Chip value estimation

Gold Refining Forum

Help Support Gold Refining Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
I think the confusion is converting troy ounces, but starting out with regular pounds and grams and not troy pounds. I am not sure if I am doing this right.

700 pounds x 453.44g = 317,408g / 1000 = 317.40g net weight. 317.40 / 31.1 = 10.2 oz troy.

Or more simply put: 700 lb = 10208.333 oz(troy) divided by 1000
 
I took 700 pounds X 16 Ounces per pound = 11,200. Then took 11,200 ounces X 28.39 grams per ounce for a total of 317,968 grams. Then I divided that by 1000 for a total of 317.968, then I divided that by 28.39 grams per ounce for a total of 11.2 ounces. :|

Edit: You're right, if I divide by troy ounces it comes to 10.2
 
Nope you guys are right.I was doing ounces to pounds,not grams to ounces......don't forget it is still monday.
So yes if someone had 700 pounds of these,and they yielded 1 gram of AU per kilo,then it would yield 10.2 ounces TR.
 
OK - so if 10.2 ozT per 700 lb was a good estimate - that would make them worth around $22.46 / lb for gold alone, at today's spot. I've crushed a bunch of this kind of chip package up, trying to get an idea of PM value. I would guess the value to be considerably less than this - but I haven't actually processed any of them - so it's just a gut feeling I get picking the gold bond wire out of the rubble using a microscope. Remember that the copper in the mix makes up a good proportion of the weight - copper has a density 13% greater than iron.

samuel-a said:
I ball park the black IC's at <0.001% , gold wise.

The package material is not composed from ceramics as mic mentioned. calling it plastic might not be very accurate term as well (though much more relevant term).
The molding materail composed mainly from a resin of sort (organic material of course) that act as binder and fused silica powder cured togather with high heat.

There are other element (either organic or non) added to the molding material to effect the physical behavior of the package. This info is not really shared in full by manufacturers though.

Edit:
If you have a very large amount and relatively homogenous, you'd be better off pay to assay them before selling them.

I would say your description is a very accurate one for - plastic. Most structural plastics, or ones that have to withstand thermal cycling - as a chip does - are reinforced with ceramic binders such as tabular alumina or some form of silica, as cab-o-sil, microspheres, glass fiber, etc.

With so many pounds of the material - your advice to assay is a sound one. Otherwise, at any price he sells at, he'll forever wonder if he ripped himself off. With that kind of volume, it's likely that there's more coming from the same supply stream. An assay may very well pay for itself in the first batch sold.
 
I ball park the black IC's at <0.001% , gold wise.

Not to beat a dead horse but, to eliminate confusion, 0.1% = 0.001, as Pat said, whereas 0.001% = 0.00001. Therefore, at .001%, they would only contain 454 x .00001 = .0045 g of Au/pound = $0.22/pound, based on a $1545 market. At the actual 0.1%, the value is $22/pound

what i meant to say my ball park numbers are: per 1000 grams of chips, there's less or equal to 1 gram of gold.

That would be 1 part per thousand = 0.1% or, expressed as a decimal, 0.001. To convert the decimal to a percent, you always move the decimal point 2 places to the right (or, multiply by 100). To convert percent to a decimal, move it 2 places to the left (or, divide by 100).




To make math easier, I would highly recommend memorizing as many relevant conversion factors as possible or, keep a list handy and keep adding to it. Of course, this must be done with understanding. Personally, I hate those online plug-in type calculators and rarely use them unless the math is extremely involved and I don't want to take the time to figure it out. Why? Because I don't learn anything by using them. Also, on some of them, it's easy to get a wrong answer if you don't know exactly what values to plug in where. When I grind out the math, I can truly visualize what is really happening. Also, I often try to find at least one other way to do the math to double check my answers, as in the 2 examples immediately below. Also, think about your answer. I've learned that, if the answer doesn't sound right, it usually isn't right. Remember that there are many different ways to approach each problem and still get the same correct answer. Probably, the more correct ways you can think of, the better you understand what you are doing. On this problem alone, we, as a group, have come up with about half a dozen good, workable methods.

One of the convenient conversions I've always remembered is that an avoirdupois pound (a standard US pound) contains 14.583 troy oz. Therefore, 700# @ 0.1% = 700 x 14.583 x .001 = 10.2 tr oz.

Another conversion permanently implanted in my mind is that a 2000 avoirdupois pound ton contains 29,166 tr oz. Therefore, for 700# @ 0.1%, (700/2000) x 29166 x .001 = 10.2 tr oz.

Forgive me, this is off the track of the thread but, the reason I've memorized this last conversion is that, when assaying such things as ores or low grade pulps, a sample size of 29.166 g is commonly used to make the final calculations easier. This weight is called an "assay ton (AT)". When using a 1 assay ton sample weight of 29.166 grams, the final result, in milligrams, is equal to the number of troy oz of gold (or, whatever) per 2000 avoirdupois pound ton of material. For example, with a 29.166 g sample, a final gold weight of 0.003 g is equal to 3 tr oz/ton. I had a set of AT weights that ranged from .05 AT to 1 AT - the smaller weights were used for richer materials. The sample weighings were made on a special small 2 pan swing balance called a "pulp balance." With AT weights and a pulp balance, the weighings went very fast. Probably less than a sample per minute. Probably faster than a modern digital scale, unless you could program it to directly weigh in assay tons.

