Processing MLCC's

[9kuuby9]

I think we need to talk straight preferably with photos or manufacturer marks(names), model numbers instead of things like High end MLCC's, high yield, mil specs, super high yield, and similar constructions.
What one consider super high yield or high end may not be the case or it is considered as such only by that person and all this only lead others to confusion what we talk about.

Sam asked me what I'm going to process by melting. And I replied High end MLCC's

The "average yield" that I have presented above are for common Pd containing MLCC's.

The Yields are discussed my times in the past.

 

[patnor1011]

How do you define common, average and high yield may be clear to you yet not many other people will know what is being discussed here.
Please do not be offended, I talk about it only because the same happens with cpu processing where we do have many types and they do have different yield. The same apply for what you discuss here.
Can you give us some examples of what you call high end MLCC's?

 

[9kuuby9]

How do you define common, average and high yield may be clear to you yet not many other people will know what is being discussed here.
Please do not be offended, I talk about it only because the same happens with cpu processing where we do have many types and they do have different yield. The same apply for what you discuss here.
Can you give us some examples of what you call high end MLCC's?

That's True, To quote myself as I did many times;

Any mid-grade to high grade components are used in quality products. For instance a Computer Motherboard. Low-grade components (mostly nickel and tin) are used in poor quality products or those that do not need very good quality components, such as a computer power supply.

To determine the grade of the component is to look at what the particular scrap is used for and what it's purpose is in circuitry.

And again;

Let me also enlighten everyone that their are a lot of different types and sometimes with completely different yields, this is said to try prevent unnecessary posts on this topic.

What I process are High-end MLLC's, How do I know for sure? Because I have a supplier and has some datasheets on it.

You can not really Compare small components such as MLCC's and SMT's to Processors. Because you can not really look up the MLCC's. Their is no serial number present or anything else.
While CPU's can be assorted by Type, For instance Pentium pro's. In Pentium Pro's itself you have different yields. Because they are produced by different manufacturers. So for very small components, Like MLCC's You will have to take a guess to determine if it's low, mid or high-end. (look at the quotes above)

And by processing or assaying it, you will have a definitive yield. and thus the ability to say if it's low, mid or high-end. How do you know it's low, mid or high-end? You collect the yield data of MLCC's and compare your assay data to it. then you are able to rank it as in low, mid or high-end.



And if you try to chip off MLCC's from boards, as many do to acquire MLCC's. Then you will have a mix of MLCC's that could have many yields. By using a rare earth magnet, you'll have the ability to separate the nickel based ones from the Palladium based ones.

I hope this somehow helps explain it.

 

[niteliteone]

SMD resistors Have a similar yield results as the ones I presented above.

Can you (or anyone) substantiate that claim?

I've gone out of my way to only remove MLCCs rather than all the SMDs on the boards I've stripped down. If the SM resistors have similar yields, WOW have I wasted some time with that.

http://www.venkel.com/docs/part-pdfs/resistors-mini-catalog-venkel.pdf

Page 6,-General Purpose Resistors- under "Termination" you have 2 sorts, Namely Tin over Nickel and Silver over Palladium.

Let me make it clear that they have similar yields to MLCC's but still have a general lower yield. Because MLLC's have a capacitance, therefore Electrodes. So the outer termination and Electrodes contain PM's. While Resistors only have the outer termination that contains PM's. Having at it's core a substance that offers resistance and so PM's are not fit for that role, because PM's have a good conductivity with a very small amount of resistance. Which will not act as a resistor given it's primary role.

Let me also enlighten everyone that their are a lot of different types and sometimes with completely different yields, this is said to try prevent unnecessary posts on this topic.

Are you going to use the General Purpose type from page 6 as you have indicated, or are you going to use the specialty ones from page 3 that contain Palladium, unlike the general purpose ones from page 6 :?:
This is an important question as most of what every one here will ever see will be the common general purpose ones and not the specialty ones. Most of what we do comes from Asia, where every thing is built as cheaply as possible :shock:

 

[9kuuby9]

I do not know if some of you really read what I write? or just the portions you like?

When I speak about the yield of MLCC's or resistors, I always say "the ones that do contain Pd" Not all MLCC's or Resistors.

for example;

SMD resistors Have a similar yield results as the ones I presented above.
You also have SMD array's which have sometimes more yields, because they have more layers.
However take notice that not All MLCC's contain Pd Sometimes they only contain Ni & Ag
If you have mixed MLCC's the yield may vary for the Pd (one half has 2% yield and other half has 0% yield which results in a total 1%)
So when we speak about "Pd 1.5 - 3%" this is only an average yield number for the MLCC's that DO contain Pd.
When you acquire MLCC's try to get them as per kind and not mixed. So when you assay them per type u have a stable yield.

I even give examples and explanations, yet some don't seem to understand?

I somehow got tired from continuously pinpointing what I said.

