Refining Pentium Pro with AR

[g_axelsson]

I just tested one of my ceramic Pentium Pros with a magnet. The pins are magnetic. The heat spreader is not.

Second that, checked 81 ceramic PPro and all had magnetic legs. So did all the 386, 486, MMX, K6, i960, pentium and P3 processors I tested too. Kovar seems to be the rule... now, can anyone name a ceramic or fiber CPU with non kovar pins? I'm going to check the stash of 3000 CPU:s that I have...

... back now. I checked 286, 386, 387, 486, pentium, and overdrives made by Intel, IBM, AMD, Cyrix, Fujitsu, ST, SGS Thomson and Siemens and I only found one non-magnetic. IBM 286 with aluminum can was made with non-magnetic legs, but that's a rare chip. Mostly found in IBM PS/2 systems from the early 90:es.

For CPU:s packaged in plastic capsule, DIL and surface mounted ones with lead frame it seems that both magnetic and non-magnetic alloys are common.

Göran

 

[Anonymous]

Goran I refer you to my post above yours mate.

Jon

Edit: If as you say 99% of all ceramic pins are based upon Kovar then of course I'm not going to disagree with the facts that you've found. If that is the case then logically the recommendation to HCl wash these Pentium Pro chips would apply to all the ceramics we process. We know it isn't necessary in other ceramics so therefore it isn't necessary here. I rest my case gentlemen unless someone can give me a reasoned and logical explanation otherwise.

 

[g_axelsson]

Goran I refer you to my post above yours mate.

Jon

I saw that and have no more comments than we all prefer different styles of refining.

If you prefer to drop your CPU:s in AR and extract the gold from a solution heavy in base metals, it's ok and it works. Some of our members are doing that and it works for them.

If you prefer to remove as much of the base metals first with either nitric or hydrochloric acid it's okay. It adds a step and takes longer time, but the final drop is from a solution not as heavily in base metals and that could be important for some of our members.
As most scrap we refine has been used the surface of the plating is probably damaged so the acid can get to it even when fully plated. When it starts to dissolve the foils easily breaks. Some pins might survive but a lot base metal will be removed.
This process should work and I'm certain that some of our members are using it just fine.

My favorite is what I call modified AR, add AR to the CPU:s but don't use enough to finish the job. Then in the end the gold will cement out on the base metals and you can pour off the barren liquid with most of the base metals. Add more AR (and possibly also more CPU:s) while removing any ceramics where all the metals are gone from, it just takes up space. In the end you have a quite concentrated gold solution with less base metals.
This method takes a bit of experience and longer in time but suits me well. It's not a process I would recommend to a newbie.
This process is used by some of our members and it works great for them.

I would like to add a disclaimer to this. I'm not that experienced in refining CPU:s but I'm gearing up for the 80 kilo CPU lot I bought this summer. All I've made so far is started a 1 kilo test run of black MMX fiber CPU:s to assay how much gold there is in them. For these I decided to boil them in HCl to remove the legs, then I'm going to incinerate to get to the bond wires.

We all have different styles that suits our self, so chill out, there is more than one way to refine gold.

Göran

 

[Metaphore]

What I'm getting is that it would be best to separate Pentium Pros into 4 components before processing:

• Heat spreader
• Silicone chip
• Pins
• Ceramic base

Silicone chips and pins processed in AR.
Ceramic base crushed and processed in AR.
Heat spreader - I'm not too clear on. Is it nitric?

 

[Anonymous]

Goran I agree with you. I don't need to chill out because I am completely chilled, I can promise you. Maybe my post didn't read that was because of our language differences but it was meant to be succint not rude.

My whole diatribe was based upon the advice to "boil in HCl to remove the iron." That advice was incorrect, and not just a "difference in refining methods." You can boil a gold plated item in HCl if you like but I would love to see how the HCl is going to reach the iron substrate?

 

[Anonymous]

What I'm getting is that it would be best to separate Pentium Pros into 4 components before processing:

• Heat spreader
• Silicone chip
• Pins
• Ceramic base

Silicone chips and pins processed in AR.
Ceramic base crushed and processed in AR.
Heat spreader - I'm not too clear on. Is it nitric?

No. This whole thread has set you on the wrong path my friend. My initial advice is correct. Remove the black base, break up the processors with a hammer and process in AR complete. I am talking from hands on experience here and have lost count of the amount of these processors I have done successfully.

Regards

Jon

 

[Metaphore]

No. This whole thread has set you on the wrong path my friend. My initial advice is correct. Remove the black base, break up the processors with a hammer and process in AR complete. I am talking from hands on experience here and have lost count of the amount of these processors I have done successfully.
Jon

But that is exactly what I did and had the never ending reaction going. Poured off the first AR and had the same happened with the second. Something in my process needs to improve, no?

 

[Anonymous]

Yes I agree. I questioned the advice that you were given initially and nobody who had offered that advice came back and altered theirs based upon the information you gave. Ok then here goes:

Take your 2 litre beaker. Place 5 prepared processors into it carefully (prepare as I said previously.)