When you need to know the answer in tr oz/metric ton, the same logic can be used but, in this case, an assay ton sample would weigh about 32.15 grams.

pinwheel said:
700 pounds x 453.44g = 317,408g / 1000 = 317.40g net weight. 317.40 / 31.1 = 10.2 oz troy.

Or more simply put: 700 lb = 10208.333 oz(troy) divided by 1000

Correct answer on both counts. The first method is excellent and very logical. Of course, instead of dividing by 1000, you could have multiplied by .001, the decimal equivalent of 0.1%. Same thing.

I don't think the 2nd example is simpler to the reader, since you didn't explain where you got the 10208.333 and the way you wrote it could be confusing. I know what you're saying, but others may not. To make more mathematical sense, you could have said something like: 700 lb = 10208.333 oz(troy); 10208.33 divided by 1000 = 10.2 tr oz., based on a 0.1% yield.



I hope everyone will excuse me for the long boring math posts. I must admit that, when I see math being done, especially when there are errors made, it draws me like a moth to a flame and I start blathering (as my wife adroitly calls it). To prevent making costly errors in this complex PM field, it is imperative that everyone can somehow do the math (or use the correct internet calculators correctly). I really do try to make the math I present as simple and logical as possible, although I often fail in this. If I say anything that confuses anyone, let me know and I'll try to simplify it. I also can make mistakes, so please correct me when I'm wrong.
 
Thanks for that last paragraph. The assay office is going to give me back a reading in milligrams and I was wondering how to reverse engineer that into something usable for larger weights.
 
pinwheel said:
Thanks for that last paragraph. The assay office is going to give me back a reading in milligrams and I was wondering how to reverse engineer that into something usable for larger weights.
If you're talking about my direct response to what you wrote, please note that I changed that paragraph a bit.

And that is x number of milligrams per WHAT - per ton (if per ton, what kind of ton?), per pound, per gram, per unit, per WHAT? You must know per WHAT.
 
Thanks Chris.
When you put it that way i see where i was wrong.


patnor1011 said:
samuel-a said:
I wrote <0.001%, meaning one thenth of a percent and even less.

0.1 is one tenth of 1
0.01 is one hundredth of 1
0.001 is one thousandth of 1

I stand corrected Sir.
 
Claudie said:
I think they have gold wires inside but I believe you need a microscope to see them. The subject of which chips have gold and which do not is somewhat confusing here I think. Some members say they have no gold, others say they do. I am sure that some of them do, but some may not. I suppose this should be a project for someone to find out about PM content and end the confusion. I think the best way to process them might be to incinerate them, then crush them, separate the magnetic parts out, then maybe go with AP. Suggestions?
FWIW
Lets not confuse the types of chips we're discussing - the only ones I remember that we have established PM content on are these http://goldrefiningforum.com/phpBB3/viewtopic.php?f=34&t=4383 which are a 2-part BGA (Ball Grid Array) with foils (the geometric patterns) in the green fiberglass part and gold bonding wire imbedded in the thermoset plastic top. See my photo in the thread.

Those are not what pinwheel is displaying - his is upside down, BTW - the usual package looks like this, in situ 100_0900.JPG - I'm a little dubious of the gram per kilo yield, or are we already past that? I had to go take a picture.
 
I will be running some BGA's with a new contest soon.Then we can all see a current yield record.Of course it won't be extremely accurate because of the different manufacturers.But I'll show a pic and weight to begin with and go from there.
P.S. Detectr can you grab me some peter pan peanut butter from that ad paper?Thanks I'll paypal you the $. :mrgreen:
 
About 90 percent of today’s chips are still packaged via wire bonding using gold wiring. The projections are all over the map, but gold is quickly losing its luster, as some believe that about 70 percent of all devices packaged via wire bonding will use copper instead of gold by 2013 or so.



http://www.eetimes.com/electronics-news/4207542/Copper-turns-to-gold-in-IC-packaging-
 
samuel-a said:
Thanks Chris.
When you put it that way i see where i was wrong.


patnor1011 said:
samuel-a said:
I wrote <0.001%, meaning one thenth of a percent and even less.

0.1 is one tenth of 1
0.01 is one hundredth of 1
0.001 is one thousandth of 1

I stand corrected Sir.


The only thing which was confusing is that you used % symbol where I marked it red. That confused me, and gave different meaning to numbers. 8)
 
"[la-la-la!] Shop Hy-Veeeee - where - there's - a -- Helpful smile in evr'y aisle!"

It'll be waiting for you at the bus station !!! :lol: :lol: :lol: But truly choosy moms choose Jif!

