• This Forum is for adults 18 years of age or over. By continuing to use this Forum you are confirming that you are 18 or older. No content shall be viewed by any person under 18 in California.

Induction brass annealer redux

Re Mr Vander Voort, compare fig 4 and fig 10 with the latter using 30 minutes. Cold working to 50% reduction took hardness from just over 50 to just under 200 HV. 30 minutes at 900F didn't return the cartridge brass to its starting state. As I said, it takes a hell of a lot of annealing to get back close to the state pre cold work hardening. Luckily it would appear the level of cold work hardening tested was rather severe - or at least I hope it is. It would have been interesting if Mr Vander Voort had tested more practical work hardening scenarios, eg how much hardening occurred in 1, 2 or 3 firings, and then produced fig 10 for each.

I never said the formula makes sense. I don't think anyone here has a clue if the formula is valid or not. (And it isn't particularly complex as far as formulae go.) What I did say is the statement that anneal time is temp dependent makes sense. Honestly, no surprises there. There was always likely going to be a temp for which required anneal time was extremely short versus a temp for which required time was going to be much longer. The question has always been "what is that temperature?"

Indeed, so far Ambrell using 1250F is the only corroborating support of Reese's formula that I have seen which is why I said:

About the only defense of Reese's formula I've found is from this Ambrell case study. Plug in their target temp of 1250F from their PDF and you need only hit it for a tiny fraction of a second (which seems to be supported by their video).

Note the formula is also dependent on the T1/t1 assumptions: that an hour at 700F provides an adequate anneal. Reese is explicit that he is providing merely a translation of this assumption to a higher T2 working temp.

All we can do is make a leap of faith (the Ambrell case study helps us make that leap) and use it until something else comes along. But don't kid yourself that it's currently much more than that. What is clear from the discussion and the Ambrell case study is that touching 750F for a moment is not going to be enough. People here can choose whether to follow the Ambrell example and simply use, say, 1200F Tempilaq (Reese's formula suggests a mere 6 hundreds of a second for the required anneal time**) or whether they wish to use a more complex GinaErik implementation to hold a lower temperature for a period of time measured in seconds.

** Any implementation requiring visual observation of Tempilaq color change, whether using 1000F or other values, is very unlikely to be able to resolve the time at which that change occurs to hundreds of a second and quite likely not much below even 1/2 second accuracy. So one needs to be realistic in thinking that 1.88 secs - or whatever - is any more accurate than using simply, say, 2.
 
Last edited:
"brass annealing is time-temperature dependent."

Yes, that's obvious with even cursory thought. Very, very little time at some temperature and longer at a lower temperature. The question has always been "how much time at what temperature?" This has typically been considered in the form of 'touching' a target temp, ie trying to understand what the "very, very little" time end of the scale is.

"The Metal Engineer"

This we are left to take on faith. Unfortunately he doesn't provide any credentials in the video. Again, it's a shame he doesn't provide a little background on himself and his metallurgical skills. (The rest of the video could easily have been a quarter of its length to accommodate this.) Perhaps he discussed his background more in another video?

"His formula (may be he is the author or comes from his text books)"

That's my point. Unfortunately we have no idea where it came from and he doesn't make an attempt to explain it. As a result it's disappointing in this regard.

Ambrell Corporation make, amongst many other things, custom commercial annealing machines, including those for ammunition casings. They come with a lot of credibility. Two good things come from the little case study they posted on their website. The frequency their demonstration annealer operated at, 113kHz (something that's been debated here before) and their target temperature, 1250F, each stated in their pdf.

About the only defense of Reese's formula I've found is from this Ambrell case study. Plug in their target temp of 1250F from their PDF and you need only hit it for a tiny fraction of a second (which seems to be supported by their video).

