Refining of Gold from Mines

[4metals]

Gentlemen,

While there is appeal in visiting far off exotic places all in the name of employment, there would be little to gain on the part of forum members who rarely see this degree of detail in an operation like Ashapura's refinery.

I applaud Ashapura's willingness to share the details of his operation in order to get valuable advice to aid his company. Hopefully others will learn to give detailed descriptions and photo's of their operations in order to get meaningful and helpful advice. Then the true spirit of the forum, Refiners helping refiners, can show through.

 

[ashapura refinery]

Firstly saying thank you will never be enough when it comes to the amount of help we have received. After reading everything and understanding things we have learnt a lot. As for the amount of money we have lost - we have no idea where to begin from. Today in the afternoon we were calculating on how much money we could have saved in doing the entire setup.

Coming to the process - we completed the batch how much ever we could. The washing of the refined powder was done well as per the instructions of Lou in our filtration vessel itself. There was no problem with that at all. Sadly everyone was so tired from working for hours continuously that we forgot to take pictures (Sorry about that.)
Now for the leftover solution - the pictures of which we had put up in our last post - we added SMB to a small batch of it but there was no reaction in that. Then we collected another small part of the solution in a beaker and added zinc to it and got black color.

For now we have kept that solution aside as it is. ( Pictures of the solution with SMB and with Zinc addition have been put up.)
Now for the purity - we dried the fine gold powder, melted a sample and did an assay and got the purity of 9992, the same as the previous batch. So yes the purity of the gold did come out to be good.

We next moved on to making a large bar and then making the grains of the bar and started the melting process for the Kilobars. This time we are moving on with 999.0 and not converting it to 995.0 as this was the request of the client.

Well the Bars are coming out to be good on the back (picture attached) but on the front (picture attached) we have these reddish spots which generally get removed by Hand Buffing and cleaning with sulphuric acid.

Tomorrow morning we will be marking these bars. Once done we will post pictures of the bars immediately.

Coming to the advice given for improvising:
Heating of the mixing reactor - As 4metal suggested - we will be contacting the company who supplied the glass equipments and check with them if they can take back the cylindrical vessels and provide us with a heating device for the mixing reactor. We will make arrangements for this at the earliest.

Other problems faced:
One small problem we faced during the Aqua regia process was that we had undissolved grains of approx. 302.00 gms. We had the stirring in the mixing reactor going on for nearly 4 hours. What we have noticed during the stirring process is that in the beginning the material starts dissolving very quickly but, like almost 80-85% gets dissolved easily but later the process slows down - is this because of too much of silver chloride being formed?

What other material we process:
We are getting the entire batch of dore gold from West Africa. We get it from one particular supplier in batches of 7-13 Kgs each batch and this dore generally contains gold, silver, copper and iron. Apart from this we do not process any other material at the refinery.

Apart from all this - one thing which is bothering us is that there are bigger refineries than ours in U.A.E wherein they follow the process which we had mentioned in the very beginning. And they say that their losses are way less as compared to ours. Its not like we do not want to go ahead with the method we followed now or anything of that sort. It is just something which we have been wondering about because the place from where they receive their material is also the same place from where we are receiving.

We talked about asking for consultancy by Lou or 4Metals today morning when we completed the process. So yes we will be contacting them to know their fees and the way to go about the entire thing.

And as 4metal said about sharing details, we can share every single detail we have no problems with that because we believe that learning has no end and as he mentioned that our details maybe helpful to the members of the forum – trust us we will be way more happy if anyone else can benefit from all the advice being given in this post.

Some more details about us:

When we decided on setting up this refinery in India (This is our first business in India) we consulted with our friend who is a chemical engineer and had previously helped in the setup of another refinery. We didn’t have 100% knowledge in the topic of refining but we did have some idea on how to go about it because we have previously worked with our friend and seen the process.

Our aim when setting this refinery was to make it one of India’s best refineries and also to reach the acceptance level of LBMA. India is a large consumer of gold and on a daily basis a lot of gold is being imported into the country. We had set this entire unit keeping large quantities in mind. Which is one of the reasons of having so many cylindrical vessels, a big tank for collection, a filter press etc.

But sometimes things do not work as planned, we had a lot of issues in setting up. A lot of delays took place and a lot of problems came as well but we worked our way through and then imported our very first batch. Now in the very first batch we had our friend who took care of the entire process with ease but there was a big loss, along with him we had an assayer who was excellent in assaying and did have some knowledge about refining but again not completely.

