My new AMP Annealer

[Jefferson]

In Canada we have a dealer in Winnipeg xtreme shooting centre I think Jason is the guy bringing them in. Great more money to spend soon. Jeff

 

[Bob L.]

Does anyone have experience with the longevity of ferrite cores used in induction annealing machines? It seems that most applications require water cooling or have a limited duty cycle like the AMP.

 

[jrb572]

I recently received mine. I like it a lot. I noticed where someone had said something about brass inconsistencies. They actually recommend sending your brass in to them. They actually said my batch of Black Hills 300 win mag brass was different than what they had tested. If I remember their setting is 65 and for my brass is 62. They always recommend sending your actual brass to double check the setting is correct.

 

[jlow]

I recently received mine. I like it a lot. I noticed where someone had said something about brass inconsistencies. They actually recommend sending your brass in to them. They actually said my batch of Black Hills 300 win mag brass was different than what they had tested. If I remember their setting is 65 and for my brass is 62. They always recommend sending your actual brass to double check the setting is correct.

The point of my comment about brass inconsistencies relates to being able to get a perfectly consistent anneal from case to case because case weight and brass content varies, even if you are using Lapua. Case in point, here is a photo of a bunch of .260 Lapua brass sorted according to weight. Keep in mind that the weight variance could be MUCH worse if you are using cheaper brass. So a setting for a 171 grain case is going to anneal a 173 grain case differently because the heavier case will have more brass and will need more anneling time to get to the same softness.

 

[Ketosan]

Received Bryan Litz's new book--"Modern Advancements in Long Range Shooting" Vol 2 yesterday and one of the topics he covers is annealing. He has some concern with flame annealing due to the potential inconsistencies from flame temperature and point of "aim" of the flame on the brass. He used an AMP for his tests and thinks this technology is a step in the right direction for annealing. He only tested 223 and 308 rounds and the 308 annealing data was not posted due to an error in the annealing setting on the AMP. But he indicated that the 308 data was quite similar to the 223.

The test was done firing 10 rounds per setting and repeating this 10 times. Groupd A was never annealed, Group B was annealed after the fifth firing and Group C was annealed after every firing.

For the 223 using Lapua brass, the average SD after 10 firings, was Group A, 7.4, Group B, 7.5, and Group C, 6.9.

For the 308 using Lapua brass, the average SD after 10 firings for the never annealed brass was 8.5. As mentioned above he did not post the annealed 308 data due to an error in the setting on the AMP.

He did stress that this testing was minimal for now but does show that "maybe" annealing is not always necessary to get good SDs and therefore good target results.

Lots of good info in his book--as always.

 

[XBBR Shooter]

[...] For $1,000 the annealer would need to have a hopper and auto feed the brass before I'd even consider it.

I feel the same way - I currently anneal 2,000 to 3,000 cases a year, and I started annealing (torch and pan of water) back in ~1970's.
I now use a torch and electric screwdriver (no pan of water ), and heat to a dull red for 3 to 4 seconds.
Before I would spend big money on a machine (torch or electric), it would have to feed from a hopper (Dillon 1050 style), and handle the cases mechanically.

For that money, I am not going to handle cases one at a time - I am already doing that now !!

 

[jlow]

Received Bryan Litz's new book--"Modern Advancements in Long Range Shooting" Vol 2 yesterday and one of the topics he covers is annealing. He has some concern with flame annealing due to the potential inconsistencies from flame temperature and point of "aim" of the flame on the brass. He used an AMP for his tests and thinks this technology is a step in the right direction for annealing. He only tested 223 and 308 rounds and the 308 annealing data was not posted due to an error in the annealing setting on the AMP. But he indicated that the 308 data was quite similar to the 223.

The test was done firing 10 rounds per setting and repeating this 10 times. Groupd A was never annealed, Group B was annealed after the fifth firing and Group C was annealed after every firing.

For the 223 using Lapua brass, the average SD after 10 firings, was Group A, 7.4, Group B, 7.5, and Group C, 6.9.

For the 308 using Lapua brass, the average SD after 10 firings for the never annealed brass was 8.5. As mentioned above he did not post the annealed 308 data due to an error in the setting on the AMP.

He did stress that this testing was minimal for now but does show that "maybe" annealing is not always necessary to get good SDs and therefore good target results.

