Homemade Induction Annealer

T

TSloper

Private
Minuteman
Oct 1, 2019
76
142
Over the last couple of days I put together an induction annealer setup. All components minus the coil were sourced from Amazon. I turned the coil myself from solid copper wire to ensure proper inductance and distance to the case. The power supply is 24V/15A. Switching frequency of the induction coil is ~120 kHz. The coil supplied with the ZVS unit was far too large in diameter and would not sufficiently heat the brass with the 24V/15A supply. Proper coil sizing is key. I have not found a need for any additional cooling. The PCB and components are barely getting warm.

The 1/8" rod under the coil is the release mechanism to allow the case to drop. I am contemplating adding a push/pull solenoid to automate the drop at the end of the cycle but at my processing rate it isn't necessary.

Using 750F Tempilaq as an indicator my run time for a .223 case is 3.4 seconds. A tenth of a second can make a significant difference. The picture showing both cases is only a delta of 0.2 seconds. A run time of just 7.5 seconds will have the brass glowing red hot as seen in the pic. That is far too hot for annealing but I provided the pic to illustrate that it doesn't take long to over anneal. Average power consumption during a cycle with the current setup is ~260W. With a different power supply and additional cooling the system can run to 1000W.

This is a sub $100 build and the results are consistent and repeatable.

Tim
 

Attachments

  • IMG_0167.jpg
    383 KB · Views: 5,100
  • IMG_0166.jpg
    336.5 KB · Views: 4,903
  • IMG_0171.jpg
    321.3 KB · Views: 5,092
  • IMG_0173.jpg
    266.7 KB · Views: 5,583
T

TSloper

Private
Minuteman
Oct 1, 2019
76
142
Here are the primary components...

ZVS Board ($36.62) - https://www.amazon.com/gp/product/B01C70G7Y8/ref=ppx_yo_dt_b_asin_title_o08_s00?ie=UTF8&psc=1

5-36V Timer/Relay Board ($12.99) - https://www.amazon.com/gp/product/B07V24WJ4S/ref=ppx_yo_dt_b_asin_title_o05_s00?ie=UTF8&psc=1

24V/15A Power Supply ($21.99) - https://www.amazon.com/gp/product/B0196PXMTU/ref=ppx_yo_dt_b_asin_title_o07_s00?ie=UTF8&psc=1

Momentary Switch ($8.99) - https://www.amazon.com/gp/product/B06XSBYNM7/ref=ppx_yo_dt_b_asin_title_o04_s00?ie=UTF8&psc=1

750F Tempilaq ($30.49) - https://www.amazon.com/gp/product/B00PL7SEUU/ref=ppx_yo_dt_b_asin_title_o06_s00?ie=UTF8&psc=1

Copper Wire ($17.87) - https://www.lowes.com/pd/Southwire-25-ft-8-Gauge-Soft-drawn-Copper-Bare-Wire-By-the-Roll/50372842

Cord for power supply - Lowes carries a wall plug cord that is has the L/N/GND wires exposed. Just go to where they sell extension cords and you will find it.

Feel free to ask questions. The coil winding is key. My coil is 10 turns with an ID of 15-16mm. I wouldn't go more than 20mm if you adhere to the setup above. If you aren't sure your coil dimensions/turns will work message me and I can tell you if it will work acceptably. I will need to know coil wire diameter, coil turns, and ID. The opening of the case mouth is positioned to be right about mid way in the coil.

Other timer boards can be used but make sure it supports the voltage/current needs. The timer board I used is using two N-channel FETs in parallel rather than traditional relays. It is working very well. Other ZVS boards can be used as well but you will have to take note of how much capacitance the board has when deciding the coil specs. I'm an EE btw so this project was right up my alley.

Quick tip on turning the coil with that particular wire... use a wooden dowel or similar and turn the coil as tight as possible. Once it is turned take 5-6 business cards and slide them through the coil windings. This will then give you uniform spacing between the coils. Just make sure none of the coils touch and you will be fine. At the mounting ends of the coil I just soldered the 8 AWG wire into the center of a couple pieces of 1/4" copper tubing. This allowed the coil to mount as intended with this particular ZVS board.

If you opt to mimic the wooden block approach the proper drill bit size for 223 is 13/32". You want just enough clearance that the case will drop freely but not tilt excessively. I made the base board hole under that 1/2". Everyone will have their own approaches but just make sure you dimension the case support such that the case mouth opening is about mid way in the coil.

