Motherboard yield report.

[NoIdea]

I thought a more lower estimate of 0.05 - 0.1g per CPU would be more realistic, well maybe?

Deano

 

[Geo]

when your talking about averages, i guess its all relative. ive seen motherboards that was loaded with flatpaks and ive seen the cheap Chinese boards that didnt even look like the pins were plated.ill get some better yield info when summer gets here and i can set up a copper parting cell. honestly, thats the only way to know for sure. cast all the metal from a random sample in an anode and part the copper and refine the metal.

 

[rucito]

30% of 450 = 135gr. metals in 1 board
0.1% of 135 = 0.135gr. gold

 

[voeckel]

I thought a more lower estimate of 0.05 - 0.1g per CPU would be more realistic, well maybe?

It was a generous deduction to see what's left for the boards themselves once you take off the CPUs. But there is still a mistake in my calculation due to applying the values to 100% of the total weight, not to the 30% metal content! That's why the numbers still seemed to be very high to me.

Hence below is now a final calculation to clean up the mess a bit.

Baseline:
1 MT of boards with CPU = 2000 boards
1 MT = 300 Kg of metals
0,1% Au of 300 Kg = 300g Au (including CPU on board)
Au content of average CPU = 0,07g
2000 x 0,07g = 140g Au in CPUs

300g Au board in total - 140g Au CPU in total = PCBA Au content = 160g/MT

Should be a good guideline now for motherboard yields, although it does not mention any board grade. Let's assume therefore that it is a colorful mix of grades and the numbers come from an average of 100 MT of processed boards. :idea:


EDIT:

0.1% is about right from my experience. individual components yield different totals but given the overall weight of a motherboard,it sounds right. of coarse you would have to include the CPU with the motherboard.

Geo, are you sure that this would be really right on the overall weight of the board (so the 100%, incl the 70% non-metal fraction)?? If so, then my previous calculation would have been right but honestly, I doubt it a bit. Just want to be sure on the statements here so that we don't publish a lot of wrong calculations for newbies.
Please let us know once you made your test.

 

[samuel-a]

Voeckel

160g per MT is very low.

Typical assay for mixed MB types ranges from 240g to 350g Au.

Though, maybe colorful MB could yield this low...
Yield are a tricky business... You have to define the source material very particularly.

 

[zzz]

Processed with nitric the skeleton of this chip:
recovered some silver, i think 2 to 4 grams. I'll make and weight the button in 2/3 days.
See you.

 

[Geo]

i have three months of winter left and from the looks of it,its going to be a wet one. when you talk about yield of a particular board, there are variables from board to board even in like models due to costumer specifications. on older models,0.1% may be low and on newer models it may be a little high. you have to consider that most computers were built very specifically for each individual and components from board to board will vary. the hidden gold on a motherboard may exceed the visible gold by 2 to 1. there are motherboards that appears to have no CPU but rather are covered in flatpak chips. these have no VGA chips (north/south bridge) but still contain much more gold than boards with a processor.

this can go on and on. even if you took two of the same boards from the same computers,they could still contain different amounts of precious metals.

 

[johnjohny]

Hi,
apologies for reviving ancient thread. Got bored so started reading forum and found this topic with a bit of inconsistency.

"In general, waste PCBs contain approximately 30% metals and
70% nonmetals. The typical metals in PCBs consist of copper (20%),
iron (8%), tin (4%), nickel (2%), lead (2%), zinc (1%), silver (0.2%),
gold (0.1%), and palladium (0.005%). The purity of precious metals in waste PCBs is more than 10 times higher than that of rich-content minerals."

Anyhow I am calculating I can't get it right.
If copper + other listed metals are listed as % of 30% then copper would amount to 20% of 30% = 6%. That would mean that the other metals listed (and non listed) woudl have to make the remaining 94% (out of 30%). It doesn't add up.

If copper + other listed metals are listed as % of total weight then just the listed metals would consist more than 30% (20+8+4+2+2+1...) and apparently metals are 30% of total weight. It doesn't add up either.


Don't know where the information came from, but looks very fishy.
There is a possibility that I am missing something and if so I'll be thankful if somebody points out my mistake.

Regards

 

[etack]

I think you are taking "approximate" numbers literal. Approximate will have some deviation to them. (+-) 6% on metal content of PCB MB is very possible.

Eric

 

[g_axelsson]

I can't see aluminum in that list. It would be a large part of the metals in a PCB.

Göran

 

[Smack]

It depends on pre-processing of the boards before they are processed and made into dore bars. There is very little aluminum on the boards I take in.

 

[g_axelsson]

Even electrolytic capacitors are removed? But if this is a test made on dore bars then I can understand it. Aluminum is really hard to alloy with some metals and I guess most of it will oxidize and end up in the slag.
Other metals I would expect but didn't see is tantalum and beryllium, not so big parts but still metals that exists on boards. But then that list never claimed to be complete.

Göran

 

[Smack]

Yes even those, I would say at least 85 to 90% of them and the Ta caps too. For me if I have time, after cherry picking the boards for the stuff I want to process myself, I take off as much stuff as I can that I will not get paid for from the refinery. For every pound I take off in material they don't pay me for I save a dollar ($1.00 per pound processing fee) and if I can take that same material and sell it for lets say $.25 per pound, I'm $1.25 to the good. The electrolytic caps I sweat to get the aluminum and sell that as Old Cast Aluminum for $.50 per pound and whats left after that process I throw in with the sheet steel and make an additional $.12 per pound. The only thing I'm taking off that I don't have a market for any more is plastic but I'm still saving $1.00 per pound. You might say, (Well how much does all that time cost you?) and like I said before, if I can do it and still keep up with the flow of material then it costs me nothing but my time. I have a couple of retired guys that like to come and tear stuff apart for something to do. Gets them out of the house and keeps the relationship with their wife good as it gives the women a break from the men too.

 

[g_axelsson]

That's nice, Smack, and I would do the same with my scrap if I sent it for refining and not selling it.
But as a basis of calculate metal contents in general motherboards you are skewing the numbers. Taken to an extreme I could state 100% Au in motherboard, as that is what I get in the end. :mrgreen:

Göran

 

[Geo]

it cant be a true analysis of percentages unless there can be an agreement on what the beginning measurement will be. if you detract from the weight of the motherboard straight out of the computer, your percentage will vary from each individual that tries to quantify them. did you remove the heat sink? did you remove the CPU? did you remove capacitors? of coarse there are computers like the early Apple's that did not have a CPU and there weren't any electrolytic capacitors. if the subject was divided into models or even series, then data could be reliably obtained. even then, like i said before, even computers with the same model numbers may have different components installed. i believe that instead of trying to get a definitive percentage amount, it would be easier to set a baseline amount (e.g. no motherboard should have less than X amount). even if you or i swear that the stack of motherboards SHOULD contain this much gold that you know for sure that it DOE'S contain at least this much gold.

 

[Smack]

I was, I guess on a slightly different topic and not so much percentages of metals in motherboards. Nothing is skewed, I was talking about the stuff I take off all boards.

 

[AlchemyGeorge]

Hi gentlemen,
Merry Christmas everyone!

Regarding the yield of motherboards, below is a paragraph from a chinese research paper. read carefully for the numbers bug me all the time here and can't be right.



This would mean that there would be typically 1 KG of Au in a MT of PCBA scrap!

What can the ones with commercial experience tell us please? I think, this must be way, way off. Maybe 1/10 of this or even less?
As you can see they do not even provide details on the board grade used for the tests to get these numbers. :roll:

I think that’s 0.1% of the metals recovered…not of the weight of the entire board. Just a thought.