Anyone working with a simpler GinaErik implementation would likely do well to use 1200F or even 1300F Tempilaq as their target. (Given the vagaries of deciphering color changes in Tempilaq 1200F or 1300F is likely accurate enough as attempting greater accuracy is likely spurious.) It also supports the idea that you want to deliver a great deal of power very quickly to the area of interest.
Link to Vander Voort's biography below - interesting read. Think his subject matter creds are pretty impressive:

Biography of George F Vander Voort

Biography – George F. Vander Voort​

March 31, 2009 by Alan Charky

george-vander-voort-3


George received his BS in Metallurgical Engineering from Drexel in 1967 – the last class as Drexel Institute of Technology and received their Distinguished Alumnus Award in 2005. At Drexel he was twice editor-in-chief of the Drexel Technical Journal, including the three years in a row when it was named the number one college technical journal by the Engineering Colleges Magazines Association. He was a member of Blue Key National Honorary Leadership Society and in 1967 he was listed in Who’s Who in American Colleges and Universities. Sigma Alpha Mu fraternity named him their outstanding senior in 1967. After graduation, George joined Bethlehem Steel Corporation in Bethlehem, PA. He received an MS degree in Metallurgy and Materials Science from Lehigh University in 1974. After six years in the Bethlehem Plant metallurgy department he was transferred to the Homer Research Laboratories as a research engineer; nine years followed in the Metallurgical Service and Investigations group with involvement in all areas of microscopy, failure analysis work and alloy development. In 1983, George joined Carpenter Technology Corporation as supervisor of Metal Physics Research in their research center. He was responsible for all areas of microscopy and mechanical testing. From 1996 to 2009, George worked at Buehler Ltd. as Director of Research and Technology. He edited their newsletter, Tech-Notes, produced their annual microstructure calendar and was responsible for education, laboratory service, and new directions, e.g., EBSD and SEM imaging. In 2010, George became a consultant to Struers Inc, Struers A/S, Latrobe Steel and Scot Forge, plus president of Vander Voort Consulting.
George is the principal author of more than 280 publications, including the book Metallography: Principles and Practice (McGraw-Hill, 1984, ASM Intl., 1999, 752 pgs.) and Buehler’s Guide to Materials Preparation. He has edited eighteen books. He is the author of 29 articles in various editions of the ASM Metals Handbook series, was editor for the 2004 revision of Vol. 9, Metallography and Microstructures, and made eleven of the fourteen videotapes in the ASM video course Principles of Metallography. He is the author of nine ASTM standards and holds six patents. His micrographs have been within or on the covers of 137 books, magazines, newsletters, brochures or calendars. He has won 34 awards for his work in metallography contests including the Jacquet-Lucas Grand Prize. He taught physical metallurgy at Pennsylvania State University (Allentown branch) for nine years, a graduate course on microscopy at Lehigh University, and was a visiting lecturer at the University of Canterbury in New Zealand in 2003. He has taught 72 one-week courses for ASM’s Metals Engineering Institute (MEI), 110 courses for Buehler and 28 for other societies.
George has been active with the International Metallographic Society, IMS, since 1973 serving on the board of directors, as membership chairman, secretary, vice president and president (1981-1983). He was general chairman of their 12th annual meeting in 1979, and their 32nd, 36th and 38th meetings in 1999, 2003 and 2005. He was series editor of Microstructural Science from 1983 to 1989 and he has chaired nine of their symposia. He was associate editor of Materials Characterization (1991-2004) and is a member of the editorial boards of La Metallurgia Italiana, Praktische Metallographie/Practical Metallography and the Intl. Journal of Microstructure and Materials Properties. George received the IMS President’s Award in 1987 and the IMS Henry Clifton Sorby Award in 2004.
