synthetiklone
Well-known member
Here is the complete process how I run 386/486 cpus.
I have 50 CPU's, which I am measuring the results of.
There is a substantial amount of Au inside these CPU's, maybe some you are unaware of.
I will hopefully be able to attach some pictures of my processing.
To obtain the most from each my procedure is as follows.
Heat 4 CPU's with pins up on barbeque center plate, using a rotational method so first on - last off. This economises the waiting time for them to heat up.
Using long screwdriver to hold CPU, and sharp knife blade, pry off the gold lid. Put lid aside to cool off.
Carefully pickup the almost red hot CPU WITH pliers/tongs!, and put somewhere to cool off. These are ceramic, and can take awhile to cool down, so be careful.
Put another fresh CPU on the barbie, and do the last one which should be hot enough by now.
Carry on until all done.
I ground an old small router tool bit to make a custom micro chisel-like scraper tool. Use this to scrape off the fine gold wires running around the edge of the silicon wafer chip. Do this over a clean white surface to catch the fine gold wires that may inadvertently fall off. Scrape linearly along the ceramic/gold edge first, and then when they are all cut, run it around the silicon wafer edge. Try and get it close to the welding points, but dont use too much pressure as to damage the glass wafer edge. Try not scratch the wafer at all, as you will end up with bits of silicon in the mix, and also, if the wafer is in good condition, you can sell them for more than the value of the gold content of each CPU! They are sought after by collectors, and people in the microscopics and education industries. The gold wires will bundle up, and get stuck with some still attached to the weld points. Invert the chip, and tap the fine gold wires off onto a clean white surface (a porcelan plate, for example). Check with a loupe or stronger magnification eyeglass that you have as much fine gold wires removed. Complete all the CPU's.
I used the flat clean end of my scraper tool to manually compress all the wires into a cube, and obtained about a 3x3x2mm cubic block of gold from them. I am sure this is PURE gold, no refining may be necesary, but when you melt this block, it may shrink a bit as there must still be some void air space within the block. There is also, you will notic in the zoom pic of the block, that frags of silicon are amongst the wires, so processing in AP/AR might be necessary.
The CPU lids must be cut into quarters or more with tin snips to maximise the area required for the nitric to work on. I placed all the bits in a large beaker, and added nitric acid.
I first tried making nitric using LS's cooling process available here, and from potassium nitrate, with ok results. By ok, I mean that it would be days before even a small etching of the base material would take place. I tried heating the nitric mixture, and ended up accidentally overheating, causing alot of red fumes, and overboiling. It was lucky I had a large beaker, as the fizz almost left the container !
I am sure this nitric would be fine, if made properly. I found I was wasting alot of potassium nitrate, I am not sure about how much white crystal (salts) should be left after the cooling process, but for me there was alot, and maybe something wasn't correct in my procedure. I have since been given some 70% nitric acid, and this only took one day for the foils to float away loose from the lid base material.
I warmed the foils in nitric one last time to free up all the gold, and then tipped off the nitric. Refilled with water, and emptied into a plastic tray, where I quickly manually seperated the gold foils from the lid base material (nickel?) which had only the edges slightly etched.
Once this was done, I discard the base material, and boil the foils in water a couple of times to remove the remaining nitric. I will now put these foils in some hydrochloric acid as a final cleanup, as I can see there is still some very fine copper and nickel colour discoloration on some of them.
I am impressed with the quantity and quality of the gold recovered so far.
Get the barbecue back out again, and begin the same process as you did removing the lids, but we are now going to carefully remove the silicon wafer chips.
While CPU is still in contact with blue flame (must be VERY hot), use a VERY sharp blade and screwdriver to hold the CPU, to pry off the silicon chip. Be gentle, as it is like thin glass, and can chip/break easily. Have a file on hand so you can quickly resharpen the blade, you need a fine egde to get under the wafer and flip it up off the molten glue/solder base. Sometimes it will flip up clear off the barbecue and onto the floor. Make sure your floor is fireproof, as they are hot hot hot.
After doing a few, you obtain a good method. Dont get the wafer chip TOO hot, as at extreme temperatures you will force the silicon dioxide top layer to discolor (brown).
