Mixed e-scrap recovery yields

Gold Refining Forum

Help Support Gold Refining Forum:

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

Wingedcloud

Well-known member
Joined
Nov 6, 2014
Messages
96
Hello fellow refiners,

Been gathering some amount of e-waste for refining, in order to process each fraction individually, so I could make some notes on recovery yields, for future work, and I decided I should post them here on the forum. Give a little information back, for the huge amount the forum has given me :)
These values are probably well known by now, as they are from usually recovered materials, but I will post them anyway, if not more, to receive some inputs as to how good (or bad) my recovery work is.

Sem título.jpg

So I tried to make it as much condensed as possible, while providing with the most important information, regarding ratios, mostly.

The processes used for recovery were the traditional ones:

- AR for the ceramic CPU;
- Heatgun to remove pins, HCl bath, water rinse, HNO3 bath and AR for the fiber CPU pins;
- AP for fingers, and ball mill, followed by blue bowl for IC. AR to dissolve gold from both sources.

From the table, I could take these conclusions:
- On the CPU's, the gold/ratio turned out to be a little below the ratio provided by some tables and files found on the internet, on CPU gold yields. For example, some of them claim that the INTEL Pentium's shown here have around 0,12g Au/unit, which does not seem to be the case, unless I made a mistake in the recovery that caused my to lose around 50% of the gold, which I find to be very hard to happen;

- The green fiber result (just noticed it says "fibre" in the picture :oops: ) is also a bit under the value provided by said CPU yield list, but since most of them where the smaller type, I'm not too "schocked" with the yield. These are known for very low yield;

- As for the RAM, I was a little bit disappointed with the yield. After the wash, the gold in IC's is fairly visible and, on this case, it wasn't much, to be fair. I even reprocessed the waste to make sure I did not lose most of it, but there really was all there. IC's from RAM are talked about being rich in gold content, but it does not seem to be the case, at least in this recovery, with the RAM I had.

So, if anyone is willing to take a look at these results and share some ideas and thoughts about it, would be appreciated.
Also, I hope that this can help someone :)

Thanks in advance!

Winged.
 
Whilst I like the info, I have to be frank and tell you that your ceramic Pentium yield appears to be much lower that I would expect. The AMD ceramic is also a little light on yield but it's certainly in the right ballpark 8) 8)

Jon
 
RAM yield is also too low. E-scrap buyers prices are usually around 25 eur/kg, so gold content must be way higher.
 
Dr.xyz said:
RAM yield is also too low. E-scrap buyers prices are usually around 25 eur/kg, so gold content must be way higher.

It would be nice to know the weight of the fingers and for the chips from the ram. That way we could be more definitive as to what the results should have been on the ram.
 
denim said:
It would be nice to know the weight of the fingers and for the chips from the ram. That way we could be more definitive as to what the results should have been on the ram.
Hey denim,

You have fingers and IC mass on the picture. 520g for finges and 2,583kg for IC.

Dr.xyz said:
RAM yield is also too low. E-scrap buyers prices are usually around 25 eur/kg, so gold content must be way higher.
Well, I am sure I did not lose any gold on the fingers, so I must have lost alot on the IC, considering RAM are sold at that price, and I know they are.
Just to break even, there would have to be 5g. So, in reality, there should be 6 or more grams to be worth the recovery work.
Maybe it was just the RAMs I got that had poor IC, if that's is even a possibility.

Winged
 
anachronism said:
Whilst I like the info, I have to be frank and tell you that your ceramic Pentium yield appears to be much lower that I would expect. The AMD ceramic is also a little light on yield but it's certainly in the right ballpark 8) 8)

Jon

Hey anachronism, thanks for the input.
I saw a video on Youtube from Owltech where he recovers gold from AMD Duron CPU, similar to the ones I worked with, and he got the exact same ratio that I did (and was disappointed with it too :lol: )
Also, since the ceramic Pentium are usually bought by gold refiners within the same category as the AMD ones (and thus at the same price), it would be okay to conclude that they should yield similar amounts of gold, even though Pentium ones have that additional amount of gold at their core (and, in fact, the total amount recovered from Pentiums was greater than the AMD). On the other side, they weigh much more, so the gold/kg ratio meets the AMD one. That's why they are paid the same. Or at least that's what I think :)

Winged.
 
Wingedcloud said:
denim said:
It would be nice to know the weight of the fingers and for the chips from the ram. That way we could be more definitive as to what the results should have been on the ram.
Hey denim,

You have fingers and IC mass on the picture. 520g for finges and 2,583kg for IC.