 

[niteliteone]

I (we ? ) understand pretty much every thing you have posted. Just don't understand why questions and feed back are not met very well. :shock:

What " I " don't understand is, if this test deals with devices we see on a daily basis, this thread could give some good information. But if it will use parts we may see only once or twice in our life times, what real good is it in the scheme of providing actual usable data we can use on a daily basis.

Most here deal with cheap Asian electronics, not NASA's latest toys, so any parts or processes that addresses this common material is beneficial, whereas data from processing "once in a lifetime material" does nothing for the average person processing common scrap.

That is, other than the "smelting" process being tried instead of the wet chemical process. I see this as being of benefit as an alternative method, as well as using common SMD's that we all deal with, would be of benefit too.

I am interested in this project and I see it very beneficial. Especially if is uses devices we all see on a daily basis.

 

[9kuuby9]

Well I'm glad to see that this thread is somewhat being useful.

Also note that I depicted a computer mother board as mid to high-end and not some NASA or satellite equipment. :lol:

See for yourself and Just get your hands on an old or new motherboard, Chip and collect all the SMD's. And do a Chemical leach, test with stannous and the brown spot appears which changes into a green spot. :mrgreen:

Anyway we somehow reached a common ground. 8)

 

[9kuuby9]

Considering the high percentage metals...
15% Kryolite
35% Borax (anhydrous)
25% sodium nitrate
12 1/2% soda ash
12 1/2% slaked lime

This would be a Good flux recipe I guess? The material I'm dealing with is around 10% composed of metal and the rest is a mineral oxide( Alumina, Zicronia, etc...)

So A flux ratio of 1 to 2 (roughly 30% by weight) would be a good ratio for a lower metal percentage yielding ceramics.

Would be a temperature of 1200°C be sufficient to melt this mix? and how many hours does it need for it to have melted the metals and be in a fluid state?

The ceramics are also crushed to a fine powder instead of it's whole form. Would be adding 100g of Silver to this mix serve well as a collector and a means of lowering the overall temperature?

Because I don't really know at what temperature this flux mixture melts, and also the lack of knowledge in this field. Is their any elaborated literature concerning the use of fluxes and metal collectors?
Would 4metals, GSP or any experienced member in this field chime in on this matter?

Any help is kindly and greatly appreciated! :mrgreen:

Edit* I also forgot to ask if these fluxes release harmful gases when melted? I'm using an induction furnace, like those used in a fire assay but then a bigger version.

 

[steyr223]

Niteliteone said

Most here deal with cheap Asian electronics, not NASA's latest toys, so any parts or processes that addresses this common material is beneficial, whereas data from processing "once in a lifetime material" does nothing for the average person processing common scrap.

I disagree
10 years from now nasa toys will most likely be our everyday material
(Or sooner depending on how long it takes the collapse of our wonderfull government to finally happen)

Just my 2 cents i will but out now
Steyr223 rob

 

[9kuuby9]

Considering the high percentage metals...
15% Kryolite
35% Borax (anhydrous)
25% sodium nitrate
12 1/2% soda ash
12 1/2% slaked lime

Is it possible to use other chemicals in this flux recipse?

It's main purpose would be the bring the temperature down if I'm not mistaken? and cryolite to get the ceramics fluid?

For instance can Potassium nitrate nitrate be used instead of sodium nitrate?

Thanks in advance!

 

[Lou]

He recommended a flux that is basic and will attack silica/alumina and slag off.

I'm not a huge fan of nitre with copper based melts.

You can use KNO3 in lieu of NaNO3 (although the latter is cheaper).

 

[9kuuby9]

Thanks Lou!

I'm using silver here as a collector instead of copper. I have only PGM's and silver in the mix and no base metals. I will play with the ratios and see what's the best ratio. I'll try to add silver until I have a ratio of 1 Pgm to 10 or 12 silver.

I will try both out, and try to compare the results.

The thing is that here In Europe their are things that are regulated or viewed as dangerous. So even if it's in general a cheap material is gets really expensive due to the regulations and it's is sometimes completely not sold anywhere.


I'm soon going to try this out, probably in the middle of next week when I get my supplies in.
Just got my door type furnace Which reaches up to 1300°C 8)

The crucible used with those harsh flux, is a clay graphite based one called "Salamander"

I hope to see some more communal help so I can try many possibilities, Because you can not deal as an individual with such variety of materials. :mrgreen:

These threads contain a lot of valuble information when it comes to melting ceramics and fluxes.

-- http://www.goldrefiningforum.com/phpBB3/viewtopic.php?f=49&t=11490#p112545
- http://goldrefiningforum.com/~goldrefi/phpBB3/viewtopic.php?f=60&t=5871#p50658

 

[Lou]

I'd recommend silicon carbide over clay-bonded graphite for that flux mixture.

 

[9kuuby9]

Thank you for the advice Lou.

I'll do some research And See if I can get some over here in Europe.