Put each silicon wafer chip somewhere to cool off, and same with the ceramic, remember it will take 20 minutes or more for the ceramic to cool down.
When done and cool, use a new sharp blade to gently linearly scrape the edges of the silicon wafers, removing the silvery/white glue/solder material. (this may be silver solder?) As already mentioned, these wafers are worth more than the gold content of each CPU, and can be easily sold for $5 each. You will find they have a gold back to them, this layer is quite thin. Anyhow, as I said, the chip is worth something, and I would rather sell it with a nice golden back, than trash it to recover the gold on the back.. up to you though.
You will see that under the silicon chip wafer, on the ceramic part, there is a heavy layer of gold, it is usually quite flat, but I've come across some quite lumpy/bumpy ones.
Build a reverse-plating cell. I purchased new battery acid, and heat simmered down lots of 600mls until simmering stopped, and quantity was just under 200mls. I am guessing I had 33% strength, and evaporated the remaining 67% H2O. As you should know, the simmering slows, and stops because the boiling point of 96%+ sulphuric acid is alot higher than with the water. At any stage DONT boil sulphuric acid, noxious fumes are then given off.
I bend all the outer edge CPU pins down and towards the middle row, and then flatten them in a clockwise direction. This helps them make contact with each other, and make the reverse plating operation alittle quicker, as it will hopefully act on all the connected pins. I end up with between two and ten pins left un-deplated.
Also, you must use a sharp blade to quickly scrape some small scratches into the gold that was under the silicon wafer. It must have an insulating layer of something (fluxes?) left over. You want the probe and acid to make contact with the gold plating to be most effective. You dont have to scrape a complete layer off, just enough for the current and acid to begin its process.
I built a custom lowering mechanism to hold the CPU and lower it vertically into the acid, under the pressure of a weighted copper probe. Everything that goes into the acid must be pure copper. Anything else will dissolve, including impurities in the copper.
See attached picture for ideas.
For contacting the pins that dont get deplated with the pointy copper probe, I made a small fine copper wire brush like device (like a copper bristled art paint brush!) so I could use it to probe and make contact with as many pins as possible at once. Probing individual pins one after another is tedious. Remember, this brush/probe also must be connected to the anode of your power source!
The first brush/probe I made was from fine stranded copper wire in an audio cable, but on first use, most of the copper began fizzing and dissolving away. I'm guessing that either it had impurities in the copper, or the strands were to thin to handle the current, and simply melted away. I've since made a better brush/probe with thicker copper, and it works perfectly.
Fire up the cell, and place the pointy probe (not the brush one!) at one side of the gold wafer area. Let the weight of the probe force the CPU under the surface of the acid. Bubbling will commence immediately. NOTE. I have a lamp connected in series with my 24VDC power source to lower the voltage to 12V, and limit the current. While this bright lamp is operating, I sometimes notice a fine mist rising from the bursting bubbles. This is a pure sulphuric acid acid mist, and can settle anywhere in your lab, including, you clothes, and your nasal passages! BE AWARE and careful ! Sulphuric acid disolves cotton fabrics. Also, some fine drops of sulphuric end up on your work bench around the cell, so use a disposable base, and be aware of this acid getting transferred to other items, including your hands, and whatever they may later come into contact with.
I operate my cell on a timer, and can time the operation. It takes 15 to 20 minutes, and peaks at 8 amps to completely remove the gold plating from the CPU. The timer is a must for me, as it gets alittle boring watching the bubbles after an hour or so! It times out by itself, removing power, and allows me to get some other thing done at the same time.
- - I also notice that some pins are completely dissolved in this process ! - - I dont know why.
After a few minutes, and while the CPU is in the cell, use another copper probe (not connected to power) to quickly rub over the gold plated area which was under the wafer. This helps speed up the process of deplating this area. As it must form a new insulating layer. I do this twice if I can, first 5 minutes, and second 5 minutes while in the cell. Also of interest is the miniscule sparks flashing around on that middle flat area as the last of the gold is deplated of the nickel (?) base. Most noticable when in a darkened enviroment. DONT operate your cell in such an enviroment! If you must make an observation of this phenomonen, dont try under complete darkness, you'd be asking for serious spillage troubles. BE CAREFUL!