The losses remained unexplained and after that we did not get in touch with him, the assayer also left after a month because he was not able to adjust with the environment and also his knowledge in refining was less. We had done two batches of refining before he left.
We kept our search for a good chemist on and waited for sometime till we got somebody better. Upon recommendation we hired another chemist who was earlier in a good refinery. Based on his past experiences we hired him but had the same problem of losses which even he failed to explain.

Well this is just a part of the chemist. Apart from this we faced problems with the chemicals in the beginning. The acids were easy to find but none of them were of good quality, we had to then find a manufacturer in another state whose chemicals are much better percentage and grade wise.

Also the governmental policies changed due to which our importing licenses were cancelled and we had to wait for nearly 2 months without any work going on. So yes we have invested a lot of money in this setup and we have also incurred losses.

Anyhow this was our last big batch. Now we will not be processing big quantities and will instead process with a small quantity to make sure we are getting on the right path.

A small background info on us:
The person managing the entire process now is my father, who has experience in Gold Jewellery manufacturing. He has been in the field for almost about 35 years now. And he loves advancing with the technology and also never says no to learning something new. This refinery was something he wanted to do long time ago and that is how we have started off by getting the support from his friend and everything.
Even though things did not work as planned nor did we become one of the best refineries in India, but we are still glad we got to learn so much .

Tomorrow morning we will be stamping our bars and we will definitely put up the pictures of the final finished goods Hopefully the finishing would be good. Once we are done with our delivery we will be back on the recovery of the left over gold.

 

[nickvc]

I hate to say this but he fact your chemist walked out and also your assayer is rather suspicious, this business is full of crooks and thieves I just hope the losses you incurred are not down to them!
Not fully understanding the processes you have left yourselfs open to be robbed and have no comeback, how do you prove it?
Your aims are admirable but your ability to achieve them is low, sorry but he truth is often unpleasant.
You need to fully understand the whole process and methodology of refining to be able to see what's happening or not, this unfortunately you do not, if money isn't the problem I suggest employing a competent refining chemist or even just a competent refiner but you really need to know what they are doing and why.

 

[Palladium]

Cameras !!! Cheap and easy! Cameras everywhere! It might not catch them, but it sure keeps an honest man honest when he thinks someone is watching and keeping a record. It's also good to have to look back on if theirs an accident or questions about anything.

 

[Justincase]

I hate to say this but he fact your chemist walked out and also your assayer is rather suspicious, this business is full of crooks and thieves I just hope the losses you incurred are not down to them!
Not fully understanding the processes you have left yourselfs open to be robbed and have no comeback, how do you prove it?
Your aims are admirable but your ability to achieve them is low, sorry but he truth is often unpleasant.
You need to fully understand the whole process and methodology of refining to be able to see what's happening or not, this unfortunately you do not, if money isn't the problem I suggest employing a competent refining chemist or even just a competent refiner but you really need to know what they are doing and why.

Not actually as difficult as you might think.
A simple credit check and a little field work to see if there has been any out of character purchase will yield wonders.
I have spent twenty years in the security and investigative service's and I have to say the idiot's always expose them self's

 

[4metals]

The key to profitability in the refining industry is your analytical chemistry. Knowing how much gold you start with is critical as you can imagine.

Before you beat yourself up about yields and how much gold you lost, lets discuss how you determined how much was there to begin with.

So if you will, lets discuss your assay lab and review a few things that may help you to determine how good your lab is.

I assume you do both classical fire assay and have some sort of XRF device for quick assays, is this correct? Then, based on the fact that you said your purity of the most recently refined lot was .9992, are you using either an AA or an ICP for checking your purity?

Starting with some basic analytical concepts. A good analyst will always run multiple samples of the same material, using whatever method your lab has available. So to start with lets talk about accuracy and precision. A common mistake is people thinking they are the same thing. Accuracy is how close your result comes to the actual number. This is often difficult to pinpoint because human nature has us exaggerating these things or understating them depending on which scenario best serves our position. But a good assay is very close to the real number. Precision, on the other hand is how close your results are to each other. Remember I said a good assayer runs multiple samples? I prefer at least 3 samples and if the results are very close to each other, there is a high degree of precision. If I can get 3 samples to assay within .0001 of each other, I see that as a high precision assay and I would take the result to the bank and assume high accuracy. By looking at the spreads between analyses your assayer performs on the same sample, you can see how precise he or she is. Also submitting the same sample under a different lot number a few days later and seeing the result is a fair test as well. Sneaky.....maybe, but helpful!