Top the beak up to 1 litre volume with your 36% HCl and put on heat. Warm it thoroughly through. Don't boil it, just get it warm for now.
Add 5ml of Nitric and leave it. The process will start a lot faster because your HCl is hot.
Once your Nitric is working you'll see the solution darken, fizz, bubble and do all the usual things. when it stops add another 5ml. Rinse and repeat until the metals have disappeared. don't be afraid of your last 5ml nitric addition requiring denoxing. You want to make sure the whole process is complete.

Once you have seen the whole 5 digest properly you will be able to reduce your initial quantity of HCl for the next batch OR use the same solution for more processors but that takes a few more steps along the learning curve. I have suggested this volume to allow you to see the process through from start to finish with no problems.

The breaking up of the processors is essential as it give the acids a better foothold on both the lids and the gold plate under the die. 6 to 8 pieces each should be enough though, you don't need to go too mad with the hammer. 8) 8)

I'm assuming that you're following all relevant safety procedures. Enjoy.

Jon

 

[g_axelsson]

Goran I agree with you. I don't need to chill out because I am completely chilled, I can promise you. Maybe my post didn't read that was because of our language differences but it was meant to be succint not rude.

My whole diatribe was based upon the advice to "boil in HCl to remove the iron." That advice was incorrect, and not just a "difference in refining methods." You can boil a gold plated item in HCl if you like but I would love to see how the HCl is going to reach the iron substrate?

I'm sorry too, I should have added a few smileys to my post. It was more in a friendly jokingly tone I added the "chill out" and it wasn't just directed to you, more to the thread generally before people start taking too firm positions.

I'm looking forward to kadriver's video as it will show yet another persons way of solving how to refine ceramic CPU:s. If he runs into any issues along the way I know he's proficient and systematic enough to solve any problems along the way.
I've just looked through Sam's two part video of ceramic CPU refining. I'm missing any mentioning of denoxing and most newbies will use too much nitric. He also used what seems to be a metallic spoon to mix the SMB in the gold chloride, something I would never do out of principle. Neither does he mention what he is doing with the gold plated lids. But it works for him and that's the direct AR method. (He actually uses poor mans AR)

When time allows me I'm going to test boiling a PPro in HCl just to see how hard it is to break through the plating. Then I can answer your question. :mrgreen:
Right now, winter is coming and my refining season is mostly over, but I'm looking to purchase a small house in the countryside just for storage and refining space.

Still friends? 8)

Göran

 

[Anonymous]

Hehe always friends Goran.

Disagreeing isn't a declaration of war mate 8)

Jon

 

[Barren Realms 007]

Pre treatment of ceramic CPU's in HCL is really not a necessary step that needs to be done. It serves no use in refining them. Mostly all it does is create extra waste solution to deal with and add more cost and time to processing them.

 

[kadriver]

Ive never broke these up before. Is there a certain way to do it or do I just put them on the concrete and hit with a hammer?

I have zero experience with these CPUs, I held one in my hand for the first time today, they are much heavier than I thought they would be.

Ive got 12 CPUs on the way. I'll make one video just like spaceships recommends with 6 of the CPUs since he has experience with these.

Then I'll do the other 6 using HCl treatments before the AR is added. This way we can see if there is a difference.

kadriver

 

[Anonymous]

Kadriver

I'll break one up tomorrow and post a pic up for you with pleasure 8)

Jon

 

[Anonymous]

Broken with a hammer. You can break it further if you wish but this was a quick "demo" pic hence the mess.

Take note of the heat spreader being in pieces, and the gold plating under the die being visible as some of the die has broken off. That's where you want to be before processing. Sorry about the mess around it!

Jon

 

[g_axelsson]

This two part video is missing denoxing, but except for that it's quite good and shows the process.

https://www.youtube.com/watch?v=tQ_R8Gr26kk
https://www.youtube.com/watch?v=Hv4CcaTVJqY

It shows what spaceships is talking about. As PPro:s are almost twice as big, 6-8 pieces should be fine, but that's not a critical number.

As for your original problems, what did you do with the heat spreader? Did you try to dissolve it? It can take a lot of acid and it's hard to dissolve thanks to the tungsten. The heatspreader is made from a tungsten copper alloy. It might also be able to cement gold out of solution. As I haven't refined any PPro yet I don't have any recommendations here.

Shameless plug : I'll sell PPro:s to anyone that want's to test refining for $15/+postage. It's more than the gold is worth but compensates me for the extra work of packing and shipping.
Edit : All sold.

Göran

 

[kadriver]

Got it Jon, thanks for that tip. I'll break them up just like you showed.

Metaphor, I just thought of something with respect to the process you used.

You added HCl and small doses of nitric.

After heating for several hours, and after several doses of nitric, and after there are no more fumes being produced, you could check the solution with stannous to make sure that you have no gold in solution.

Then siphon off the acids and put in fresh HCl and more nitric doses. This should prevent the build up of base metals that you experienced.