Once deplated, there is still one area not done. This is where the square ring where the lid was soldered in place. Raise the CPU, and re-place the pointy probe onto the square ring lowering it into acid. You should see a new bubble reaction begin. If no reactions, try placing probe in a new spot. This will deplate the gold UNDER the solder (silver?) I also noted some of the solder is removed in this process as well. Sometimes the square ring comes completely loose, as the thin gold layer in between has all been eroded away into your black gold soup mixture.
Now, what remains on the CPU is a few odd pins which weren't making contact with the anode. Use a brush probe, or a pointed probe to make contact with each, and dissolve the remaining plating. Each pin takes up to 5 seconds.
Carefully remove, and do the next one! This is a long process, which is why I employed the use of a timer, and can do other jobs while cell is running. It requires a bit of attention, and definately no serious distractions while operating this device. Remember your gloves, and eye protection! I've not yet had a bad splash, but dont want one in my eye !
Once the cpu is gold barren, I remove it with another copper gripper widget I made, give it a gentle dip into the sulphuric, to remove some black powder, touch a corner of the chip to the inside of the glass cell to let acid run off, and quickly move to a dunking jar of water, give it a swirl in there to remove more black powder, and I then drop it into a bucket of water, so when I'm done I can give the lot a good swish about removing more black gold powder. A substantial amount gets caught up in the rugged remains of the nickel pins, and when your done you will have quite a bit in the first dunking jar. this can be added to your dilution stage. The last bucket looks like it has lots of odd bit of remains amongst the powder, and I would preprocess this lot to remove the bits that got into the mix from all the shaking.
I've completed all 50 CPU's, and so far, the quantities look very good. After the cell settled overnight, I had about 18mm of black powder settled in the bottom. I know there is some impurities in there, as some of the CPU pins have completely dissolved, some of the solder has dissolved, I have some copper from my first test brush/probe, and I have some lead particles from when the live copper anode probe accidentally touched the lead cathode, and erroded some away into solution. Once all 50 CPU's are done, I will be doing some final cleanup stages of this recovered material, following information obtained on this forum, and will post quantitive and qualitive results! (including what I sold the wafer chips for !!
)
Regards to all.. and feedback most welcome..
SK
I have 50 CPU's, which I am measuring the results of.
There is a substantial amount of Au inside these CPU's, maybe some you are unaware of.
I will hopefully be able to attach some pictures of my processing.
To obtain the most from each my procedure is as follows.
Heat 4 CPU's with pins up on barbeque center plate, using a rotational method so first on - last off. This economises the waiting time for them to heat up.
Using long screwdriver to hold CPU, and sharp knife blade, pry off the gold lid. Put lid aside to cool off.
Carefully pickup the almost red hot CPU WITH pliers/tongs!, and put somewhere to cool off. These are ceramic, and can take awhile to cool down, so be careful.
Put another fresh CPU on the barbie, and do the last one which should be hot enough by now.
Carry on until all done.
I ground an old small router tool bit to make a custom micro chisel-like scraper tool. Use this to scrape off the fine gold wires running around the edge of the silicon wafer chip. Do this over a clean white surface to catch the fine gold wires that may inadvertently fall off. Scrape linearly along the ceramic/gold edge first, and then when they are all cut, run it around the silicon wafer edge. Try and get it close to the welding points, but dont use too much pressure as to damage the glass wafer edge. Try not scratch the wafer at all, as you will end up with bits of silicon in the mix, and also, if the wafer is in good condition, you can sell them for more than the value of the gold content of each CPU! They are sought after by collectors, and people in the microscopics and education industries. The gold wires will bundle up, and get stuck with some still attached to the weld points. Invert the chip, and tap the fine gold wires off onto a clean white surface (a porcelan plate, for example). Check with a loupe or stronger magnification eyeglass that you have as much fine gold wires removed. Complete all the CPU's.
I used the flat clean end of my scraper tool to manually compress all the wires into a cube, and obtained about a 3x3x2mm cubic block of gold from them. I am sure this is PURE gold, no refining may be necesary, but when you melt this block, it may shrink a bit as there must still be some void air space within the block. There is also, you will notic in the zoom pic of the block, that frags of silicon are amongst the wires, so processing in AP/AR might be necessary.