I think it is critical for you to determine how good your own starting numbers are. It is also critical to do a proof corrected fire assay as results can vary by 1/2 to 1/3 of 1 percent based on the correction factor calculated. All of this leads to attaining the right starting numbers. If you would like we can discuss the techniques for a corrected fire assay here. We have a good number of members who have hands on fire assay experience here, it wouldn't be an exaggeration to say easily tens of thousands of individual assays if we added them all up, so detailing this process for you can come from others as well as me if I become a little too long winded! And I heartily encourage discussion here.

I would also like to know how you attain your silver assay. Is it by difference from running a doré bead? And how does your assayed result compare to your fine silver recovery?

Moving on, assuming you started a lot with 10 Kg of assayed material with an assay of 92%. You are confident the assay is good because you have reviewed the assayers notebook and the precision was acceptable. That means you expect (32.15 x 10) X .92 = 295.78 ounces to come out at a 100% yield. We know because you found values in downstream liquids and ducting that 100% is not attainable, don't lose sleep over it, nobody attains 100% accountability.

So you shot all of the gold recovered and it weighs 294.89 ounces. It assays at .9992 so your actual recovery was (.9992 X 294.89) = 294.65. The difference between what you started with and ended with is what you lost (temporarily hopefully).... (295.78 - 294.65) = 1.13 ounces.

dividing what you lost 1.13 ounces by what you expected 295.79 equals .382% lost.
Subtracting that from 100% you get 99.618% which is your lot accountability.

These were numbers I made up for example, most refiners running decent sized lots are coming up with in house accountability of 99.85 to 99.9% or better. And the losses do turn up in ducts or spills or in the silver chloride which in this case was not even calculated.

Now that you know how this is calculated, you can calculate your refining lot accountability for every lot (a very effective management tool!) and in discussions with your Middle East refining buddies find out their actual refining lot accountability numbers. Tracking these numbers routinely is critical to loss control and continual improvement. There is nothing that workers notice more than the boss coming out the day after a lot and letting them know that he knows EXACTLY how good or bad they performed. I always reward good performance and also let them know when there are issues. It is quite possible that a lot high in silver can hang up excessive gold which is not reduced and assayed for a few days, but when your workers know you have that under watchful eye they respect that.

Cameras can only go so far.....the workers knowing you know what is happening gives you better control.

If you can get back to us with a discussion of your analytical abilities in the refinery we may be able to help you fine tune this operation further.

 

[4metals]

Glowing hot gold bars, now that's a picture you don't see often. I assume you weigh out your shot and melt and pour while using a torch to prevent it from cooling too quickly? You should consider "toughening" your bars with a pinch of niter, it may be all you need to solve the copper color on the bar.

One small problem we faced during the Aqua regia process was that we had undissolved grains of approx. 302.00 gms.

For now I would attribute that to lack of heat driving the reaction. If the buildup of silver chloride buries the gold shot it will slow down the reaction so the agitation needs to keep the silver chloride moving. In any event, when you have undissolved material mixed with silver chlorides you need to separate the two.

For simplicity sake you can screen the silver chloride / undissolved shot in a polypropylene mesh screen. The screens I like (because they work well, are inexpensive, and last for a long time) look like this:


The screen sits perfectly in a 5 gallon plastic pail and you place the chlorides and undissolved on top and wash through the majority of the fine silver chloride particles. Remaining on the screen will be the undissolved, often crusted over with silver chloride. These can be put in a 4 liter beaker and heated in ammonium hydroxide. This must be done in a laboratory hood. The silver chloride will dissolve and the gold color of the shot will be apparent.

Rinse the undissolved material, dry the undissolved and melt it into a bar. Assay the bar and subtract the gold content from your expected yield as discussed in the previous post. It can be added to the next lot for refining. The ammonia solution now needs to be addressed. You should add Hydrochloric acid to it slowly after it has cooled and you will generate silver chloride as all of the silver chloride that was encrusting the undissolved shot which the ammonia dissolved, will drop out again as silver chloride. Filter this chloride and add it to the silver chloride you collected with the screen and all of your silver will be together for reduction.