Tim
 
Last edited:
T

Tx_Aggie

Gig 'em
Supporter
Full Member
Minuteman
Sep 3, 2017
846
818
VA
Welcome to The Hide!!

That has to be one of the best first posts I've seen in a long time (on any forum).

I think I just found one of my reloading room projects for the winter. Thank you for posting this!
 
Last edited:

Ryridesmotox

Hood rat
Full Member
Minuteman
Mar 15, 2019
814
505
Are you wiring the momentary to power everything on and then the timer shuts the power to the coil off? I'm pretty bad with understanding electrical crap.
 
T

TSloper

Private
Minuteman
Oct 1, 2019
76
142
Are you wiring the momentary to power everything on and then the timer shuts the power to the coil off? I'm pretty bad with understanding electrical crap.

Close. When plugged in the power supply and timer board turn on. There is no power going to the ZVS board at this point. The momentary switch triggers the timer board at which point it turns on the MOSFETs (think of them as relays) and it begins counting down. Once it reaches 0 the MOSFETs are turned off cutting power to the ZVS board. This timer board has a resolution of 0.1 seconds. That is critical as 0.1 seconds can matter.

The trigger terminals on the timer board are TRG and TRG_GND. 24V goes from the power supply into the momentary switch. The output of the momentary switch goes to TRG of the timer board. GND from the power supply connects to TRG_GND. When you press the button it triggers the timer. The tigger input is isolated from the rest of the timer board via an opto-coupler. This is a good thing.

Tim
 
Last edited:

Ryridesmotox

Hood rat
Full Member
Minuteman
Mar 15, 2019
814
505
Close. When plugged in the power supply and timer board turn on. There is no power going to the ZVS board at this point. The momentary switch triggers the timer board at which point it turns on the MOSFETs (think of them as relays) and it begins counting down. Once it reaches 0 the MOSFETs are turned off cutting power to the ZVS board is removed. This timer board has a resolution of 0.1 seconds. That is critical as 0.1 seconds can matter.

Tim

Oh I see. Seems simple enough. Much easier to control than the torch with a drill
 
T

TSloper

Private
Minuteman
Oct 1, 2019
76
142
Oh I see. Seems simple enough. Much easier to control than the torch with a drill

Maintaining consistency with the torch and drill approach is what I see as its main deficiency. Flame placement will vary and time will definitely vary. But it has worked for many for a very long time.
 
Reactions: Ryridesmotox
T

TSloper

Private
Minuteman
Oct 1, 2019
76
142
I have plently of the 8 AWG wire left. I wouldn't have any issue making up some coils for those that don't want to do it themselves. Just cover the shipping cost. The coil dimensions should work just fine for .308 and 6.5.
 

whatsupdoc

Major Hide Member
Full Member
Minuteman
  • Dec 12, 2017
    2,110
    2,698
    @TSloper couple of questions.

    Have you tried 308 or larger cases with this setup is 360 watts enough?
    Number of turns for the copper coil?
     

    whatsupdoc

    Major Hide Member
    Full Member
    Minuteman
  • Dec 12, 2017
    2,110
    2,698
    Thinking out loud here, as far as the aneal time goes I wonder if you used a sacrificial case, time it so that it just gets red hot then just back down the time say 2/10ths.
     
    T

    TSloper

    Private
    Minuteman
    Oct 1, 2019
    76
    142
    @TSloper couple of questions.

    Have you tried 308 or larger cases with this setup is 360 watts enough?
    Number of turns for the copper coil?

    I have not but it will just take a little bit longer over 223. The other option is to increase to a higher voltage power supply and calibrate the time appropriately.
     
    T

    TSloper

    Private
    Minuteman
    Oct 1, 2019
    76
    142
    @TSloper couple of questions.

    Have you tried 308 or larger cases with this setup is 360 watts enough?
    Number of turns for the copper coil?

    My coil is 10 turns. The coil specs are about matching the inductance of the coil to the ZVS capacitor bank while keeping the coil to case gap from being to large as you will loose efficiency there. Too little inductance and you will send the resonantly frequency too high.
     
    T

    TSloper

    Private
    Minuteman
    Oct 1, 2019
    76
    142
    Thinking out loud here, as far as the aneal time goes I wonder if you used a sacrificial case, time it so that it just gets red hot then just back down the time say 2/10ths.

