. So that makes that one of the main concerns after incineration is leadoxide.
As most oxides are almost insoluble in water I guess it would be ok to settle out the dust and pour off the bleu bowl water (if used) down the drain?

There is almost 0% chance of running into Lead if you are processing any DDR RAM onwards, which is about 99% of what you come across now.

When I do my wash procedure after incineration I pour the wash water through a big sand filter I made by filling a old rubbish bin half through sand.
 
Topher_osAUrus said:
I dont mean to skew the thread (but here I go..)

So, I am curious what that says about the 8 pin plastic DIP. I mean, there is just NO way to have accurate data on that for the composition of its contents.

let me elaborate

There are just far too many components that can and do come in an 8 pin chip. Op Amps, 555 timers, constant current drivers, schmitt triggers, buck converters, boost converters (of course peripheral components are needed to complete those circuits), ....this list goes on an on... even resister arrays are in 8pin plastic DIP chips, and they have literally no value internally what so ever.. in fact, most all of those chips I listed have next to no value on their innards. ...maybe if we were talking about nothing but PIC chips or ATTiny chips, then it would be feasible to have an accurate composition of them entirely- but then it would get drawn into the discussion of. - what plant were they made at, by which company, etc... etc...

Im not trying to be overly critical here, (it just seems like it), but I believe FPGA and BGA chips, will be in the same boat too... like the N/S bridge bga chips are wayy different than the graphics BGA chips.
Can you be more specific in what you are talking about? 8 pin plastic DIP what is it and what has this to do with my topic?
 
archeonist said:
Can you be more specific in what you are talking about? 8 pin plastic DIP what is it and what has this to do with my topic?

It was simply an example
Its listed right under tzoaks picture of the BGA in his app, suggesting that yield for it is available from the program as well.

The example being. -- Yield isnt a "one size fits all" thing... there are 100's of chip types that come in BGA designation... 10000's of chips that come in PDIP 8pin. If all of them come with high PM value, why do these sell for (literally) pennies from China?

some have value

a lot dont.
-You dont need any precious metals to make some semiconductors, or any at all for resistor arrays and so on and so forth

What it has to do with topic is this- Unless you are searching EACH part numbers datasheet, you wont know the composition of it. Period
 
Topher_osAUrus said:
I dont mean to skew the thread (but here I go..)

So, I am curious what that says about the 8 pin plastic DIP. I mean, there is just NO way to have accurate data on that for the composition of its contents.

let me elaborate

There are just far too many components that can and do come in an 8 pin chip. Op Amps, 555 timers, constant current drivers, schmitt triggers, buck converters, boost converters (of course peripheral components are needed to complete those circuits), ....this list goes on an on... even resister arrays are in 8pin plastic DIP chips, and they have literally no value internally what so ever.. in fact, most all of those chips I listed have next to no value on their innards. ...maybe if we were talking about nothing but PIC chips or ATTiny chips, then it would be feasible to have an accurate composition of them entirely- but then it would get drawn into the discussion of. - what plant were they made at, by which company, etc... etc...

Im not trying to be overly critical here, (it just seems like it), but I believe FPGA and BGA chips, will be in the same boat too... like the N/S bridge bga chips are wayy different than the graphics BGA chips.

It all depends of the way of look. There are many components that comes in PDIP-8 package, and if that information is enough for anyone to believe that it proves that there is no way to have accurate data - problem is solved - there isn't (for them...or better said from their point of view).

From my point of view - i believe in results of gold i managed to recover from specific types of IC chips. And - those numbers are always slightly lower than numbers (data) from manufacturers. Anyone could try this and see if it is true. For example - take some amount of PDIP-8 package IC chips, process it, and compare your result with this i will post on a next picture, and you will see results will be very close. When i compare my results with these - it was always same - a little less than expected, even with PDIP-8 package.

Of course, it all depends what "accurate" means to you in this case. For me - having information that i can "take out" 85-90% of gold (from expected amount of gold) is indeed accurate enough, or better said closest enough.

There are many plants, companies, time that it was made etc... but take some amount of PDIP-8 package IC chips from all mixed times, plants and companies, take out the gold from it and compare it with these data - you will see the point for yourself. The results will vary little more or little less...but never that much - that predicting yields this way would not have sense (for me). I believe in this because i compared my numbers with these and for me - it is close enough i could roughly predict yields.

I am talking only about IC's from PC's - motherboards, PCI cards, graphics, HDD boards etc... I am not sure these data could be used on military, medical or any other equipment...but for PC's - i checked it for myself and that is why it makes a lot of sense (to me).

ddds.jpg
 
Topher_osAUrus said:
Yield isnt a "one size fits all" thing
That is true, but production is very similar. One bonding wire per leg, 1 dye. Wire thickness is more or less the same, unless is some power device, which are very rare in PDIP packages. Optocouplers use same package and have probably higher yield, but majority IC have very similar. Open few and have a look.
 
shmandi said:
Topher_osAUrus said:
Yield isnt a "one size fits all" thing
That is true, but production is very similar. One bonding wire per leg, 1 dye. Wire thickness is more or less the same, unless is some power device, which are very rare in PDIP packages. Optocouplers use same package and have probably higher yield, but majority IC have very similar. Open few and have a look.