Lots of good info in his book--as always.

Thanks for posting Brian’s test. A couple of comments.

1. I can certainly agree that flame annealing using the socket and drill approach can be inconsistent due to “potential inconsistencies from flame temperature and point of "aim" of the flame on the brass”, but this cannot be true with a machine like a BenchSource. Hopefully his comment was related to the socket and drill method.
   
   
2. Just looking at his SD numbers, those are likely SD numbers for a single run and without SD numbers for multiple runs and a standard error number for those multiple runs one cannot determine if they are actually statistically significant which is the ONLY thing that matters in making the comparison i.e. is it noise or is it real? Being as close as those numbers are, I would say not likely.

The thing to keep in mind when looking at those numbers is a lot of work hardening is not related to firing but to how you work your brass when you size the brass/neck. The more you size, the more the work hardening. I am not saying his study is not useful, but without information on how he treated his brass, I would say take it with a big grain of salt when comparing it to your own processing.

 

[Kyle Schultz]

The point of my comment about brass inconsistencies relates to being able to get a perfectly consistent anneal from case to case because case weight and brass content varies, even if you are using Lapua. Case in point, here is a photo of a bunch of .260 Lapua brass sorted according to weight. Keep in mind that the weight variance could be MUCH worse if you are using cheaper brass. So a setting for a 171 grain case is going to anneal a 173 grain case differently because the heavier case will have more brass and will need more anneling time to get to the same softness.

I'm betting all the observed differences in weight are concentrated in the backend of the case which we deliberately don't anneal. Case necks are turned to a consistent thickness. Shoulders are bumped and case lengths trimmed. So the area of the case that does get annealed is pretty consistent.

In a related matter, the fiasco with the Norma 6 Dasher case rims is all the proof anyone needs to recognize that weighing cases as a proxy for internal volume is a complete waste of time.

 

[jlow]

I'm betting all the observed differences in weight are concentrated in the backend of the case which we deliberately don't anneal. Case necks are turned to a consistent thickness. Shoulder are bumped and case length trimmed. So the area of the case that does get annealed is pretty consistent.

In a related matter, the fiasco with the Norma 6 Dasher case rims is all the proof anyone needs to recognize that weighing cases as a proxy for internal volume is a complete waste of time.

You can see in my original post on this that I address the neck and shoulders with these comments:

"The reason is simple, because there is two sides to an anneal, the machine and the brass. Unfortunately regardless of how good the annealer is, the brass we use is not consistent. Weight your brass and notice that even with Lapua, the weight is inconsistent. Inconsistent weight means some cases have more brass in it than others. More brass means more annealing time to get to a specific anneal and so if your case brass content is inconsistent, the annealer no matter how good cannot fix this.

I buy Lapua cases by the 500 count and weight sort them so that each batch I use is less than 1 grain different but that only decrease the problem not fix it. Even AMP clearly acknowledge this since they have different program setting for different makes of brass but their number does not take into account variation within manufacturer which we 100% know exists.

If you use a ball micrometer and measure neck thickness, yes even Lapua will have variation and this affects the anneal for the above reason. One may argue that one can turn the neck – true, but how does that fix the brass content variation at the shoulders?"

As to weighting cases, you are unfortunately 100% wrong there. I have LOTS of experience weighting cases and they do in fact correlate very well if you know what you are doing. Always remember, if someone cannot prove something it could just be because they are doing it wrong.

 

[normmatzen]

I wouldn't worry about the ferrite cores used in the AMP power oscillator ( or any induction heater application). Ferrite is not a new, "state of the art" material and is universally used in any power application operating above line frequency. This is due to its high permeability and efficiency, especially at high frequencies. High frequencies meaning from over 60 Hz to hundreds of MHz. Only the size changes. Ferrites are used in MIL SPEC equipment down to cheapie consumer goods.
Heavy duty cooling requirements are mainly due to the copper losses in the coil wires. At higher frequencies, current flow in copper wire suffers from skin effect. That means all the high frequency current flows in a thin layer of the copper wire at the surface. The depth of this "skin" is inversely proportional to frequency and causes really fat wires to have an effective resistance similar to a wire way smaller. That is why Litz Wire is often used for transformer and coil windings at higher frequencies. Litz Wire is a fat bundle of relatively fine, insulated wire. The skin effect is usually about as thin as the wire so multiple strands (hundreds or thousands) allow the wire to perform like a solid wire at very low (like DC) frequencies. In any case, the cooling required in an induction heater is to cool the wire, not the Ferrite! The Ferrite will get hot too, but no way as bad as the copper wire.
As a metric, we are talking hundreds of amps in the work coil of an induction heater! The tuning capacitors have to be very stout as well as most capacitors of the capacity used in an induction heater are usually made for relatively low power circuits. To lower the resistance of the capacitor, quite often the "capacitor" is actually a bank of many larger capacity capacitors in parallel. And, yes, these capacitors must be kept cool also!

