Express method for determining the amount of gold in microcircuits

Gold Refining Forum

Help Support Gold Refining Forum:

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

Francois

Member
Joined
Sep 7, 2021
Messages
22
Hi)
Dear forum members, what methods do you use to quickly determine the amount of gold in unknown chips?
I found two different methods on YouTube videos.
First method is with a laser -
https://www.youtube.com/watch?v=Zx3w5ylmssQ
Second method with chemicals -
https://www.youtube.com/watch?v=mT1FStxAVz4
Both methods allow you to estimate the approximate number of gold hairs in a chip.

Which method do you use if you need to infer the content of gold hairs in a single chip?

Thanks! :)
 
Francois said:
Both methods allow you to estimate the approximate number of gold hairs in a chip.

Which method do you use if you need to infer the content of gold hairs in a single chip?

if you want to know "how many" gold bonding wires are in a chip all you have to do is count how many legs are on the outside of the chip (the legs that attach the chip to the board) for every leg on the outside of the chip there is gold bond wire inside the chip

that's what the gold bond wires do - they attach the BIG legs on the outside of the chip to the silicon die on the inside of the chip - which is what allows the chip/die to communicate with the rest of the computer/internet/world

Kurt
 
kurtak said:
Francois said:
Both methods allow you to estimate the approximate number of gold hairs in a chip.

Which method do you use if you need to infer the content of gold hairs in a single chip?

if you want to know "how many" gold bonding wires are in a chip all you have to do is count how many legs are on the outside of the chip (the legs that attach the chip to the board) for every leg on the outside of the chip there is gold bond wire inside the chip

that's what the gold bond wires do - they attach the BIG legs on the outside of the chip to the silicon die on the inside of the chip - which is what allows the chip/die to communicate with the rest of the computer/internet/world

Kurt
Is there not a great variation in length and diameter of the wires used?
Even a slight variation will lead to discrepancy, the larger the number, the greater the difference.
How many wires in a twenty or thirty kilo lot?
 
kurtak said:
Francois said:
Both methods allow you to estimate the approximate number of gold hairs in a chip.

Which method do you use if you need to infer the content of gold hairs in a single chip?

if you want to know "how many" gold bonding wires are in a chip all you have to do is count how many legs are on the outside of the chip (the legs that attach the chip to the board) for every leg on the outside of the chip there is gold bond wire inside the chip

that's what the gold bond wires do - they attach the BIG legs on the outside of the chip to the silicon die on the inside of the chip - which is what allows the chip/die to communicate with the rest of the computer/internet/world

Kurt

Before you decide to buy a large number of printed circuit boards, you need to find out the precious metal content of the chips.
If you dissolve one or two chips in acid, the accuracy will be poor.
You can easily compare different types of chips to a reference chip. Just look at the number and length of hairs after removing the top layer. This way you can already get an idea of the situation.
 
11657-01.jpg

and
HTB10HjGE7SWBuNjSszdq6zeSpXa1.jpg


How do I know how many hairs are in which chip? :wink:

If you carefully remove the top layer of the chip, you can already judge the number and length of hairs. And whether there are any hairs there at all :)
 
Each pin will have one bonding wire.. a quick count... I see 52 legs on one side, so times 4 = 208 pins..
if I read the datasheet of this Spartan Chip correctly, (https://www.xilinx.com/support/documentation/data_sheets/ds312.pdf) there are bonding wires.. not sure if they are golden.
Wires encased in plastic or epoxy can not be made visible to count or guess length by taking the top off. That would be too easy. And there would be no need for incineration of the chip if they were visible.
 