I would venture to say that a lot of then use aluminum bond wires instead of gold.

Ive even seen ceramic processors (gold cap ones none the less) that had Al bond wires.
(I will add the photo when i find it)

Whats a good analogy for the point Im trying to convey..
A sandwich is a sandwich, but not all sandwiches are the same?
Maybe that works

Edit to add

And production is similar to a point, basically the aesthetics of it, while the die / interior is 1 of 1000000 options.

Im not trying to be "that guy", but, one of these days a guy is going to get next to nothing on a yield, and come back wondering why this lead him astray.

All done on my soapbox now
 
Topher_osAUrus said:
I would venture to say that a lot of then use aluminum bond wires instead of gold.

Ive even seen ceramic processors (gold cap ones none the less) that had Al bond wires.
(I will add the photo when i find it)
Most of ceramics DIPs i have seen use aluminium bond wires, but I have not seen epoxy (plastic) yet. I don't say that there aren't, but all I opened had gold bonding wires.
Maybe recent types use AL or even Cu wires as most of scrap I deal with is 10+ years old.
 
shmandi said:
Topher_osAUrus said:
I would venture to say that a lot of then use aluminum bond wires instead of gold.

Ive even seen ceramic processors (gold cap ones none the less) that had Al bond wires.
(I will add the photo when i find it)
Most of ceramics DIPs i have seen use aluminium bond wires, but I have not seen epoxy (plastic) yet. I don't say that there aren't, but all I opened had gold bonding wires.
Maybe recent types use AL or even Cu wires as most of scrap I deal with is 10+ years old.

indeed, it is a possibility. That was my only point. When dealing with escrap, those possibilities are pretty close to endless.

My post was not meant to degrade or belittle the hard work that Tzoax has done, in fact I find his selfless giving and sharing of yield info quite admirable.

I guess I just had to pick nits and throw in the bit about "they may all look the same, but in fact they are (almost) all different"
 
This is one example - a couple of years ago i tested PLCC-68 package type IC chips.
This is a link from my thread about it:
http://goldrefiningforum.com/phpBB3/viewtopic.php?f=33&t=22951#p240804

ooo.jpg

Result was 0.49g of gold per kg.

When i compared my results with these (0.47g/kg) - it almost matched.

fdfdfdfdfd.jpg

I was surprised - because i expected less gold than predicted... And then i realized why that happened - because i removed legs before weighting. Data for IC chips from my application is with legs - 4.583g.
And IC chips i tested was 4.371g (without legs). So, again - my own results are slightly below than predicted, and like you see on the pictures there are many different manufacturers from different time, but - finally - results are very, very close.

And again - it is only matter of point of view. i could not believe in something without testing, and i proved to myself that my data is slightly lower than these, always, with any package. I am not asking from anyone to take it for granted, i am inviting you to check for yourself. Make some tests with some specific packages and believe in your results.

shmandi said:
Most of ceramics DIPs i have seen use aluminium bond wires, but I have not seen epoxy (plastic) yet. I don't say that there aren't, but all I opened had gold bonding wires.

Shmandi, i agree with you, i am talking about epoxy (plastic) packages here, not ceramic, and like you - i never saw aluminum bonding wires there - only gold bonding wires.

Topher_osAUrus said:
Whats a good analogy for the point Im trying to convey..
A sandwich is a sandwich, but not all sandwiches are the same?
Maybe that works

Topher, specific sandwich package will contain specific amount of mayonnaise.
 
but, see, I don't like mayonnaise, and some manufacturers of sandwiches dont like to spend money on it, since they see it as frivolous waste of pennies. When you make 10000000000 sandwiches at a time, those pennies add up to pounds
 
theitalianhenchman said:
Hi Tzoak
Can you please tell us (me?) in which internet site you can find the data about the chips?
Thanks

Of course, i will send you links for data source and application (it's all free) in PM. Just keep in mind that there are many manufacturers there so yields could/will vary...but based on my own tests with IC packages - they will not vary much.

Alex
 
Tzoax said:
theitalianhenchman said:
Hi Tzoak
Can you please tell us (me?) in which internet site you can find the data about the chips?
Thanks

Of course, i will send you links for data source and application (it's all free) in PM. Just keep in mind that there are many manufacturers there so yields could/will vary...but based on my own tests with IC packages - they will not vary much.

Alex


I've received the PM and sent one back. Thank you very much!
 
Back
Top