I'll be interested in your processing - I've run the bottoms thru a shredder, burnt (I can't honestly call it incineration - I used a reducing rather than oxidizing atmosphere, before I knew better. I was worried about the neighbors getting PO'd) then to Poor Man's AR. The excessive carbon crap made filtration a mess.

I've had good luck with them (green bottoms) in AP, though.
 
Since the assaying of these parts was mentioned a few times, I would like to put my 2 cents worth in.

Please note that, in this discussion, I am assuming that the parts are all the same. By this, I am talking about the package and not necessarily the chip contained in the package. In general, for most ICs, whatever value is there is mostly in the package and bonding wires and not the chip. The part number is determined by the chip. Therefore, you could have different part numbers but, if the package is the same, the values in each part should be in the same ball park. If there is a mix of different packages, it is a different story and a much more involved sampling method is needed than the one I am suggesting.

There's a lot of talk about assaying on the forum but little about sampling. Assaying is easy but sampling is difficult. No matter how accurately you perform the assay, if the sampling is done poorly, the results are next to worthless. The assay only tells you what is in the sample.

If the parts are the same, I would strongly advise not grinding them up before sampling, unless your assay involves recovering all the PMs from the entire amount that you ground up. Once ground, the material tends to segregate (especially stuff like those tiny wires) and the sampling becomes more inaccurate. If you do grind and run fire assays, you will need to split the ground material down with a sample splitter and then run a whole bunch of duplicate assays, especially when you have 700# of material.

The easiest and, probably, most accurate way on identical parts (at least, in appearance), such as these, is to fire assay the whole parts separately. To sample them, since they all appear to be identical, I would probably just grab a part here and a part there - I can't see how splitting down the lot by coning and quartering or using a sample splitter would give much improvement. In this case, you can probably run 2 parts per fire assay but, if the total copper and nickel is more than about 2 grams, you may only be able to run 1 part per assay. Since organics are involved, the parts must be incinerated before fluxing. This is easily done by weighing the 2 parts and placing them in an empty assay crucible. Then put them in a hot (say, 1850F) assay furnace, with the door opened just a crack to let in a little oxygen, and burn them until they fall apart into white ash. This might take 45-60 minutes. or more. Occasionally, you might have to pull the crucible out and give it a stir with a small diameter steel rod. If they aren't burned completely and any carbon remains, you'll probably end up with a too large lead button. After incineration, pre-mix the flux in a separate crucible, pour it onto the ash, stir it all up, and run a normal fire assay.

A big question is how many samples should you assay to get reliable results? For 700#, using the methods above and assuming the parts are all the same in appearance and size, I would probably start with at least 6 samples, which may be adequate. This is only a guess. The 6 assay results can be simply entered into the standard deviation part of the statistical mode on most any $20 scientific calculator. From this, you can compute the probabilities of the correct answer being between a certain range of values. If this isn't good enough for you, you need to run additional samples for better accuracy. Actually, from the results of the 6 samples, you can compute approximately how many samples you need to get a desired reliability. Please note that, no matter how you sample or how many samples you run, you'll NEVER know exactly what you have unless you refine the entire 700#. With samples, all you can do is get close. The more samples you run, the closer you'll be

There are many other ways to sample this lot. Probably the best way is to incinerate the entire 700#, then ball mill it, screen it, melt the metallics and drill the bars, and sample the ash with either a sample splitter or a grain sampler.

Then, of course, you have to deal with the refiner. He could care less what your assay results are or how you got them. The refiner is usually your biggest stumbling block to getting what you should get.

And you thought the assaying process was easy. What I have outlined is a simple case. Ores are probably the worst, since they are the most heterogeneous. In any case, you can't just grab 1 sample and assay it with any expectation of reliability whatsoever. Even with karat golds, you need to run at least 2 samples taken intelligently and most pros prefer 3. Refining is probably as much art as it is science, but sampling is almost pure science and it's not very intuitive. Correct sampling and assaying is where the money is made. It puts you a step above the person you're dealing with. Whether buying or selling, he who knows the true value of the material best, most always wins - old GSP proverb.

Chris
 
That information on sampling from GSP Is worth more than all the gold in the 700 pounds of chips. Thanks for putting that out there GSP.
 
i do a lot of these types of chips-----never obtained less then 1gram per kilo and most of the time also silver. some of these black chips that i refined have more then 5grams of gold per kilo;
i might get in conflict with some members of the forum but i do not wish to disclose from what type of scrap comes these high yeald black chips
regards
Arthur
 
arthur kierski said:
I might get in conflict with some members of the forum but I do not wish to disclose from what type of scrap comes these high yield black chips
It is perfectly understandable and acceptable in my opinion that you wish to keep a unique high yield scrap source private. Often I am in the same position as it is a clients material I have processed. They consider the scrap source and yields proprietary trade secrets.

Thank you for sharing your minimum and maximum yields from this type of scrap.
 
Glondor said:
That information on sampling from GSP Is worth more than all the gold in the 700 pounds of chips. Thanks for putting that out there GSP.
Thanks a lot for the compliment. It's always gratifying to know I haven't wasted my time.
 

Latest posts

Back
Top