Why do ceramics go for more than gold tabbed BGAs?

[mysticrey2414]

So I’ve gone over this in my head multiple times and I feel I must be missing something. So if you go to boardsort or any other ewaste buying website, the price they pay for ceramic CPUs is way higher than the price of gold tabbed BGAs and this baffles me, because it seems that no matter how I calculate it, (using data from many different resources, especially these forums) the tabbed BGAs appear to be far superior to any ceramic CPUs in regards to yields. For example, even with the fiber still attached, I’ve noted people yielding up to 5 grams of gold per 1 kilogram of chips. Counting just the chips and no fiber, up to 10 grams per kilogram of chips. While on the other hand, the highest yield I can find data for with even the best ceramic CPUs is anywhere from 2-4 grams per lb or 4-8 grams per kilo (just rough estimates from things I’ve taken note of)
Now boardsort pays 15$/lb of gold cornered BGAs, and anywhere from 67-190$/lb of ceramic CPUs depending on the type of cpu and that’s not counting the lower grade ones like k6.
What gives? Am I forgetting to calculate something? Just surprised they only pay 15$ per lb of the tabbed BGAs

 

[kurtak]

The black epoxy chip of a BGA is where MOST of the gold in a BGA is - some (but very little) in the green fiber part of the BGA

The black epoxy (only) chip has "about" 4 grams gold per pound --- the black part of a BGA is close to half the weights of the whole BGA (black chip plus green fiber)

So the recovery on the full BGA chip (black epoxy plus green fiber) is actually only about 2 grams per pound

Boardsort is buying the BGAs as full chips (black epoxy plus green fiber) so buying on 2 grams per pound - not 4 grams per pound

That said - I agree - their price is low

I just wanted to clarify that what they are buying is a 2 gram per pound recovery & not a 4 gram per pound recovery

My above numbers are based on doing 20 - 30 pound batches BGAs

Kurt

 

[eaglekeeper]

So if you go to boardsort or any other ewaste buying website, the price they pay for ceramic CPUs is way higher than the price of gold tabbed BGAs and this baffles me

Chris from Boardsort pop's into the forum from time to time, maybe he can answer that question. But Boardsort being a business with competition they're yield data may be proprietary info.

Am I forgetting to calculate something? Just surprised they only pay 15$ per lb of the tabbed BGAs

I'm sure that the recovery of all the base metals may be a consideration, how much...I don't know.

 

[Cheffrey17]

So I’ve gone over this in my head multiple times and I feel I must be missing something. So if you go to boardsort or any other ewaste buying website, the price they pay for ceramic CPUs is way higher than the price of gold tabbed BGAs and this baffles me, because it seems that no matter how I calculate it, (using data from many different resources, especially these forums) the tabbed BGAs appear to be far superior to any ceramic CPUs in regards to yields. For example, even with the fiber still attached, I’ve noted people yielding up to 5 grams of gold per 1 kilogram of chips. Counting just the chips and no fiber, up to 10 grams per kilogram of chips. While on the other hand, the highest yield I can find data for with even the best ceramic CPUs is anywhere from 2-4 grams per lb or 4-8 grams per kilo (just rough estimates from things I’ve taken note of)
Now boardsort pays 15$/lb of gold cornered BGAs, and anywhere from 67-190$/lb of ceramic CPUs depending on the type of cpu and that’s not counting the lower grade ones like k6.
What gives? Am I forgetting to calculate something? Just surprised they only pay 15$ per lb of the tabbed BGAs

I've also been curious about this as well, every video I've seen they always mention how the ceramics are much better than bga's in recovering gold, but the price to even buy them at just seems way to outrageous for me to justify it. i have about 4 ceramics cpu's I need to process but id like to save them up until I get about a pound to see what it's all about.

Just wanted to mention that and post on here so i could get some answers as well!

 

[mls26cwru]

its relatively simple... its the same reason they pay what they do for gold fingers... They would rather have the boards come to them as a complete unit and not picked over. They know the numbers and returns of whole boards and base their numbers off that. If they offer too high a price, they incentivize people to depopulate the boards and turn them into a commodity that is now "unknown".

in addition, they are board buyers... they turn over a semi full of boards a week (probably more)... so they are looking to turn material and meet quotas/lot minimums as quickly as possible. If they buy 5 lbs. of fingers, they have to wait until they can get a batch large enough to send in for refining and tie up the capitol until then... its not a big deal if its a 10lb minimum, but what if its a 500lb. minimum? (i don't know what their minimum batch sizes are, but its another thing to think about in the economics of running a business)

 

[speed]

On the open market ceramics command a premium (to the point of being more then thier gold content is actually worth!) because they are easy to process and quite frankly most people are ignorant and don't realise there's more gold in a bit of black plastic than in a shiny gold plated package.

 

[GoIdman]

The recovery of gold from ceramics are straight forward, the total cost of the recovery and refining (materials, time, consumables ect.) is somewhat lower then processing black epoxy BGA, therefore if you want to end up with the same price of your recovered gold, you have to compensate somewhere. In this case the lower buying price compensates the investment.