Dr.xyz said:
RAM yield is also too low. E-scrap buyers prices are usually around 25 eur/kg, so gold content must be way higher.
Well, I am sure I did not lose any gold on the fingers, so I must have lost alot on the IC, considering RAM are sold at that price, and I know they are.
Just to break even, there would have to be 5g. So, in reality, there should be 6 or more grams to be worth the recovery work.
Maybe it was just the RAMs I got that had poor IC, if that's is even a possibility.

Winged

Thanks for the correction Winged. So by those numbers you should have gotten maybe 2.5g from the fingers, and 4.5g from the chips? Maybe total about 7g from the ram. Going on what my yields have been in the past.

Dennis
 
Wingedcloud said:
Also, since the ceramic Pentium are usually bought by gold refiners within the same category as the AMD ones (and thus at the same price), it would be okay to conclude that they should yield similar amounts of gold, even though Pentium ones have that additional amount of gold at their core (and, in fact, the total amount recovered from Pentiums was greater than the AMD). On the other side, they weigh much more, so the gold/kg ratio meets the AMD one. That's why they are paid the same. Or at least that's what I think :)

Winged.

Let me rephrase to avoid any misunderstandings. The amount of gold you got from the Pentiums was not enough. The recovered yield was wrong. You haven't got all the gold. One or all of the three 8) 8)

Also as mentioned your RAM yield is light. If people are suggesting to you that you're low on your returns and they've processed quite a lot of this stuff I'd thoroughly recommend taking their advice and checking out your processes. We've all got things to learn. I still learn every single time I go out to process gear so I'm genuinely passing on knowledge- not trying to be smart for the sake of being smart. 8)

Jon
 
anachronism said:
Wingedcloud said:
Also, since the ceramic Pentium are usually bought by gold refiners within the same category as the AMD ones (and thus at the same price), it would be okay to conclude that they should yield similar amounts of gold, even though Pentium ones have that additional amount of gold at their core (and, in fact, the total amount recovered from Pentiums was greater than the AMD). On the other side, they weigh much more, so the gold/kg ratio meets the AMD one. That's why they are paid the same. Or at least that's what I think :)

Winged.

Let me rephrase to avoid any misunderstandings. The amount of gold you got from the Pentiums was not enough. The recovered yield was wrong. You haven't got all the gold. One or all of the three 8) 8)

Also as mentioned your RAM yield is light. If people are suggesting to you that you're low on your returns and they've processed quite a lot of this stuff I'd thoroughly recommend taking their advice and checking out your processes. We've all got things to learn. I still learn every single time I go out to process gear so I'm genuinely passing on knowledge- not trying to be smart for the sake of being smart. 8)

Jon

Hello,

I inicially decided to publish these results, because I had the feeling that they were somewhat short on what the actual yield should be.
Following your statement that my Pentium yield was wrong, I decided to do some digging on some other refiner's yield and came across some valid information: according to yield document from Samuel-a (which I believe is an overall accepted yieldlist), Pentiums like the ones I processed yield about 0,072g Au/unit. My results have given 0,064g Au/unit. A bit below (and I recall losing some solution when I thought about the whole process, and the whole mess that led to the cementation topic I created :D), but perfectly within "acceptable" range.
About the RAM's, and even though I respect other people's results, I can't understand how they can recover those values from RAM IC, when I constantly get low yields. On the other hand, I get very good yields from BGA chips, which are considered high grade, using the process I always use.

I am not trying to deny other people's results, because they surely haven't created the gold when it wasn't there. But, as I believe that my process is being made correctly (and I am being as humble as I can here), I try to find explanations to why my results are below what other people are getting on the same material.
Picking up on denim's statement on RAM yields, 4,5g from 2,583kg of chips, gives a 1,74g Au/kg, which is perfectly within the amounts present in some yieldlist, as for the fingers. But I don't know why, in my material, it's just not there.
Also, just saw a yield list that states that in 100 RAMs, there are 0,6-0,9g of gold. Well, my results are well above that range.

Winged
 
Winged...as someone who is not vested in this conversation.

If Jon says you are missing gold, you should be looking for it, or very specifically pointing out your procedure for others to critique...not trying to defend your current numbers.

He makes his living processing literal tons of it, whereas the hobbiests producing and sharing the data have maybe processed pounds.

It’s really in your best interest.

Otherwise, impressive post.




Sent from my iPhone using Tapatalk
 
I will add...I understand peoples love of the panning, blue bowl, gravity separation, etc. The reality however is that when you get in to RAM and other small chips, unlike BGA's, is that the length of the wire is very very small. It's very easy for these pieces of wire to catch a ride on a 100 mesh piece of ash, and wash right on out of your collection riffle.
 
snoman701 said:
Winged...as someone who is not vested in this conversation.