George is a member of the American Society for Testing and Materials, ASTM (now ASTM International), since 1979, as a member of committees E-4 on Metallography and E-28 on Mechanical Testing. He was chairman of E.04.14 on Quantitative Metallography (1982-1998) where he developed and wrote 9 standards for both manual and automated quantitative metallographic measurements. He served as second and first vice-chairman of E-4 and a four-year term as chairman of E-4. He chaired two international symposia for ASTM E-4. He received the ASTM Award of Merit in 1987 and is a fellow of ASTM. In 1990, he received the Anthony DeBellis Memorial Award from ASTM E-28 for his work on microindentation hardness testing. In 1994, he received the L.L. Wyman Memorial Award from E-4. In 2006, he received the J.R. Vilella Award for his revision of Vol. 9 of the ASM Handbook, Metallography and Microstructures (2004 edition). Since 1989, George has been the USA representative to the International Standards Organization, ISO, sub-committee on tests other than chemical and mechanical (SC 7 of TC 17). He revised two ISO standards and wrote one new standard.
George is a member of the American Society for Metals (now ASM International) since 1966 (Life Member); and served on the executive committee of the Lehigh Valley Chapter (secretary, 1971-1974). A past member of the MEI and the Academy Committees, he chaired the Technical Books Committee and was secretary and vice chairman of Publications Council; now he is on the Handbook Committee, the Events Committee and the International Committee. He has been an MEI instructor since 1977 (72 courses taught). He was made a Fellow of ASM in 1990; received the Bradley Stoughton Award of the Lehigh Valley Chapter in 1993. In November 2001, he was elected to a three-year term as a trustee of ASM International. He has given 356 lectures (six honorary lectures) in 39 countries; spoken 81 times at 51 ASM Chapters, and at 74 universities. He received the Roland Mitsche prize of the Montanuniversität Leoben, in Leoben, Austria, for his contributions to metallography. In 2008, he was named a Distinguished Life Member of Alpha Sigma Mu honorary scholastic society for materials science majors (and was elected to their board in 2009, and is in-coming vice president). The September 2009 issue of Practical Metallography was dedicated to George’s 65th birthday.
George is well-recognized for his work in failure analysis, both through publications and investigations. He has published 16 articles on failure analysis in journals, 8 by-lined articles in the ASM Handbook series and has edited 8 books related to failure analysis. He has also published 44 papers on quantitative metallography on subjects such as inclusion and grain size measurement, clustering, banding, and phase measurement. In 1993, he was hired by the Nuclear Regulatory Commission to assess the metallographic study performed at Argonne National Laboratory and Idaho Falls National Engineering Laboratory on the lower head of the Unit 2 nuclear reactor at the Three Mile Island site that failed in 1979. George’s recommendations led to his measurements that quantified the precise degree of thermal exposure experienced at different head locations under the debris pile. George also examined heavily eroded structural steel from the World Trade Center for FEMA after the 9/11 attack to determine the reasons for the severe deterioration, particularly for Building 7 which was not impacted by the terrorists. George has worked on a number of cases involving semi-submersible and fixed oil drilling platforms in the Gulf of Mexico, locomotive axles that broke and caused derailments, collapsed Loran Towers, etc. He is a court-certified expert witness in microstructural aspects of failures.
George is also a member of TMS of AIME, the International Society for Stereology, The Polish Society for Stereology, the Microscopy Society of America, the Microbeam Analysis Society, State Microscopy Society of Illinois, Midwest Microscopy & Microanalysis Society, Deutsche Gesellschaft für Materialkunde, and the Royal Microscopical Society (UK). He has two children. When he has some spare time, George enjoys foreign travel, food and wine, watercolor painting, photography, hiking and backpacking. His wife, Dr. Elena (Manilova) Vander Voort, is a physical metallurgist and electron microscopy expert on high-temperature alloys with 28 years service at the Pulzonuv Central Boiler and Turbine Institute in St. Petersburg, Russia, in the failure analysis and life extension department.
 
