IBM ceramic CPUs and their boards

Gold Refining Forum

Help Support Gold Refining Forum:

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

kjavanb123

Well-known member
Joined
Apr 1, 2009
Messages
1,743
Location
USA
Hi

I was at a scrap yard and these old goodies got my attention. Total of 10 IBM p160 CPUs and 1 Pentium CPU.

They were in these boards. They will be processed this weekend and I will post results.

Any estimate on CPUs yield? Last time I processed these and got 0.5g of gold per CPU.

CPU collection
image.jpg

Boards that CPUs were on them
image.jpg
 
Don't destroy these!
Sell them on eBay via auction, one at the time and quadruple your profits.
Collectors love them.
 
Hi

I agree with you on prices they can go on ebay for a collector, but currently I am in Iran and ebay is filtered here.

Best regards
KJ
 
kjavanb123, If you are not familiar with it, there is this PDF document here "GNS - e-Scrap Yield List v1.0" by member Samuel-a, It has a lot of yields on different CPU's and it seems to be trusted universally here - incidentally it lists Au in the IBM 6x86 in the array of 0.45-0.55 g/unit - so about the same as your verified yield... I haven't processed any gold-cap CPU myself, but before putting any up on e-bay I usually check this list for the minimum asking price :)
 
Hi Arthur

Yes I know about that list. The IBM CPU there is P228 model and these were P150+ models.

The primary precipitation doesn't look good at all to be similar to those numbers in the list.

I will get the gold melted tommorow and also try to get the silver chloride converted to metal ans report my findings here.

Best regards
KJ
 
Hi

Finally the result of chemically processing the 10 IBM 6x86 p150 is in.

Total gold recovered 0.32 grams from 10 units

Total silver recovered 1.01 grams from 10 units

Exteremely disappointing results which means on average each unit has 0.032 grams of gold and 0.1 grams of silver.

Here is the photo of crushed CPUs after process conpleted
IMG_4237.JPG

This shows the precitpited gold from AR solution
IMG_4236.JPG

This is the gold bar, NOTE: the weight is combined weights of gold from 10 CPUs and another 2 gram gold.
image.jpg

This is recovered silver, converted from silver from 10 CPUs
IMG_4238.JPG

This is silver bar from above
image.jpg

The entire process took 5 hours to dissolve gold and silver from 10 units of ceramic IBM CPUs

Hope that helps people who might come accross these types of CPUs and as it is advised selling them on eBay would be the only viable choice.


Best regards
KJ
 
I've sold several CPU:s like that for $10 each on ebay. It's clear that I did the right thing.

Göran
 
kjavanb123 said:
Hi

Finally the result of chemically processing the 10 IBM 6x86 p150 is in.

Total gold recovered 0.32 grams from 10 units

Total silver recovered 1.01 grams from 10 units

Exteremely disappointing results which means on average each unit has 0.032 grams of gold and 0.1 grams of silver.



The entire process took 5 hours to dissolve gold and silver from 10 units of ceramic IBM CPUs

Hope that helps people who might come accross these types of CPUs and as it is advised selling them on eBay would be the only viable choice.


Best regards
KJ

No youve got your process wrong or something drastically incorrect. Those yield figures are completely incorrect. That's the plain truth of the matter.
 
My take on this type:

https://www.youtube.com/watch?v=irMDfWA5SM8&t=1s

slightly better than your result but still very disappointing
 
I don't know guys, seems awful low...

kjavanb123, did you process lids also?


Someone should address this, is this http://www.goldnscrap.com/yieldlist.pdf list that wrong...

Doesn't make that much difference to me, since I still mostly sell them on e-bay anyway, but I kind of got the feeling that Sam's list was GRF approved :)

Or could it be because if the model type (6x68MX pr233 vs 6x68 p150+)... checking Owltech's video it looks like he had 6x68L pr166+, at least one of them
 
Owltech, so in average about 0.12 g/cpu.

It seems like IBM used a lean manufacturing process compared to Intel. I've seen them using glue (epoxy) instead of braze to fasten the cpu:s in ceramic bodies. Maybe that is a good explanation why the results for these chips are low.

Makes it an even better deal to sell the CPU:s on eBay.

Göran
 
I`m no CPU-guy, but this intrigues me, mostly because of the GNS - e-Scrap Yield List v1.0

Judging by this link there is not that much of an age difference for these variations of IBM 6x86.... https://cpumuseum.jimdo.com/museum/ibm/6x86-6x86mx/#swap

but they were produced both at IBM (America) and SGS-Thomson (Europe)....
 
Owltech said:
My take on this type:

https://www.youtube.com/watch?v=irMDfWA5SM8&t=1s

slightly better than your result but still very disappointing

There's a world of difference between your 0.1225 per chip and Kevin's results, it's not just slightly different. I'm not sure where Kevin's gold went.
 
IMHO, It seems like some of the solder on the gold cap was not dissolved, and also the gold solder inside the silicon chip...
I believe there's still some gold in them processed CPU's.
Sometime ago, I processed similar CPU's, Pentium Pro's to be exact, I had left behind some gold...
https://goldrefiningforum.com/phpBB3/viewtopic.php?f=52&t=12845&hilit=pentium+pro&start=210#p128293
I hope this helps.

Take care!
Phil
 
Hi

After trying to re-dissolve crushed IBM cpus, I could squeeze another 0.1g of gold picture below,
image.jpg

And another 0.3 g of silver. So the final yield from processing 10 IBM CPUs 6x85 p150+ is as following

Total gold: 0.420g
Gold per CPU: 0.0420g

Total silver: 1.31g
Silver per CPU: 0.131g

Conclusion if you find them sell them on eBay.

I am processing the boards and will post yieldsin a different topics.
 
Thanks Kevin

I will say though that my yields for these are almost identical to Owl's yields so I still suggest that something has gone awry.

Jon
 
Pins alone should reach Kevin's numbers.

For example, take AMD K6-2: same Socket 7, same number of pins but with alu-lid and flip chip, and still yields 0.03-0.04g/CPU.
 
Back
Top