Usually, the Ferrite cores and tuning capacitors, as well as any connector blocks, can be easily cooled with small fans.

Yes, I did spend some time designing switched mode power supply controllers!

 

[Bob L.]

Norm,

Thanks. That answers my question. I am familiar with using ferrites for radio frequency applications, but not at the amperage used in induction heating. I was afraid the ferrite, being brittle, was more limiting than the wire.

Bob

 

[3081FTR]

Thanks for posting Brian’s test. A couple of comments.

1. I can certainly agree that flame annealing using the socket and drill approach can be inconsistent due to “potential inconsistencies from flame temperature and point of "aim" of the flame on the brass”, but this cannot be true with a machine like a BenchSource. Hopefully his comment was related to the socket and drill method.
   
   
2. Just looking at his SD numbers, those are likely SD numbers for a single run and without SD numbers for multiple runs and a standard error number for those multiple runs one cannot determine if they are actually statistically significant which is the ONLY thing that matters in making the comparison i.e. is it noise or is it real? Being as close as those numbers are, I would say not likely.

The thing to keep in mind when looking at those numbers is a lot of work hardening is not related to firing but to how you work your brass when you size the brass/neck. The more you size, the more the work hardening. I am not saying his study is not useful, but without information on how he treated his brass, I would say take it with a big grain of salt when comparing it to your own processing.

probably best you read his book

 

[jlow]

probably best you read his book

I probably will, but the point is not everyone here reading the post can or will read the book, thus the comment about what was presented here.

 

[dmoran]

...As to weighting cases, you are unfortunately 100% wrong there. I have LOTS of experience weighting cases and they do in fact correlate very well if you know what you are doing. Always remember, if someone cannot prove something it could just be because they are doing it wrong.

jlow -
Question for you, from these two scenario's:
1. Say you have 2 cases that weigh a difference of 2-grains from one another, and one of the cases has a rim thickness variation of say 0.002"
2. Same again, but say you mill section into 2 cases and one at the primer web is .006" thicker on one side then the other side, but the other case is more equal from side to side?

Neither of these scenario's would have any baring on the capacity, but definitely would on the case weights. How do you correlate such indifference?

-----------------------------------------------------------------

Opposite of jlow, more times then not I have seen poor correlation from case weights to the actual volumes.
Measurements from internal dissection as well as the external measurements prove repeatedly why gross case weights have variations that can have little to no effect on case volumes.

Sorry for the hi-jack...
Donovan

 

[jlow]

jlow -
Question for you, from these two scenario's:
1. Say you have 2 cases that weigh a difference of 2-grains from one another, and one of the cases has a rim thickness variation of say 0.002"
2. Same again, but say you mill section into 2 cases and one at the primer web is .006" thicker on one side then the other side, but the other case is more equal from side to side?

Neither of these scenario's would have any baring on the capacity, but definitely would on the case weights. How do you correlate such indifference?

-----------------------------------------------------------------

Opposite of jlow, more times then not I have seen poor correlation from case weights to the actual volumes.
Measurements from internal dissection as well as the external measurements prove repeatedly why gross case weights have variations that can have little to no effect on case volumes.

Sorry for the hi-jack...
Donovan

Donovan – I know we disagree on this.

I know theoretically it is possible to have differences in weight between two cases as you have correctly pointed out but have no effect on case capacity – I completely agree on this. However, theory aside, I know from actually measuring case weight and volume that they do correlate the vast majority of the time, the times they do not correlate is when the case has been fired and fired with different charges resulting in different pressure, thus expanding the cases differently.

We believe what we believe in based on the results of our studies. The result of which clearly is different, thus the difference in what we believe.