While there are other ways of going about this, this is direct and pretty quick, and it doesn't cause you to end up with an alloy high in silver to deal with.

For now we have kept that solution aside as it is. ( Pictures of the solution with SMB and with Zinc addition have been put up.)

I do not understand why you added additional metabisulfite to the solution and then zinc. Your friend in a refinery is stannous chloride. (although it was not your friend the way you were using it) You should be testing the solution as you are dropping the gold from solution. When it tests negative to the stannous chloride test, all of the gold is down.

What did adding the extra metabisulfite do for you? And the zinc, did you add it to recover values from stannous testing? It does not warrant adding zinc, and producing waste treatment issues down the line, if your testing shows no values to be recovered.

It is encouraging to actually hear someone received bars from Africa that had gold in them. Here in the US refiners get calls daily by sellers of African gold and it is always a scam. I have melted quite a bit of "African gold sponge" which looked like the real thing, only to be an alloy containing no precious metals. We quickly learned to give the sponge a quick nitric test before wasting time melting.

 

[4metals]

I made a spreadsheet while on the train this morning.

Fill in the yellow boxes with assays and weights and it gives you the accountability.
As you add in values from the silver bar and undissolved the accountability goes up
Make one of these for each lot you run and fill in the blanks

View attachment Gold refining lot accountability.xlsx

 

[g_axelsson]

That's great!

I made it even greater... :mrgreen:

There is a pale yellow field after the unaccounted gold holding the unit. Now you can change it in grams, kilos, grains or even elephants. :shock: Just write your unit of choice there and it will change all the other places on the spreadsheet.

Göran

 

[4metals]

Elephants. I like that! International accountability.

 

[ashapura refinery]

Details of our assaying:

We have our own assayer in the refinery who has some good experience and has worked in assaying labs for more than 6 years. (The new person we hired after the old one left.)

Now generally we take 3 samples of the dore bar - 2 back from back side and one from the front. These samples are both taken for the fire assay test which takes place in our refinery itself. And we divide the three samples before we take it for the fire assay so that we can also get an xrf test done on this (XRF testing is done outside our premises as we do not have the equipment for that.
Dore pictures with sample taken -

For the fire assay test we have muffle furnaces and also the 6 digit scale for proper weighing of the samples. Most of the time fire assay and xrf result have a difference of 0.0005 (example 0.9296 and 0.9291), we calculate the average and then proceed ahead.

Today we did the accounts of the process we have done and here it is:


The silver chloride process and the checking of the drain valves, the filter cloths, the pipelines is pending.

Also here is a picture of the gold bar:


During the melting though we faced a problem. As you can see in the picture below, after the melting we cut a part of the bar and inside there is a shape which looks similar to a honeycomb. Generally when a bar is cut this kind of a shape does not come.

Also after the pouring of the gold in the mold, there is a small point where it gets sucked inside and then there are some spots on the bars. Probably this is because of the presence of some other material.

 

[goldsilverpro]

I've worked in many refineries and internal theft occurred in most all of them, especially the small to medium sized ones, most of which had lax internal security. In some cases, the thieves rationalized that since the refinery was stealing from the customer, it was OK to steal from them. In most cases, they just fired the employee. Too hard to prove anything, legally.

When I had my own refineries, a method I used to help avoid theft (at least in my mind) was to know exactly what all the in-house scrap materials were worth and I also kept track of scrap inventory and weights. Thank God for fire assay, AAs, a good lab, and a variety of good balances and scales. I let the employees know I was doing all that. They always saw me pulling samples and running assays. They had no doubt that I had the ability to evaluate anything.

At a big plating shop where I once worked, they caught a guy that had been stealing nickel (primarily) and copper anodes daily for about 7 years in his lunch box. Adds up. A gram or two of gold per day also adds up fast.

When I worked at a big refinery in Hong Kong, all the employees wore fairly skin tight overalls with no pockets. There was also a metal detector when they left for the day.

 

[4metals]

Generally a refiner does not accept a melt done outside of their facility. I have seen all sorts of games played with gold coated steel and other devious methods used by sellers who want to be paid quickly. When you get a bar in from an outside source, you should melt the bar, sample with a vacuum pin dip, and pour back into a bar. This will help you to avoid some potential issues.

