My new case annealing machine

[TC260]

So when I anneal with the BenchSource, using two torches to get more even annealing because it covers both sides of the case, I generally run about 2-3 seconds. If you agree that I can keep the flames consistent and their distance from the case also consistent, then the absolute degree of annealing and therefore the resulting “softness” is completely dependent on the annealing time.

Now, if you stop and look back at the neck turning which we are using to control neck thickness and let’s just say the necks are 0.012” thick and we can turn them to 0.0003” accuracy then the % error would be about 2.5%.

So going back to annealing, if the annealing time is say 2.5 seconds, to get the same % error for annealing you would need to be able to reproducibly anneal your case with an error of 0.06 seconds. Even if you assume that using one torch and you go double the time i.e. 5.0 seconds, you still need to be able to reproducibly anneal within 0.13 seconds. I don’t know about you but I know I cannot get that kind of precision where I am looking at a stopwatch or listening to a metronome and get that case onto the flame AND get it off with the same degree of precision.

The reason that you can get that degree of accuracy using an automated rig like the BenchSource is because the dwell time at the flame is in fact automated.

BTW, I only include the dead soft case to show that over annealing can make a significant difference in neck tension. You obviously don’t need to go all the way there to have an effect just as you don’t need to cut your neck so unevenly that one side is twice as thick as the other.

Not saying this is right but at least thinking through it logically, this is what I came up with.

Logically what you're saying makes sense. However, to some degree it's based on an assumption that I'm not sure is entirely accurate. In particular, the longer the time in the flame, the softer (more ductile) the necks become. There is a relationship between heat, time, and ductility, but I don't believe it's a continuous and direct relationship like with turning necks. In neck turning, wall thickness is exactly related to the depth of cut regardless of how deep the cut is. Same with case trimming. Annealing is a little more complicated than that.

Essentially all we're doing with partial annealing is applying a sufficient amount of energy (heat) to the material so that the point defects and dislocated rows of atoms can rearrange themselves from a high energy state (hard) to a lower energy state (soft). In a sense, the material is "fixing" itself. Get it hot enough for long enough and very rapidly the atomic structure rearranges itself internally. It's not a situation where you have a perfect linear relationship whereby for each degree warmer the material then gets an x percent softer. That's why we can anneal a given case multiple times and it won't get any softer than it does after the initial annealing. Once all the high energy bonds are broken, there aren't any more to fix so there's no further change to its properties. By the logic you’re using, second and third annealings would continue to make the case softer and softer with each successive treatments which isn’t the case.

As long as the cases are heated hot enough and long enough for the internal changes to take place, the material doesn’t care what the heat source is. I believe that's why we don't see much if any difference in the finished product between automated annealing and proper annealing technique using hand tools. Again, if you are finding large differences, where’s the data? How are you coming to that conclusion other than reasoning it out?

Now when you’re using a machine with two torches, then yes of course, time control is very important. It’s important in large measure because you’re moving the part through the temperature window quickly. From what I’ve seen of induction annealing, the time control is critically important for the same reason. If using one torch turned down low or some other less intense heat source then the part would be coming up to temperature much slower and the time factor is much less critical. Again, once the material gets hot enough long enough for all the high energy bonds to rearrange themselves then the process is completed. There’s no more increase in ductility that can be gained within the sub-liquid temperature range.

 

[TC260]

The reason I suggested that we ignore "dead soft" cases is because when the brass is heated into the liquid region, there's a completely different reason for the case softening than there is as a result of recovery annealing

 

[DanConzo]

A lot of people anneal the neck and don't time it at all, they watch the colors and/or the glow and some use tempilaq and they shoot great. Assuming two torches annealing more even is a heck of a good sales pitch but where is there any evidence. I would make an assumption "guess" that a piece of brass .010" to .015" thick would be real even pretty fast especially if its spinning. I guess if you believe that two torches is better than one and it works for you, that's great, but others may not believe your assumption in the first place, so it's a dead end debate. The end result is what counts.

 

[jlow]

TC260, Thanks for writing in, it’s good. Lets talk more about what you said. You said “to some degree it's based on an assumption that I'm not sure is entirely accurate. In particular, the longer the time in the flame, the softer (more ductile) the necks become. There is a relationship between heat, time, and ductility, but I don't believe it's a continuous and direct relationship like with turning necks.”

As I see it, the problem is exactly what you said i.e. “I’m not sure”. My concern is in fact that we don’t know what that relationship is and as far as I can tell, most people appear to be working on the assumption that “it does not matter much” which to me is dangerous.

Lets take a look at what those relationships could be. Here is a chart that I made up to show four possible theoretical relationships:

The first one which is the “GREEN” line assumes a linear relationship where for every increase in time, there is a linear increase in softness.

The second one, the “RED” line is the “exponential” relationship where at least during the beginning, increasing the annealing time causes very little change in softness but past the inflection point i.e. 7 seconds, there is a rapid increase in softness.

The third one, the “BLACK” line is the “logarithmic” relationship where during the start of annealing, there is a rapid non-linear increase in softness but then this slows down past the inflection point i.e. ~4 seconds where there is little increase in softness with increase heating time.