Note that Fig 4 is from Vander Voort's work.
Hardness.png

Fact - full annealing is achieved when the hardness is below 100.
Fact - 30 minutes at temperature below 800 F are not enough to achieve full annealing (Reese formula requires minimum of 3 hrs to do that).
Fact - After 30 minutes at temperature 950 F full annealing begins. Note that Reese formula requires 16 minutes and exponentially lower minimum time for higher temperatures.
Your "case study" mentions that at 1200 F for 0.1 sec, Reese formula is correct.

Conclusion: It is quite clear that Reese formula follows the pattern for the above experiment. It is also quite clear that prolonged annealing time (30 minutes in this case) at constant temperature does not damage the brass.

Practical opportunity for better results: Since my build is capable of delivering constant temperature ( thanks to VenatusDominus, who introduced us to the flame sensors), I can easily prolong (with limits of couse) the annealing process and cover possible deviations (material or sensor calibrations for example) .

.
 
"Fact - full annealing is achieved when the hardness is below 100."

Why does Norma anneal their brass to numbers above 100...? So the above is likely opinion/conjecture rather than fact.

I'd suggest you watch Reese's video again. He makes no claims as to the requisite hardness/softness. He merely points out that ideally one should land at an area where the delta hardness relative to time is modest. In that area one is more likely to have broadly consistent hardness than if working to a steeper part of the curve. Pick a point on the Vander Voort curve (one of many such curves) and go from there. Reese doesn't define which point you ought to pick. He used 700F as an example. The 700-900F area on the Vander Voort graph still looks pretty steep to me, albeit not as steep as 500-700F.

We can however, observe one good thing about the 700F point: it would seem to eyeball relatively close to the level Norma brass is produced to - hence the relevance of this other paper I linked to. So that would be 700F (or maybe 750F) for 30 minutes not 60. That will cause a big delta from 1.88s when using Reese's formula and 1000F.

At any rate, my point is that we continue to speculate and shouldn't be thinking that we've achieved gospel. The Ambrell case study is the only example I have been able to find of a professional disclosing their target temp and frequency goals for this exact process we are endeavoring to do. They did not disclose time at target temp nor much other detail.

I think things are very much improved as a result of Reese's discussion, the Ambrell case study, these other research papers and potentially the use of a flame sensor. But we should all realize that without proper hardness testing we merely hope we are getting the appropriate results.

So have fun and don't get hung up on second decimal place precision. A guess of 1s at 1000F is likely as good as a guess of 2s or 1.88s.
 
So have fun and don't get hung up on second decimal place precision. A guess of 1s at 1000F is likely as good as a guess of 2s or 1.88s.
You betcha' I'm having great fun indeed. Even your last statement, which was quite bitter and grossley incorrect, made me smile.

Here is my statement vs you guess.
Practical opportunity for better results: Since my build is capable of delivering constant temperature ( thanks to VenatusDominus, who introduced us to the flame sensors), I can easily prolong (with limits of course) the annealing process and cover possible deviations (material or sensor calibrations for example) .

.
 
Ha I’m far from bitter. I’m a realist. Yes, each of us, thanks to VenatusDominus, have platforms with similar capabilities. The question is how much progress have we really made without hardness testing alongside annealing scenarios? Yet again we find ourselves needing that in order to have great comfort we are achieving the results we desire. (And shouldn’t be surprised this is the case.) I think our guess is a little better than before. Touching 750F rather momentarily isn’t anywhere near enough. Yet we’ve known this for some time and have defaulted to the view that if the area well below the shoulder hits 750F then the neck and shoulder, the real area of interest, will have reached a much higher temperature. We could have tested this with higher temp Tempilaq but I don’t think anyone bothered to do so. 1 or 2 seconds or so at 1000F is likely better. Or quite possibly just touching a higher temperature like 1200F-1300F might be better also. People can decide whether to build a complex platform or satisfy themselves with the simple one and just use the higher temp Tempilaq. Anyone with a simple GinaErik annealer would certainly do well to purchase some 1200F Tempilaq from the likes of McMaster. One day someone here will have access to proper hardness testing tools such as a Vickers and do even more relevant testing…
 
Here's what Richard Ohlsson said about the annealing performed by Norma:


As seen in Table 2, the cases experience two different annealing’s of the neck during the manufacturing processes. The annealing at Norma is made by induction heating where the Win Mag at the first neck annealing is annealed for ca 14 seconds at an induction of 2,8 kW and at the second neck annealing ca 14-15 sec at 6,1 kW. The exact temperatures have not been measured but previous measurements at other cases at Norma Precision AB has shown that the temperatures lie in the range of 350-600°C. From Figure 5, it is seen that annealing at those temperatures does not change the phase of the microstructure for the brass.

Tables 7 and 11 provide the final hardness of the manufactured Win Mag and 308 Win, respectively, cases at the points discussed in the methodology.
 
Last edited:
On thin parts like necks, when we get to above 980 F, the less effective Tempilaq becomes.

At the Draper point, there begins an infrared glow in metals that would make the interpretation vague.

Also, when using higher power induction and the heat rate is fast, it also makes interpretation more difficult because you are catching a more rapid change.
 