Pete.

 

[kurtak]

The recovery of gold from ceramics are straight forward, the total cost of the recovery and refining (materials, time, consumables ect.) is somewhat lower then processing black epoxy BGA, therefore if you want to end up with the same price of your recovered gold, you have to compensate somewhere. In this case the lower buying price compensates the investment.

Pete.

Hmmmm - I am not sure I agree with that

I have processed large batches of both ceramic and BGAs (10 - 30 pound batches) & IMO it is kind of a wash on the cost for recover

Kurt

 

[Goldwasser]

its relatively simple... its the same reason they pay what they do for gold fingers... They would rather have the boards come to them as a complete unit and not picked over. They know the numbers and returns of whole boards and base their numbers off that. If they offer too high a price, they incentivize people to depopulate the boards and turn them into a commodity that is now "unknown".

in addition, they are board buyers... they turn over a semi full of boards a week (probably more)... so they are looking to turn material and meet quotas/lot minimums as quickly as possible. If they buy 5 lbs. of fingers, they have to wait until they can get a batch large enough to send in for refining and tie up the capitol until then... its not a big deal if its a 10lb minimum, but what if its a 500lb. minimum? (i don't know what their minimum batch sizes are, but its another thing to think about in the economics of running a business)

Mls26cwru hit it on the head. Discouraging harvesting is certainly a factor but not the driving reason. There is more to consider:

There are many grades and conditions of gold tabbed bga's so not all are alike, at least to us. The condition may be poor on some, the top may be missing on others, and still more may not even be qualify as a gold tabbed bga and are instead flip chips or worse. This is particularly true with the majority of the clean gold fingers we see. The quality of the fingers tend to vary widely with more fiber than gold, or a mix of flashing and plating. The quality of fingers and bgas that we buy suffers the deeper into the non-professional scrapping hole we climb ( and I mean that with no disrespect), so the price suffers as well. Missing epoxy tops, 50% or higher fiber to gold ratio, and a sprinkling of flip chips all play factors into our overall pricing on bga's and Fingers.

Contrast that to a Intel 486 or a Pentium Pro where short of someone cutting the pins (the the obvious heatsink) the returns are extremely consistent because the product is consistent. A PPro is a PPro and so on. We can comfortably pay right up to the stated value with little risk. Not so much with publicly sourced fingers and bgas where consistency is non-existent, at least in the bulk that we see. Our returns have varied wildly on bga's and Fingers which is reflected in the price.

 

[Dreamer]

This may not be the right place to throw this out there. However, I have a couple dozen intel BGAs so far. I've ran a couple runs using H2SO4, rinse and went to AR straight away. From my test runs, I calculated my yield at 3.5 grams/kg. Wondering if incineration followed by gravity separation would get me closer to the aforementioned 5gms/kg.

 

[g_axelsson]

This may not be the right place to throw this out there. However, I have a couple dozen intel BGAs so far. I've ran a couple runs using H2SO4, rinse and went to AR straight away. From my test runs, I calculated my yield at 3.5 grams/kg. Wondering if incineration followed by gravity separation would get me closer to the aforementioned 5gms/kg.

There's no law stating that there has to be 5 g/kg gold in epoxy top BGA:s and producers always tries to save on expenses. If you get 3.5 g/kg then it probably is because that was how much there was in the chips.

I recently ran (this spring) a few lots with old new stock, ie a bunch of identical chips in decent numbers so I got the data for a single chip and the two types ended up with 2.5 and 4.5 g/kg.

I did some experimenting on my methods and I ended up with a decent method that lets me run up to a couple of kilos in a day. Whenever I test new methods I always run any leftovers again to see if there is any gold left in the solids. That is the only way to know if you are losing gold along the way.

My suggestion is that you take a representable sample of anything discarded and do a leach for gold. A quick stannous test should tell you if there are any major losses in your method.

Göran

 

[kurtak]

There are many grades and conditions of gold tabbed bga's so not all are alike, at least to us. The condition may be poor on some, the top may be missing on others, and still more may not even be qualify as a gold tabbed bga and are instead flip chips or worse.

I recently ran (this spring) a few lots with old new stock, ie a bunch of identical chips in decent numbers so I got the data for a single chip and the two types ended up with 2.5 and 4.5 g/kg.

Hmmm - Interesting - I don't have time to post more right now as I need to turn off my computer & get going for the day

I will try to post more tomorrow because I have processed a fair amount of BGAs & have had different (better) results

Maybe we are talking about different things (types of BGAs)

Kurt

 

[Brett Novak]

I've been running CERDIPS for a week now. The major difference is with a CERDIP the acid can leach ALL of the gold out without grinding. In a plastic package this isn't possible due to the complete encapsulation of the plastic around the bond wires. I'm getting .7g from a 6.5g CERDIP in lost weight (haven't done the math yet on gold content, but i've yielded 4.15oz from ~8Kg of stock.)