The CPU lids must be cut into quarters or more with tin snips to maximise the area required for the nitric to work on. I placed all the bits in a large beaker, and added nitric acid.
I first tried making nitric using LS's cooling process available here, and from potassium nitrate, with ok results. By ok, I mean that it would be days before even a small etching of the base material would take place. I tried heating the nitric mixture, and ended up accidentally overheating, causing alot of red fumes, and overboiling. It was lucky I had a large beaker, as the fizz almost left the container !
I am sure this nitric would be fine, if made properly. I found I was wasting alot of potassium nitrate, I am not sure about how much white crystal (salts) should be left after the cooling process, but for me there was alot, and maybe something wasn't correct in my procedure. I have since been given some 70% nitric acid, and this only took one day for the foils to float away loose from the lid base material.
I warmed the foils in nitric one last time to free up all the gold, and then tipped off the nitric. Refilled with water, and emptied into a plastic tray, where I quickly manually seperated the gold foils from the lid base material (nickel?) which had only the edges slightly etched.
Once this was done, I discard the base material, and boil the foils in water a couple of times to remove the remaining nitric. I will now put these foils in some hydrochloric acid as a final cleanup, as I can see there is still some very fine copper and nickel colour discoloration on some of them.
I am impressed with the quantity and quality of the gold recovered so far.
Get the barbecue back out again, and begin the same process as you did removing the lids, but we are now going to carefully remove the silicon wafer chips.
While CPU is still in contact with blue flame (must be VERY hot), use a VERY sharp blade and screwdriver to hold the CPU, to pry off the silicon chip. Be gentle, as it is like thin glass, and can chip/break easily. Have a file on hand so you can quickly resharpen the blade, you need a fine egde to get under the wafer and flip it up off the molten glue/solder base. Sometimes it will flip up clear off the barbecue and onto the floor. Make sure your floor is fireproof, as they are hot hot hot.
After doing a few, you obtain a good method. Dont get the wafer chip TOO hot, as at extreme temperatures you will force the silicon dioxide top layer to discolor (brown).
Put each silicon wafer chip somewhere to cool off, and same with the ceramic, remember it will take 20 minutes or more for the ceramic to cool down.
When done and cool, use a new sharp blade to gently linearly scrape the edges of the silicon wafers, removing the silvery/white glue/solder material. (this may be silver solder?) As already mentioned, these wafers are worth more than the gold content of each CPU, and can be easily sold for $5 each. You will find they have a gold back to them, this layer is quite thin. Anyhow, as I said, the chip is worth something, and I would rather sell it with a nice golden back, than trash it to recover the gold on the back.. up to you though.
You will see that under the silicon chip wafer, on the ceramic part, there is a heavy layer of gold, it is usually quite flat, but I've come across some quite lumpy/bumpy ones.
Build a reverse-plating cell. I purchased new battery acid, and heat simmered down lots of 600mls until simmering stopped, and quantity was just under 200mls. I am guessing I had 33% strength, and evaporated the remaining 67% H2O. As you should know, the simmering slows, and stops because the boiling point of 96%+ sulphuric acid is alot higher than with the water. At any stage DONT boil sulphuric acid, noxious fumes are then given off.
I bend all the outer edge CPU pins down and towards the middle row, and then flatten them in a clockwise direction. This helps them make contact with each other, and make the reverse plating operation alittle quicker, as it will hopefully act on all the connected pins. I end up with between two and ten pins left un-deplated.
Also, you must use a sharp blade to quickly scrape some small scratches into the gold that was under the silicon wafer. It must have an insulating layer of something (fluxes?) left over. You want the probe and acid to make contact with the gold plating to be most effective. You dont have to scrape a complete layer off, just enough for the current and acid to begin its process.
I built a custom lowering mechanism to hold the CPU and lower it vertically into the acid, under the pressure of a weighted copper probe. Everything that goes into the acid must be pure copper. Anything else will dissolve, including impurities in the copper.
See attached picture for ideas.
For contacting the pins that dont get deplated with the pointy copper probe, I made a small fine copper wire brush like device (like a copper bristled art paint brush!) so I could use it to probe and make contact with as many pins as possible at once. Probing individual pins one after another is tedious. Remember, this brush/probe also must be connected to the anode of your power source!