One issue is anything foreign in the melt will show up and you will know the bar was not homogeneous. Another benefit is any melt loss will be on the new bar and as a refiner you should pay exclusively on after melt weight.

Once you know a customer, it is possible to melt the bar, take a dip sample, and then pour the shot to granulate the metal for digestion. I always like to dry it and weigh it after granulation. This way you can avoid one melt.

Vacuum dip samples are better than drillings, they are more homogeneous. This is a website of a manufacturer of the sampling tubes. http://pintubes.com/bur139/

I believe the honeycomb you mention is just the mark from the blade cutting the section, I do not usually cut fine gold bars to see the insides so I would seek a second opinion on this question.

The caved in section you mention is normal and can be prevented by playing a hot torch over the bar immediately after the pour to slow down the cooling off process. If you were to not use a torch at all it is likely on high grade gold to get that caving in, which we refer to as a pipe, along the entire center of the top of the bar. Slowing the cooling with a torch will eliminate it completely if done properly. From the looks of your bar, you do some torching after the melt and this bar was not torched long enough.

I think you can eliminate the spots on the bar by toughening with niter.

Using the spreadsheet posted above, fill in the numbers from an actual lot and let us know your accountability.

 

[Lou]

That is the formation of a pipe. This happens more and more prominently with higher and higher purity gold.
Watch this video carefully:
https://www.youtube.com/watch?v=nQ37SFKahc0
You heat the mold hot, and you bathe it in a reducing fan flame of methane or hydrogen gas to keep the gold surface shiny and control the solidification rate.

You need to bring XRF in house, even if it's a handheld thermofisher Niton gun. It's close enough for ball parking.

Also, if you're drilling the bars, sometimes iron from the drill bit will add iron to sample, decreasing the purity.

 

[goldsilverpro]

A fairly standard way (I've seen it in several old mining books) of drilling bars is to drill 3 holes on the top side and 2 on the bottom. The 3 on top formed a diagonal line - 1 at the center and 2 at the 2 corners forming the diagonal with the center. The 2 on the bottom were at the 2 corners not drilled on top. All 4 corner holes are brought in from the corners somewhat. On a 1 kilo bar with the same proportions as yours, I would bring the holes in at least 1/4", both horizontally and vertically, from the corner. On larger bars, bring them in proportionally. All holes are drilled only half way through the bar. If roughness is present on the top, which often happens with, say, 14K, I first drill very, very shallow holes on the drill press, just barely enough to clean up the surface. I remove those drillings and then drill the same holes as normal.

Since I don't want a ton of drillings, I prefer about an 1/8" drill bit for gold alloys. I put the ingot inside a small PVC tote box about 3" deep when I drilled. I then collected all the drillings and used a pair of jeweler's shears to cut them up into in small pieces. After blending the pieces together, I pulled at least 3 separate samples for assay. This method was proven to consistently give very good correlation with the triplicate assays.

With an induction furnace, though, I would give it a good stir and then take pin samples.

 

[Palladium]

I'm way out my league on them size bars, but i can speak of what i see from the experience side. The honeycomb you speak off i've seen before. Though not quiet that bad. I pour smaller bars so it less pronounced for me. That usually occurs between what i call a cooling ridge . Your bar didn't lay flat so where that rise or loaf is indicates a cooling rift. The outside of the bar cooled and formed the ridge or loaf before the middle solidified. The middle solidified as a mass at a later point than the ridge because of the thermal mass that the middle holds. The outer regions towards the edge radiates heat at a faster rate. Looks like you were almost there but not quiet hot enough, long enough. This can be caused by a cold pour, or a cold mold. You need a ox void environment, a pre heated mold, and enough btu heat to allow a controlled cooling effect. Do that with what 4metals recommended with a pinch of niter. Cool it slowly and it will make a holographic crystal pattern that will blow your mind. I would then clean them up with a 15-25% nitric bath and never have to touch them with a brush. The brush marks present a negative to me. The nitric will remove the flux and any oxides and do them more justice than that brush. Stamp them just the way they are. I've made the crystal pattern a trade mark in my gold bars for clients and they fall all over themselves for it. I assume with the money you've spent you're using induction?

 

[ashapura refinery]

@4Metals:In Africa, we have our own business setup there. We buy the dore gold directly from the miners. We get the gold in small nuggets so after the collection and checking the purity we melt it at our own office (we have a small induction melting furnace for melting all the collected gold). This way we avoid a problem of melting at someone else's place.