 

[Brett Novak]

Oh - another question - are you using NEW BGA's or de-soldered BGA's? And what "vintage" - old ones from the 80's and 90's have gold bond wires. Newer ones (post 2016) may have copper palladium or gold palladium bond wires. If they are all the same chip, you can usually go on the manufacturer website and get a bonding diagram which will tell you exactly what they are. But, again for plastic packages, burn, grind, then soak in AR.

the "balls" on BGA's are usually silver or CuPd by the way. Only some of the old exotics had gold bond balls.

I'm heading out on vacation but plan to do some experiments with QFN, BGA and TQFP packages when I get back and I'll share what I find. But, I'm expecting burn, grind and soak is going to be the optimal process with a minimal amount of return. I'll do a side-by-side of plastic versus CERDIP by weight.

 

[Yggdrasil]

Oh - another question - are you using NEW BGA's or de-soldered BGA's? And what "vintage" - old ones from the 80's and 90's have gold bond wires. Newer ones (post 2016) may have copper palladium or gold palladium bond wires. If they are all the same chip, you can usually go on the manufacturer website and get a bonding diagram which will tell you exactly what they are. But, again for plastic packages, burn, grind, then soak in AR.

the "balls" on BGA's are usually silver or CuPd by the way. Only some of the old exotics had gold bond balls.

I'm heading out on vacation but plan to do some experiments with QFN, BGA and TQFP packages when I get back and I'll share what I find. But, I'm expecting burn, grind and soak is going to be the optimal process with a minimal amount of return. I'll do a side-by-side of plastic versus CERDIP by weight.

Most balls are just plain Tin

 

[kurtak]

Oh - another question - are you using NEW BGA's or de-soldered BGA's? And what "vintage"

Concerning BGAs you posted --------

Newer ones (post 2016) may have copper palladium or gold palladium bond wires

This is not just mis-information but is in fact FALSE information

Bond wires in IC chips are ether gold - or aluminum --- nothing else - & it does not matter when they were made or who made them (other then maybe stuff made in the soviet union)

How do I know that ?

Because I used to process several hundred pounds of IC chips per year - over the course of those years I processed somewhere around 2,500 - 3,000 pounds of IC chips (ALL types of chips)

IC chips my well have copper wires in them (or Kovar) - but the copper does not bond to the silicon die - that is where the gold is used - to make the connection (bond) between the copper (or Kovar) and the silicon die

Palladium (&/or other PGMs) are NEVER used as bond wires in IC chips (other then maybe soviet union stuff)

Processing IC hips was just one type of material I processed to make a living refining PMs --- in others words - refining was NOT just a hobby for me - so I am quite sure I know what I am talking about

Kurt

 

[snoman701]

Hmmm - Interesting - I don't have time to post more right now as I need to turn off my computer & get going for the day

I will try to post more tomorrow because I have processed a fair amount of BGAs & have had different (better) results

Maybe we are talking about different things (types of BGAs)

Kurt

When you were doing BGA's, were you splitting the green from the black?

I got 2.2 g/lb a few years ago just by burning a small handful of chips then smelting with silver as a collector assaying gold by XRF.

 

[snoman701]

One thing I intended to say. Boardsort pays what they do also because refiners use them as a dumping ground, artificially lowering their assay because they got a bad batch of chips (there are Intel BGA's that show up every once in a while in bulk that don't have gold bond wires).

Same thing they say about gold fingers. It goes from close cut fingers and quickly progresses to flashed trimmings.

 

[Goldwasser]

One thing I intended to say. Boardsort pays what they do also because refiners use them as a dumping ground, artificially lowering their assay because they got a bad batch of chips (there are Intel BGA's that show up every once in a while in bulk that don't have gold bond wires).

Same thing they say about gold fingers. It goes from close cut fingers and quickly progresses to flashed trimmings.

Snoman701 gets it. And when you deal in bulk, speaking specifically to BGAs and fingers, those issues are costly to correct, through time, labor and reduced assays. And even though a customer may have the closest cut fingers and the best BGA chips made, it won't make a dent in our overall return as that premium will quickly be diluted by the aforementioned bulk.

 

[mysticrey2414]

The black epoxy chip of a BGA is where MOST of the gold in a BGA is - some (but very little) in the green fiber part of the BGA

The black epoxy (only) chip has "about" 4 grams gold per pound --- the black part of a BGA is close to half the weights of the whole BGA (black chip plus green fiber)

So the recovery on the full BGA chip (black epoxy plus green fiber) is actually only about 2 grams per pound

Boardsort is buying the BGAs as full chips (black epoxy plus green fiber) so buying on 2 grams per pound - not 4 grams per pound

That said - I agree - their price is low

I just wanted to clarify that what they are buying is a 2 gram per pound recovery & not a 4 gram per pound recovery

My above numbers are based on doing 20 - 30 pound batches BGAs

Kurt

Thanks you for your insight! It’s helpful to have solid numbers from someone with that much experience. I definitely agree, even at 2 gram/lb recovery rate, 15$ a lb for these BGAs just seems low. I’m glad I’m not alone in wondering about That.