If Jon says you are missing gold, you should be looking for it, or very specifically pointing out your procedure for others to critique...not trying to defend your current numbers.

He makes his living processing literal tons of it, whereas the hobbiests producing and sharing the data have maybe processed pounds.

It’s really in your best interest.

Otherwise, impressive post.




Sent from my iPhone using Tapatalk

Hey snoman701,

Thank you for your reply.
I am not trying to discredit other people's data and statement. Far from it. And if Jon really has the experience from processing tons of this material, his knowledge should really be taken into consideration. No question about it. And I would be delighted to expose my procedure in detail, so he or someone else can point out some mistakes, so I can improve.
I am just trying to figure out why my numbers are below the acceptable yields, using the means at my disposal.

snoman701 said:
The reality however is that when you get in to RAM and other small chips, unlike BGA's, is that the length of the wire is very very small. It's very easy for these pieces of wire to catch a ride on a 100 mesh piece of ash, and wash right on out of your collection riffle.
Now that is a reason I am willing to accept to explain my low yield on RAM IC !!!
What can I do to prevent that from happening?

Winged
 
Hey Wingedcloud, my 2c as someone who has processed many kgs of ram . . .

If all your ram chips were the type with legs, I bargain on around 1g / kg and get numbers around that very reliably, BGA RAM can vary tremendously, I've had as low as 2g/kg up to 6g/kg. Your picture looks like all "legs" ram, or was it mixed legs and BGA ?

Ram fingers I work on 4g/kg on average.

Unfortunately I save up my ceramic processors instead of processing them, so can't offer much advice there.
 
kernels said:
Hey Wingedcloud, my 2c as someone who has processed many kgs of ram . . .

If all your ram chips were the type with legs, I bargain on around 1g / kg and get numbers around that very reliably, BGA RAM can vary tremendously, I've had as low as 2g/kg up to 6g/kg. Your picture looks like all "legs" ram, or was it mixed legs and BGA ?

Ram fingers I work on 4g/kg on average.

Unfortunately I save up my ceramic processors instead of processing them, so can't offer much advice there.
Hello kernels, thanks for your input.

If I recall correctly, it was mixed "legs" and BGA, but mostly "legs".

So, using your numbers, on 520g of fingers, there should be 2,08g of gold (about half of my total amount).
And on IC, considering it was mostly leg chips, on a 1g/kg ratio, there should be 2,58g of gold, for a total amount of 4,66g, making my yield a bit below this prediction.
I accept your values, since I have made an estimate based on similar ratios, before I processed the RAM.
I am truly truly truly confident I did not lose any on the fingers (with AP), so I guess I'll just have to be a little more careful on the IC powder washing process.

EDIT:After kernels input, I went through all of the RAM leftovers and counted the BGA chips type ones, for a total amount of 24 out of 415, which represents around 5,7%. So, I guess considering the value of 1g/kg is pretty much accurate, as it is valid for more than 94% of the RAM.
Considering this, and being the value of 4.66g/kg valid for the matter at hand, how can these RAM be worth 25 or 26€/kg, when there is not enough gold (or copper or palladium) to even cover for the material costs, let alone the processing?

Let me also point out that, there where some memory sticks that weren't the traditional RAM ones we are used to, so maybe the IC on those wouldn't even have the 1g/kg kernels presented.
But since I can't be 100% sure, I will stick with the "be more careful with the washing" conclusion.

I am sorry for seeming stubborn and not wanting to accept the likely fact that I lost some of my gold (on what the RAM IC prcessing is concerned), but even with valid inputs by kernels and other members, which I truly respect, the numbers don't make sense to me, and they have to make sense for me to believe in them, because they are "ground zero" in all of this.

Thanks again!!

Winged.
 
You got 520g of fingers, i expect 800+g with that base weight, probably all the RAM were kinda new? But then again, i expect avg~18g per RAM and it is almost exactly what you got.. where do we drift off?


Edit: lots of old laptop rams with heavy parts and few gold?
 
Syn said:
You got 520g of fingers, i expect 800+g with that base weight, probably all the RAM were kinda new? But then again, i expect avg~18g per RAM and it is almost exactly what you got.. where do we drift off?


Edit: lots of old laptop rams with heavy parts and few gold?

Hello,

From memory, I recall most of the RAM were the traditional type, but some of them were the old ones you talk about, with heavy IC. Not many laptop ones, if I recall correctly. But has I said, most of them were the usual ones (I could risk saying more than 90% of them).
So I don't really know how to explain the "drifting" :lol:

Winged
 
Back
Top