Martijn said:
Each pin will have one bonding wire.. a quick count... I see 52 legs on one side, so times 4 = 208 pins..
if I read the datasheet of this Spartan Chip correctly, (https://www.xilinx.com/support/documentation/data_sheets/ds312.pdf) there are bonding wires.. not sure if they are golden.
Wires encased in plastic or epoxy can not be made visible to count or guess length by taking the top off. That would be too easy. And there would be no need for incineration of the chip if they were visible.

https://www.youtube.com/watch?v=Zx3w5ylmssQ

This is Spartan. His mini version. All the gold hairs are perfectly visible after removing the top layer with the laser)

And if the hairs are not gold, but, for example, aluminum hairs, you can see that right away, too.
And you don't have to use chemistry).
 
That's a nice video. If you have acces to a laser like that, great!
It sure is a fast way to determine if there's gold in a chip.
You have one of those?
 
Francois said:
11657-01.jpg

and
HTB10HjGE7SWBuNjSszdq6zeSpXa1.jpg


How do I know how many hairs are in which chip? 😉

First of all - when we discuss it here on the forum they are not called hairs - they are called bond(or bonding) wires because they "bond" the silicon die/chip in the chip to the legs (or pins) on the edge of the outside of the chip that attach to the circuit board

If you want to know "how many" bond wires are in a chip - as both I & Martijin said yesterday - all you have to do is count the legs/pins on the outside of the chip - for each leg/pin there is also a bond wire - so if there is 208 legs/pins on the outside of a (those) chips - there are 208 bond wires in the chip

In other words - the two methods of exposing the bond wires you showed tell you "nothing about "how much" gold is in the chip - it only shows you the bond wires & if you are want to know how many there are there is no reason to spend the time/money on ether of those methods when all you have to do is count the legs/pins on the outside of the chip --- those two methods are a waste of time/money in my opinion

The reason those methods don't tell you "how much" gold is in a chip is because of what justincase said yesterday

Is there not a great variation in length and diameter of the wires used?
Even a slight variation will lead to discrepancy, the larger the number, the greater the difference.
How many wires in a twenty or thirty kilo lot?

so knowing "how many" wires - does not tell you "how much" gold

If you want to know "how much" gold in a "particular" type of chip - then you ether have to do an "actual" assay on that particular type chip - or -----------

read this thread

:arrow: My results of specific types of IC chips, flatpacks and BGA

And if the hairs are not gold, but, for example, aluminum hairs, you can see that right away, too.

I used to process around 500 - 600 pounds of epoxy IC chips per year (in 50 - 60 pound batches) & I can tell you that the VAST MAJORITY have GOLD bond wires in them --- which the link I just provided will also tell you

so why waste time & money on ether of the methods you showed to process ONE chip at a time --- when you can use that time to run them in a LARGE batch & actually make some money :?:

Kurt
 
kurtak said:
so why waste time & money on ether of the methods you showed to process ONE chip at a time --- when you can use that time to run them in a LARGE batch & actually make some money :?:

Kurt

Let's say you have a sample of one chip. And you need to know how many precious metals the chip contains and then decide to buy a large number of chips.
How will you know if it's worth spending money on a big batch of chips if no one else will provide you with chip samples? :wink:
 
Francois said:
Let's say you have a sample of one chip. And you need to know how many precious metals the chip contains and then decide to buy a large number of chips.
How will you know if it's worth spending money on a big batch of chips if no one else will provide you with chip samples? :wink:

Per the underlined - First I "assume" you mean how much precious metal (not how many)

So - again - the two methods you showed - does NOT tell you how much gold is in a chip (or batch of chips)

those methods ONLY "exposes" the bond wires so you can see them - you still have NO IDEA how much that gold you see weighs - so you still don't know HOW MUCH gold it is you are seeing

In order to determine how much gold is in a particular type of chip you have 3 options

1) have an actual assay done on that type chip

2) look at a data base (like the link I provided) where some one has already done test batches on specific types of chips & recorded/provided the results of testing that type chip(s)

3) collect chips - sort them into types - run your own test batches - record you recovery results just like Tzoax did in the link I provided