    You could take that approach but you would still be left wondering how far the 750+ temps are conducting down the case body. Tempilaq in my opinion is the right approach. I am actually considering getting some 850F Tempilaq to see if it is getting that got at the mouth versus the extension of the 750 downwards. 850F is fine if it is it’s just another data point in dialing in the time.
     

    Ryridesmotox

    Hood rat
    Full Member
    Minuteman
    Mar 15, 2019
    814
    505
    I wonder if one could set up a trap door kinda thing to trigger as the timer ends or something. So the case dros out and you can drop the next and start over.
     
    T

    TSloper

    Private
    Minuteman
    Oct 1, 2019
    76
    142
    I wonder if one could set up a trap door kinda thing to trigger as the timer ends or something. So the case dros out and you can drop the next and start over.

    My setup will do that as is. I just need to attach a push pull solenoid to the 1/8” rod and add another timer module or switch to a different timer with two relays. I put the hooks in upfront for that.
     

    Ryridesmotox

    Hood rat
    Full Member
    Minuteman
    Mar 15, 2019
    814
    505
    My setup will do that as is. I just need to attach a push pull solenoid to the 1/8” rod and add another timer module or switch to a different timer with two relays. I put the hooks in upfront for that.

    Something like this?

    NOYITO 2-Channel Programmable Time Relay Module Dual Relays Triggered In turn or individually 2-Channel 0-99.9V Voltage Detection Control Board Cyclic Timing On Off (24V) https://www.amazon.com/dp/B07X7RK61V/ref=cm_sw_r_cp_apa_i_VymRDbGN72DND
     
    B

    Bad_Donkey

    Private
    Minuteman
    Jul 1, 2019
    39
    28
    I'm in the process of building one right now. I would link to the thread, but I'm not sure if linking to another forum is allowed. If you go to Accurate Shooter forum and search for "Induction Brass Annealer Redux" there is a 90+ page thread about how to build one with parts lists.
     

    Ryridesmotox

    Hood rat
    Full Member
    Minuteman
    Mar 15, 2019
    814
    505
    I'm in the process of building one right now. I would link to the thread, but I'm not sure if linking to another forum is allowed. If you go to Accurate Shooter forum and search for "Induction Brass Annealer Redux" there is a 90+ page thread about how to build one with parts lists.

    I just looked, that's a cool setup. A bit too complex for me right now I think.
     
    T

    TSloper

    Private
    Minuteman
    Oct 1, 2019
    76
    142
    B

    Bradv86

    Sergeant of the Hide
    Full Member
    Minuteman
    May 18, 2019
    437
    285
    TX
    I have done quite similar using the same ZVS board. I threw all the guts into a small ATX computer case with the intent of automating a case feed and auto drop but for now I am single loading it. I did opt to water cool the coil, however with the 1/4" OD tubing getting a tight enough bend radius without pinching the tubing isn't possible with the tooling I have. My power supply is 48V 12.5A and the board does get hot, however I think that is likely a result of the coil being slightly too large at 25mm ID. I plan to step down in tubing size to get a tighter bend closer to the size you mentioned.

    With my current coil it will do .308 FGMM brass in 8 seconds using 750F tempilaq. I want to drop this time down with a more efficient coil to reduce temps lower in the body.

    I am not an EE, and just love tinkering with these types of things, so excuse my crude MS Paint schematic of my setup.
     
    Last edited:
    T

    TSloper

    Private
    Minuteman
    Oct 1, 2019
    76
    142
    I have done quite similar using the same ZVS board. I threw all the guts into a small ATX computer case with the intent of automating a case feed and auto drop but for now I am single loading it. I did opt to water cool the coil, however with the 1/4" OD tubing getting a tight enough bend radius without pinching the tubing isn't possible with the tooling I have. My power supply is 48V 12.5A and the board does get hot, however I think that is likely a result of the coil being slightly too large at 25mm ID. I plan to step down in tubing size to get a tighter bend closer to the size you mentioned.

    With my current coil it will do .308 FGMM brass in 8 seconds using 750F tempilaq. I want to drop this time down with a more efficient coil to reduce temps lower in the body.