The first brush/probe I made was from fine stranded copper wire in an audio cable, but on first use, most of the copper began fizzing and dissolving away. I'm guessing that either it had impurities in the copper, or the strands were to thin to handle the current, and simply melted away. I've since made a better brush/probe with thicker copper, and it works perfectly.
Fire up the cell, and place the pointy probe (not the brush one!) at one side of the gold wafer area. Let the weight of the probe force the CPU under the surface of the acid. Bubbling will commence immediately. NOTE. I have a lamp connected in series with my 24VDC power source to lower the voltage to 12V, and limit the current. While this bright lamp is operating, I sometimes notice a fine mist rising from the bursting bubbles. This is a pure sulphuric acid acid mist, and can settle anywhere in your lab, including, you clothes, and your nasal passages! BE AWARE and careful ! Sulphuric acid disolves cotton fabrics. Also, some fine drops of sulphuric end up on your work bench around the cell, so use a disposable base, and be aware of this acid getting transferred to other items, including your hands, and whatever they may later come into contact with.
I operate my cell on a timer, and can time the operation. It takes 15 to 20 minutes, and peaks at 8 amps to completely remove the gold plating from the CPU. The timer is a must for me, as it gets alittle boring watching the bubbles after an hour or so! It times out by itself, removing power, and allows me to get some other thing done at the same time.
- - I also notice that some pins are completely dissolved in this process ! - - I dont know why.
After a few minutes, and while the CPU is in the cell, use another copper probe (not connected to power) to quickly rub over the gold plated area which was under the wafer. This helps speed up the process of deplating this area. As it must form a new insulating layer. I do this twice if I can, first 5 minutes, and second 5 minutes while in the cell. Also of interest is the miniscule sparks flashing around on that middle flat area as the last of the gold is deplated of the nickel (?) base. Most noticable when in a darkened enviroment. DONT operate your cell in such an enviroment! If you must make an observation of this phenomonen, dont try under complete darkness, you'd be asking for serious spillage troubles. BE CAREFUL!
Once deplated, there is still one area not done. This is where the square ring where the lid was soldered in place. Raise the CPU, and re-place the pointy probe onto the square ring lowering it into acid. You should see a new bubble reaction begin. If no reactions, try placing probe in a new spot. This will deplate the gold UNDER the solder (silver?) I also noted some of the solder is removed in this process as well. Sometimes the square ring comes completely loose, as the thin gold layer in between has all been eroded away into your black gold soup mixture.
Now, what remains on the CPU is a few odd pins which weren't making contact with the anode. Use a brush probe, or a pointed probe to make contact with each, and dissolve the remaining plating. Each pin takes up to 5 seconds.
Carefully remove, and do the next one! This is a long process, which is why I employed the use of a timer, and can do other jobs while cell is running. It requires a bit of attention, and definately no serious distractions while operating this device. Remember your gloves, and eye protection! I've not yet had a bad splash, but dont want one in my eye !
Once the cpu is gold barren, I remove it with another copper gripper widget I made, give it a gentle dip into the sulphuric, to remove some black powder, touch a corner of the chip to the inside of the glass cell to let acid run off, and quickly move to a dunking jar of water, give it a swirl in there to remove more black powder, and I then drop it into a bucket of water, so when I'm done I can give the lot a good swish about removing more black gold powder. A substantial amount gets caught up in the rugged remains of the nickel pins, and when your done you will have quite a bit in the first dunking jar. this can be added to your dilution stage. The last bucket looks like it has lots of odd bit of remains amongst the powder, and I would preprocess this lot to remove the bits that got into the mix from all the shaking.
I've completed all 50 CPU's, and so far, the quantities look very good. After the cell settled overnight, I had about 18mm of black powder settled in the bottom. I know there is some impurities in there, as some of the CPU pins have completely dissolved, some of the solder has dissolved, I have some copper from my first test brush/probe, and I have some lead particles from when the live copper anode probe accidentally touched the lead cathode, and erroded some away into solution. Once all 50 CPU's are done, I will be doing some final cleanup stages of this recovered material, following information obtained on this forum, and will post quantitive and qualitive results! (including what I sold the wafer chips for !!
)Regards to all.. and feedback most welcome..
SK