Here is my most recent study with Lapua .308 brass (sample size 40 cases) and volume measured with rubbing alcohol with x-axis brass weight in grains and y-axis case volume in mL. R-square value of 0.8263. I can tell you from my experience that if this was presented to a room full of Ph.D. scientists, there would be no push back on the conclusion.

 

[3081FTR]

I probably will, but the point is not everyone here reading the post can or will read the book, thus the comment about what was presented here.

I know but the comments are not helpful - some large assumptions when making statements like Being as close as those numbers are, I would say not likely.
A post asking for more clarification would help those that may not read the book. Otherwise the post comes across as something different altogether.

 

[jlow]

I know but the comments are not helpful - some large assumptions when making statements like Being as close as those numbers are, I would say not likely.
A post asking for more clarification would help those that may not read the book. Otherwise the post comes across as something different altogether.

What I meant was those numbers being as close as they are, the data would have to be really tight for them to be statistically significant, this is based on my experience dealing with data and what it takes to be significant, having said so, I would agree with you that they are also assumptions and getting the real data would be more useful. I stand corrected.

 

[dmoran]

jlow -
Just grabbed 3 cases from a qualification/segregation bin that I completed the other day by volume, that are with in .2-tenths by volume of each other.

Case 1
133.96-gr = Weight
1.557" = OAL (freshly trimmed)
0.467" = case head diameter
0.062" = case rim thickness
0.407" = extractor groove diameter
0.0105" = neck wall thickness (freshly neck turned)

Case 2
133.44-gr = Weight
1.557" = OAL (freshly trimmed)
0.4655" = case head diameter
0.0625" = case rim thickness
0.4055" = extractor groove diameter
0.0105" = neck wall thickness (freshly neck turned)

Case 3
133.08-gr = Weight
1.557" = OAL (freshly trimmed)
0.4645" = case head diameter
0.0615" = case rim thickness
0.404" = extractor groove diameter
0.0105" = neck wall thickness (freshly neck turned)

Now these are prime examples of what I see that does correlate with good consistency, and that is case weight indifference that come from small amounts of case head variations, that have nothing to do with the volume.
Pretty sure if this was presented to a room full of competent people, Ph.D. scientists or not, there would be no push back on the conclusion. Especially since it shows reasons, where, and why.

Donovan

 

[jlow]

Sorry about the room full of scientist comment – just tired last night. It was not called for…

So I have looked through your data carefully. From what I can tell, you are saying that the difference in case weight of your sample (0.88 grain) did not correlate to a difference in volume. FWIW, based on my study with the .308 brass which had a variance range of 4.04 grains, a variance of 0.88 grains only leads to approximately 7 ul difference in case volume. That is a very small volume which would be difficult to detect as slight variance in getting a flat meniscus could put enough noise in the data to make it impossible to interpret.

From a statistical standpoint, the biggest difference between your data and mine (apart from some of the case features you measured) is your sample size (3 vs. 40) and range of variance of your brass (0.88 grains vs. 4.04 grains). For example, my samples show a good correlation between weight and volume with an R-square value of 0.8263 (1.0 being perfect), but if I had only looked at a random subset sample of 3 of my cases (see graph) that is only around 0.88 grain difference in weight, you can see that the graph is much less convincing and R-Square value has dropped to 0.676.

When one is trying to determine if small changes in a sample are due to something, one needs a large sample size and large variance so that the difference stands out. The reality is this is not a perfect world and as those of us who has taken the time to measure case neck thickness knows, brass samples can vary by multiple ways and those differences can potentially be related to case weight but also can have nothing to do with case weight variance.

In terms of case head diameter and extractor groove diameter, I honestly do not know if they are responsible for the observed case weight variance, but a sample size of 3 will not convince any trained scientist that I know of. If you believe this is real and is responsible for case weigh variance, what I would suggest you do is to do what I did, which is to use samples with greater weight variance and at the same time have a much larger sample size. If you can show a good correlation with this method between weight and the variables that you think are responsible other than volume, it would be much more convincing.

 

[jrb572]

jlow I'll throw this out to you. Why don't you get six cases of the same brand you choose completely different weights and send them to the AMP people and see what they say? Easy said and done it will show to them with the test equipment if there is a huge difference.

http://www.ampannealing.com/sites/default/files/uploads/Contributor-form.pdf