I think most people are setting a particular time, annealing the case for that time and then examining the Tempilaq (cold) rather than trying to observe it in real-time - creeping up on the point of change, at least when getting to the finer side of things. Personally I find examining Tempilaq a rather vague affair. Did the color change enough to indicate it hit temp? We can control time in very fine increments and creep up slowly. However we are still left 'interpreting' the Tempilaq. Hence my note above that I doubt people can do this with very fine timing/temp accuracy. Maybe in the hands of an experienced user, having been guided by someone very experienced before them, Tempilaq can be used very accurately. For most of us I suspect, however, it's still a very approximate tool.

Of course when we introduce a sensor such as the flame sensor we are introducing yet another variable. We have to calibrate it with Tempilaq while observing the output of the ADC and then hope the sensor is stable over time and temp. (And God forbid the sensor to case distance changes for any reason.)

I'm not saying it's bad - far from it. I've built a flame sensor into my own build (and can decide whether to use it or not). But I am saying we shouldn't get too excited about whether we've reached an ideal goal. Even if we take Reese's formula for granted (and unfortunately we have very limited, if any, reason to do so) there's still doubt over the T1/t1 inputs that the formula requires, we have all the issues associated with calibrating the flame sensor to Tempilaq and hoping it's stable, and we don't have hardness testing to confirm that we are indeed achieving what we hope to achieve. Given the high likelihood for error/approximation we shouldn't get focused on spurious levels of accuracy. Onward.
 
Just a curiosity question - Since the beginning this "project" has matured quite a bit. Since I'm retired and enjoy projects (depending on what they are), just wondering if anyone else like minded would consider a higher level starting point for newcomers. Meaning, code for an Arduino that would cover options like flame sensor or time, straight flame sensor (trip when it first reaches temp) or maintain temp for a selectable time by pulsing output, trap door and case feeder. Would need to figure out what I/O would be needed for case feeder.
I'm somewhat comfortable with Assembly for PIC micros but I've barely touched an Arduino. I'd be lucky if I remember how to turn one on! I suppose I could do a shield for option select switches and I/O connections plus the header to connect to the Arduino.
So, if a newbie would buy the Arduino, a shield and all the other normal parts they would be starting way up the curve.

Just a thought.....
 
Last edited:
Been busy the last few weeks and still trouble shooting problems with this in my free time. I thought I had it "fixed," I was wrong.

Finally got a working system, now just need time to put it all back together.

In the end it was a faulty ZVS board that was the source of my issues.

1st ZVS worked half assed and died. I got it to "anneal" a few times but it never really got hot.
2nd ZVS came DOA.
3rd ZVS works.

For anyone that sees this in the future, the 2 shit ZVS boards were purchased off Amazon and had square heatsinks. The good ZVS board that actually works has heavy duty angled heatsinks and the screws on the bottom that hold the brass poles are soldered in place. It's a better quality construction board. 1000W ZVS

Origional chinese 600w PSU works.

Luckly Amazon has a good return policy.

Next time I update will be with the working system back together.

Thought I was going to make a few of these for friends. Think I have changed my mind and it is more trouble than it is worth with all the low quality chinese electronics. 2 shit ZVS's and 1 shit SSR.
 
I’ve gone through three ZVS boards two SSR’s and one power supply.
Now with an 1800 watt ZvS, SSR from Itchy and a Meanwell power supply it’s pretty solid. Probably should’ve bought an Amp
 
No need for a big SSR. Just control the two that already exist on the ZVS board (heat-sinked and all) with an iddy biddy switch.
 

Upgrades & Donations

This Forum's expenses are primarily paid by member contributions. You can upgrade your Forum membership in seconds. Gold and Silver members get unlimited FREE classifieds for one year. Gold members can upload custom avatars.


Click Upgrade Membership Button ABOVE to get Gold or Silver Status.

You can also donate any amount, large or small, with the button below. Include your Forum Name in the PayPal Notes field.


To DONATE by CHECK, or make a recurring donation, CLICK HERE to learn how.

Forum statistics

Threads
165,437
Messages
2,194,912
Members
78,882
Latest member
FIDI_G
Back
Top