 

[Palladium]

Maybe i missed something here. Why not use chemical extraction here? I've run them things through ar attached to cpu chips and they come out stripped with nothing left but the ceramic substrate. Why ball mill? If not ar because of the silver content then why not nitric to attack any palladium or silver and recover the rest in the residues? Why wouldn't chemical extraction work? If your smelting it anyway all you doing is losing some of the metals in the flux and two you are adding more metals as a collector which has to be refined out anyway so your not saving any steps just adding more waste, chemicals, and hassle to the situation. Maybe i'm wrong. It's not my area of expertize.

 

[Lou]

Might be ok on small quantities, but if you do 1000s of pounds of them, best to grind, sample, smelt, and then electrowin the copper and go after the values as slimes (much less waste).

That, and these are high in ceramics that are pretty hard on most mills.


Lou

 

[Palladium]

True!
He had only gotten to the melting part with a collector. I didn't know who much and on what volume he was trying to process. I wasn't sure if his intentions were to refine the bar by chemical method to recover the precious metals and the other metals become part of the waste solution as a byproduct or if he we recovering the metals from the solution (copper electrolysis) in which case the precious metals would become the waste and be recovered for further refining. Sometimes you go after the precious metals, sometimes you leave them alone.

 

[DuckTheQucker]

How do you define common, average and high yield may be clear to you yet not many other people will know what is being discussed here.
Please do not be offended, I talk about it only because the same happens with cpu processing where we do have many types and they do have different yield. The same apply for what you discuss here.
Can you give us some examples of what you call high end MLCC's?

That's True, To quote myself as I did many times;

Any mid-grade to high grade components are used in quality products. For instance a Computer Motherboard. Low-grade components (mostly nickel and tin) are used in poor quality products or those that do not need very good quality components, such as a computer power supply.

To determine the grade of the component is to look at what the particular scrap is used for and what it's purpose is in circuitry.

And again;

Let me also enlighten everyone that their are a lot of different types and sometimes with completely different yields, this is said to try prevent unnecessary posts on this topic.

What I process are High-end MLLC's, How do I know for sure? Because I have a supplier and has some datasheets on it.

You can not really Compare small components such as MLCC's and SMT's to Processors. Because you can not really look up the MLCC's. Their is no serial number present or anything else.
While CPU's can be assorted by Type, For instance Pentium pro's. In Pentium Pro's itself you have different yields. Because they are produced by different manufacturers. So for very small components, Like MLCC's You will have to take a guess to determine if it's low, mid or high-end. (look at the quotes above)

And by processing or assaying it, you will have a definitive yield. and thus the ability to say if it's low, mid or high-end. How do you know it's low, mid or high-end? You collect the yield data of MLCC's and compare your assay data to it. then you are able to rank it as in low, mid or high-end.



And if you try to chip off MLCC's from boards, as many do to acquire MLCC's. Then you will have a mix of MLCC's that could have many yields. By using a rare earth magnet, you'll have the ability to separate the nickel based ones from the Palladium based ones.

I hope this somehow helps explain it.

My brown MLCC's are magnetic to earth magnet, but not strongly. They come from low yield mainboards with fiber cpu's on it. Do they contain palladium?
Black ones very magnetic.

 

[goldsilverpro]

I'd recommend silicon carbide over clay-bonded graphite for that flux mixture.

I agree with Lou. The SiC is more expensive but it lasts longer and you can put about anything in it. Clay/graphite seems a tad cleaner to work with.

The only thing I preferred clay/graphite for was when I was melting silver crystal with no flux and pouring 10/100oz bars. At first, I melted about 300-400 oz. The same amount each time. After a few melts, I noticed that a fairly deep grooved ring was being cut into the crucible at the melt surface. I then started varying the amount of silver per melt and the crucible eroded more evenly. Melts, as acid solutions, are most corrosive at their top surfaces. There's more oxygen there.

 

[Sancho_n_Pedro]

Sorry to revive an old post, but I was interested here.

MLCC's are commonplace on most boards and you get both magnetic and non magnetic ones. If I were mainly going for the silver in these, and as a by-product the Pt and Pd, would it matter if I mixed them both together?

I'm sourcing the MLCC from Hard Drive boards, satellite/router boards (well any boards to be fair).

I was really hoping to get to some kind of average ratio where we could say you can expect about x% of each material, as this gives a better idea on what will be retrieved.

Also, as pointed out, is the wet method the better method that the OP mentioned, for instance if you have less than 1Kg of MLCC?

Maybe i missed something here. Why not use chemical extraction here? I've run them things through ar attached to cpu chips and they come out stripped with nothing left but the ceramic substrate. Why ball mill? If not ar because of the silver content then why not nitric to attack any palladium or silver and recover the rest in the residues? Why wouldn't chemical extraction work? If your smelting it anyway all you doing is losing some of the metals in the flux and two you are adding more metals as a collector which has to be refined out anyway so your not saving any steps just adding more waste, chemicals, and hassle to the situation. Maybe i'm wrong. It's not my area of expertize.