The fourth one, the “BLUE” line is what I call the “phase change” relationship where nothing happens with increase heating but then say at 3 seconds the work harden metal goes to perfectly annealed softness and then this would not change until 7 seconds where it abruptly goes to super soft.

If someone knows the answers to my questions, I would surely appreciate them speaking up, but as far as I can tell, the information out there on this part of annealing is mostly vague and not useful.

Of the four, the “logarrithmic” would be the worst case as you can see assuming an annealing softness of 5 is what we are targeting since a slight degree off in annealing time can cause big difference in softness. The best case would be the “BLUE” relationship since heating time between 3 to 7 seconds makes no difference.

The real question though is which relationship in fact we are dealing with. Here lies the REAL PROBLEM i.e. we don’t know or at least I know for sure that I don’t know, thus the point of this discussion.

DanConzo – I have no problem if you don’t care about our discussion but just keep in mind that some of us may not be “there” as some of the more seasons folks and well, we also like to know “why” when we learn “how”. This is how I was educated and it has served me well.

 

[DanConzo]

jlow--If there is any fruit to bear here for any of us, I do care about the discussion. You at least put some substance into your posts, and don't just take pot shots like some who don't really offer anything but bumper stickers, but I'll stand by trial and error until something better comes along even if I don't fully understand why it works. The annealer you have looks like a real nice one and probably does an excellent job. If you figure any or all this out please post.

 

[TC260]

TC260, Thanks for writing in, it’s good. Lets talk more about what you said. You said “to some degree it's based on an assumption that I'm not sure is entirely accurate. In particular, the longer the time in the flame, the softer (more ductile) the necks become. There is a relationship between heat, time, and ductility, but I don't believe it's a continuous and direct relationship like with turning necks.”

As I see it, the problem is exactly what you said i.e. “I’m not sure”. My concern is in fact that we don’t know what that relationship is and as far as I can tell, most people appear to be working on the assumption that “it does not matter much” which to me is dangerous.

When I said, "I'm not sure your assumption is entirely accurate". That was a polite way of saying your wrong in your assumption without saying it directly. I do a lot of teaching so it's a habit to correct people in a gentler tone rather than blasting someone like a lot of people do on internet forums.

I understand where the confusion comes from in these discussions. The study of how materials behave is very complicated and I'm not an expert on it by any means but I try to explain what I know in the simplest terms possible.

The graphs are nice but I'd do them a little differently. We know that there's a time/temperature curve for the annealing process to complete itself. At 450 it takes about an hr and at 750 it's takes a couple seconds for the process to finish. So if we were to plot a ductility(softness as we're using the term)/time graph, depending on how intense the heat source is, we'd see an exponential increase with time up to a plateau when the process is complete. The graph would continue flat until the temp go up to about 900 degrees when the brass goes into the liquid region. Where for our purposes it would spike straight up to some higher level of softness. Exactly how soft is based on factors that are way outside the scope of a reloading discussion but it's ruined for us.

Going back to my earlier post, recovery annealing is a finite process. Get the brass hot enough for long enough and the process is complete. That's why I believe that when it's done properly, there isn't a measurable difference in the final product between a machine annealed case and a hand annealed case. If enough heat is being applied to anneal the cases in a couple seconds then a machine is the only way to go because the margin for error is too small to do it by hand consistently. If the heat is turned down so the process takes 7,8,9 seconds then margin for error is much larger and can be done just fine by hand.

 

[jlow]

DanConzo – thanks! Just trying to flush out more information for all of us.

TC260 – Thanks for your input. For what it’s worth, I have no problem being wrong as I am here to learn. Having also been a teacher myself for many years, I have learned that having an open mind and moderated ego is the best way to learn.

Your information is interesting. So from reading what you written, I imagine that the “proper” graph may look a little like a cross between my “exponential” and “phase change” curves. It also appears that we don’t disagree about the importance of accuracy when it comes to high temp/short time annealing and the benefits of doing it with a machine and dangers of doing it by hand.

It’s all good stuff – thanks!

 

[TC260]

Yep, a cross between the exponential and phase change graphs. Although, since all of our methods are being done over such a short time period I would expect the exponential part of the graph to be essentially a straight vertical line up to it's plateau. Practically making it an all or nothing process. If we were increasing the temp over the course of minutes then I'd expect we'd see more of an exponential curve to the graph with the plateau being reached at a lower temperature and more possibility for variances in hardness with insufficient time to complete the process.

As I mentioned earlier, this stuff is confusing because basically the same term, partial/process/recovery annealing and full/recrystalization annealing, is used to describe what are two fundamentally different changes in the material. In the sub-liquid (below roughly 900) temperature range, recovery annealing is simply a matter atomic rearrangement. Get it hot enough without going too hot and we get the desired result.

In the liquid temp range above 900, then there are changes to the grain structure which control most of materials properties. Quenching, tempering, furnace cooling, etc, are all methods that are used to control the cooling rate of a material through the liquid range to achieve a desired grain structure. Sometimes those get thrown in to reloading discussions so we have terms and techniques being mix-matched all over the place.

 

[jlow]

+1! Thanks!