In other words - until you "actually recover" the gold from the chips you have no idea HOW MUCH is IN the chip(s) --- you ONLY see the gold - with the two methods you posted

When I did (& bought) large batches of chips - I bought them based on on options 2 & 3 above

Keep in mind - how much gold YOU actually recover (from this kind of material) is going to depend on how good of a job YOU DO in YOUR process of recovery --- this kind of material is not an easy process & involves "several" steps & if you are not careful in your work at EACH of those steps you can & will loose gold at EACH of those steps --- the larger the batch - the easier it is to loose gold - if you are not careful at each & every step in the process

Kurt
 
kurtak said:
When I did (& bought) large batches of chips - I bought them based on on options 2 & 3 above

Keep in mind - how much gold YOU actually recover (from this kind of material) is going to depend on how good of a job YOU DO in YOUR process of recovery --- this kind of material is not an easy process & involves "several" steps & if you are not careful in your work at EACH of those steps you can & will loose gold at EACH of those steps --- the larger the batch - the easier it is to loose gold - if you are not careful at each & every step in the process

Kurt


I completely agree with you, this method does not give absolute accuracy) But absolute accuracy is not always needed, sometimes a comparative test is enough.

If you take a motherboard with one socket and remove the top layer on the south and north bridges, you can immediately see that the north bridge is useless and the south bridge is very tasty :) And on the other socket the north bridge and the south bridge are both very tasty :)

Immediately it becomes clear which chips are better to deal with :wink:

No chemistry needed :wink:

This is just an example, it applies to all chips.
 
Francois said:
If you take a motherboard with one socket and remove the top layer on the south and north bridges, you can immediately see that the north bridge is useless and the south bridge is very tasty :) And on the other socket the north bridge and the south bridge are both very tasty :)

Now you are arguing with me just to argue with me --- I don't have time for that :!: :!: :!:

If you want to see if there is gold in a single chip (or not) take that single chip - burn it with a torch & turn it to carbon/ash - crush it - put the crushed carbon/ash in a gold pan & pan off the carbon/ash until you see the bonding wires --- that takes about 4 - 7 minutes

Edited to add; - Now you have the EXPRESS method :!:

other wise you can go buy a VERY expensive laser & use that method

or you can use the other method which will take something like an hour or more (edited to change time after looking at the video again)

That is the LAST I have to say about this because I am not going to waste more time arguing with you :!: :!: :!:

I have better things to do with my time :!: :!: :!:

Kurt
 
kurtak said:
Francois said:
If you take a motherboard with one socket and remove the top layer on the south and north bridges, you can immediately see that the north bridge is useless and the south bridge is very tasty :) And on the other socket the north bridge and the south bridge are both very tasty :)

Now you are arguing with me just to argue with me --- I don't have time for that :!: :!: :!:

If you want to see if there is gold in a single chip (or not) take that single chip - burn it with a torch & turn it to carbon/ash - crush it - put the crushed carbon/ash in a gold pan & pan off the carbon/ash until you see the bonding wires --- that takes about 4 - 7 minutes

Edited to add; - Now you have the EXPRESS method :!:

other wise you can go buy a VERY expensive laser & use that method

or you can use the other method which will take something like an hour or more (edited to change time after looking at the video again)

That is the LAST I have to say about this because I am not going to waste more time arguing with you :!: :!: :!:

I have better things to do with my time :!: :!: :!:

Kurt


I'm not arguing with you and didn't even think about it :wink:
I agree with you, your method works very well too :!:
Thank you very much for the advice!

The laser doesn't cost a lot of money, you can use a regular cheap Chinese laser on a CNC machine or 3D printer.
The best part is that you don't even have to unsolder the chip from the board for the diagnostic process. And after that, the chip will continue to work :)
In my opinion this method of approximate estimation of gold content has the right to life too :wink:

Again, I appreciate your depth of knowledge and experience and respect your opinion :!: :!: :!:

Regards, François!
 
Back
Top