Spaceships and barren realms both say that the AR approach should work fine. But you had problems with other substances in your batch. This is what I would like to avoid when I do my videos.

I looked up solubility of tungsten and the site said it is soluble in a mixture of nitric and hydroflouric acids. But it is barely soluble in hydrochloric acid alone.

It also said that the tungsten has a YELLOW component, much like the yellow substance that you formed in your reaction;



TUNGSTEN [[[symbol]] W, atomic weight 184 o (0=16)], a metallic chemical element found in the minerals wolfram, an iron and manganese tungstate, scheelite, a calcium tungstate, stolzite, a lead tungstate, and in some rarer minerals... It is soluble in a mixture of nitric and hydrofluoric acids, and the powdered metal, in aqua regia, but slowly attacked by sulphuric, hydrochloric and hydrofluoric acids separately; it is also soluble in boiling potash solution, giving a tunstate and hydrogen.

Tungsten dioxide, W02, formed on reducing the trioxide by hydrogen at a red heat or a mixture of the trioxide and hydrochloric acid with zinc, or by decomposing the tetrachloride with water, is a brown strongly pyrophoric powder, which must be cooled in hydrogen before being brought into contact with air. It is slightly soluble in hydrochloric and sulphuric acids, giving purple solutions. It dissolves in potash, giving potassium tungstate and hydrogen, and is readily oxidized to the trioxide.

tungsten trioxide, W0 31 occurs in nature as wolframine, a yellow mineral found in Cumberland, Limoges, Connecticut and in North Carolina. It is a canary-yellow powder, which becomes a dark orange on heating; the original colour is regained on cooling.



This may be an explanation for the yellow material in your beaker.

kadriver

 

[kadriver]

Goran;

Those are Samuels videos. He does an excellent job. I've watched them several times. Thanks for referring as I'll review them before I do my video.

I'll be doing the same thing with the Pentium Pro CPUs.

Looking forward to starting this project.

kadriver

 

[g_axelsson]

The question is what to do with the heat spreader? Since it's made up of a copper wolfram powder sintered alloy it's a bit difficult to process. They doesn't behave as normal metal, for example you can shatter it with a hammer as shown above. I know Lazersteve removed them with a torch and processed them separately.

Use this search to find quite a number of relevant postings about the problem above.
http://goldrefiningforum.com/phpBB3/search.php?keywords=pentium+pro+heat+spreader&terms=all&author=Lazersteve&sv=0&sc=1&sf=all&sk=t&sd=d&sr=posts&st=0&ch=300&t=0&submit=Search

I'm going to read them now...

Göran

 

[kadriver]

Excellent reference Goran.

Here is some relevant info on separating the gold foils from the tungsten on the CPUs;




glondor wrote:
If the heat sink breaks, you do not want to do them directly in A/R. That is a sintered tungsten/ copper heatsink

Samuel-a wrote:
Actually, the opposite is true.
W bearly reacts with AR, the only difference is that it will later precipitate (when cold) a small amount of yellow WO3 that is easly filtered out.

Lazersteve wrote:
Glondor is 100% correct here, putting the tungsten topped cpus directly in AR is a bad decision for several reasons:

1) The yellow powder (oxidized tungsten) requires lots of rinsing to remove the gold solution from it's structure,
2) The yellow powder's slight acid solubility allows it to make it through several stages of filtration,
3) Soaking in nitric acid first allows the tungsten plates easily to be collected in tact without breaking them up,
4) The yellow powder will give false positives to stannous chloride as a cornflower blue color,
5) In effect, if you do not remove the tungsten heat spreaders first, you are putting unnecessary garbage into your AR reaction (GIGO).
6) The yellow powder tends to stick to buckets and beakers leaving 'ring around the beaker' if it is not removed before AR treatment.
7) Unnecessarily making the yellow powder by going straight to AR with gold topped ceramic cpus adds another waste stream (the yellow powder) to your processes.

Steve


In Steve's post he removes the tungsten caps by deforming it with a hammer blow - he does not break it into pieces. Then he separates the gold plated cap and removes any pieces of ceramic that cling. The hand-cleaned gold plated caps are then soaked in dilute cold nitric acid. The foils came off nicely leaving the tungsten intact and easily retrievable before going to AR to dissolve the foils.

This is an older post and the method Steve used may have changed. Spaceships and Barren Realms both say their experience with AR right off the bat works fine.

Metaphor, this may explain how you got that yellow substance by adding AR to the CPUs. I truely want to avoid the formation of this yellow tungsten powder.

Here is the link to that post by Lazersteve; http://goldrefiningforum.com/phpBB3/viewtopic.php?f=60&t=18680&p=188873&hilit=pentium+pro+heat+spreader#p188873

kadriver

 

[Metaphore]

My purpose is not to save time or money, I'm trying to learn as much about this as I can. I believe separating the different materials and running them separately will aid in the learning process even if it lowers the yield.

One thing I'm still somewhat hazy on - what is the best way to remove the lid from a PPro? Will a map gas torch do it?