Now when we receive the gold dore at our refinery, the entire process right from the 1st melting to the final production of the gold bars is done here itself. For this we have Induction Melting Furnaces ( 20 Kg Melting, 8kg Melting, 5kg Melting and 1 Kg Melting Machines ). The only time the gold goes out of the doors of the refinery is when we need an XRF test and after the xrf test is done the gold weight is checked at our premises to make sure we have not lost anything.

@4Metals and @Goran - Thank you very much for this spreadsheet, it makes calculating even more easy. Hopefully this will benefit not just us but more refiners on this forum :lol:

@Lou and @ Palladium : We generally melt the bar, and make sure the gas torch is on when we are about to pour the gold in the mold. But we do not keep the torch flame on for a long time. So as per what you have said we should try keeping the flame on for a bit more longer than usual and reduce it slowly. Hopefully this will not result in the pipe formation and also help in getting a much more better finishing. But we must say that the finishing of the gold bar is way more better than the last few batches.

@Palladium : Yes we use induction melting machines. We have tilting induction melting machines(20 Kgs and 8Kgs) and also the normal induction melting machines (1kg and 5 Kg). We are using tilting for bigger quantities to make it easy for the staff to pour out the gold instead of risking carrying it out by hand.

@Goldsilverpro : we make holes of about 1/4" deep. But as per your suggestion of 5 samples - 3 in the front diagonally and 2 on the back - this will give a good coverage so next time we test we will make 5 samples instead of 3.

We are almost done with the cleaning of the drain valves, the pipes and everything. Once it is 100% complete we will definitely get back with the actual accountability to know how better we have done this time.
And this time we had our eyes on the staff - there were 4 staff members who were working for the entire batch and three of us keeping a watch. So this time we have ruled out the problem of theft. Hopefully our recovery is better than the before batches. We will keep you all posted on this.

 

[nickvc]

When commercially melting, recovering and refining the profits can be from the residues rather than the actual refined gold. If you are using flux in your melting make sure to keep it, there are posts here about how to recover any values within it most fom the experts. The flues over your melting will need cleaning again keep the dust, sweep and keep everything off the floors especially in the melting room, basically anywhere you have had values think if there may be anything to recover.
As an aside I had a good friend whose family were jewellery manufactures in india and the only scales in the factory were kept in the owners office, gold makes people do very odd things, I'm not saying your staff are thieves but put security in and point out that it's for their safety as if anything does go missing you can prove it wasn't them but as pointed out the best security is in you knowing exactly where your values are in what form and how to recover them, if you don't the old smoke and mirrors trick is easy to pull.

 

[4metals]

A few more questions for Ashapura,

So I gather from the response from the that their refinery has a fire assay lab, no XRF on site, and apparently no instrumentation, either ICP or AA. Is this correct?

Do you correct your fire assays? If so what result does this purity of feedstock average for a correction factor?

How is it possible then for you to state the purity of your fine gold bars for sale? XRF?

 

[4metals]

When commercially melting, recovering and refining the profits can be from the residues rather than the actual refined gold.

This is very true. Fortunately Ashapura is refining gold they have purchased at another location so their margins are not subject to the cut throat price wars happening between refiners these days. But when a refiners margins are tight, often the profits are found in the residues.

The accountability we spoke of earlier usually comes out showing a loss of values. There are exceptions though. Sometimes the difference between individual assays is sufficient and if your result comes out on the high side of the average, you can actually exceed 100%. That doesn't mean that you made gold, but it does mean your results exceeded your average assay result.

I can be pretty anal when it comes to recordkeeping in a refinery so I track all of my losses due to accountability on a spreadsheet. Then when I ship out slags for recovery, burn all the burnables and assay for values, clean out ductwork, or process stock pots from cementation, I add these numbers back in to my running totals and keep track of how much metal was "lost" to individual lot processing, versus how much metal was recovered in the facility from processing residues. If the gold retention is low enough in the silver recovery, the gold from the slimes recovery process is added in here too. (Just don't add it twice, it either goes back into the lot it came from or in when you clean up the silver cells)

In the long run, with good assaying and good housekeeping, you can get back almost all of what you expected to get from the individual lots. Unfortunately, you usually have to end up waiting for that last 0.15% until you clean up all of the drips and settles and other unscheduled events that yield values to your operation.