    I am not an EE, and just love tinkering with these types of things, so excuse my crude MS Paint schematic of my setup.
    View attachment 7167896

    I agree that the coil ID is causing inefficiency here. You have double the power supply I do. Can you tell me what the capacitors are rated to on your ZVS board (ie. .33 uF x 6)? I can tell you what coil dimensions, turns, etc will work well.
     
    T

    TSloper

    Private
    Minuteman
    Oct 1, 2019
    76
    142
    So I've spent a lot of time plowing through annealing data for C26000 brass (ie. cartridge brass) that I've found online. AMP's data was the most useful. 750F has been the target for most hobbiest annealing efforts but I do not believe getting to 750F and stopping is adequate at the neck. I have some 900F and 1000F Tempilaq on order. My conclusion based on everything I've read is that unless you are getting close to 1000F at the neck momentarily you will not reduce the hardness to a meaningful amount in that area.

    So what I hope to verify is that I can momentarily hit 1000F at the neck and not show 750F extending more than a few mm below the shoulder. If I can't hit 1000F without pushing the 750F down too far I will have to increase the supply power. I should have this tested Wednesday or Thursday. i think I will be ok with the supply I have for 223 brass.

    Tim
     
    Last edited:
    B

    Bradv86

    Sergeant of the Hide
    Full Member
    Minuteman
    May 18, 2019
    437
    285
    TX
    They are in fact .33uF x 6.


    Although a different brand, I imagine that these likely come out of the same factory.

    I am hoping to retain the water cooling, but if getting 1/8" od or 3/16" od copper tube bent properly becomes an issue I'll be open to solid wire.
     
    T

    TSloper

    Private
    Minuteman
    Oct 1, 2019
    76
    142
    They are in fact .33uF x 6.


    Although a different brand, I imagine that these likely come out of the same factory.

    I am hoping to retain the water cooling, but if getting 1/8" od or 3/16" od copper tube bent properly becomes an issue I'll be open to solid wire.

    Ok cool. So the goal is to wind the coil to achieve two things... 1) inductance that matches the capacitance to achieve the proper switching frequency range and 2) keeping the distance between the object and coil ID minimized. I went with an ID on mine that would still allow me to go to larger cases (308, etc) than 223 should the need arise. 1/4" tubing is definitely challenging with a short turn radius. 1/8" tubing would be far better in my opinion. I like the 8 AWG solid wire as it's mass is just so much greater than the tubing and while it may warm up it isn't getting hot enough to need any cooling. I can still touch it.

    I remeasured my coil... ID = 0.600 in, L = 1.55 in, 10 turns (8 AWG)

    If you go with ID = 0.750, L = 1.55 in, 9 turns (8 AWG) you will end up around a switching frequency of ~121 kHz which is right where you want to be. I picked an ID of 0.750 just because it should be easy to find a dowel that size to turn around. If you can get closer to my ID of 0.600 in go to 10 turns.

    Now if you were going to be trying to melt something in a graphite crucible or run fully automated from feed to drop then water cooling of the coil absolutely is warranted but where the duty cycle is low (manual feed) and solid 8 AWG copper is being used I just don't see the need.

    Tim
     
    B

    Bradv86

    Sergeant of the Hide
    Full Member
    Minuteman
    May 18, 2019
    437
    285
    TX
    Great information, thank you!

    The water cooling was originally intended for when I add in an automatic feed and drop system.

    I will build a .6" ID solid 8awg coil and check how warm the components get using the additional power over ~50 cases as fast as I can. I am sure the more efficient coil in addition to shorter run / heating times will keep the board far cooler than what I currently am experiencing. If I do get an auto feed system built I will simply copy the coil from 1/8" tubing.

    Do you see any benefit to a more robust mounting system for the coil itself? I suppose the three upright posts are sufficient for energy transfer but would there be any benefit at all to something like a solid piece of copper? The upright posts are the hottest component on my board and will absolutely burn skin.
     
    T

    TSloper

    Private
    Minuteman
    Oct 1, 2019
    76
    142
    Great information, thank you!

    The water cooling was originally intended for when I add in an automatic feed and drop system.

    I will build a .6" ID solid 8awg coil and check how warm the components get using the additional power over ~50 cases as fast as I can. I am sure the more efficient coil in addition to shorter run / heating times will keep the board far cooler than what I currently am experiencing. If I do get an auto feed system built I will simply copy the coil from 1/8" tubing.

    Do you see any benefit to a more robust mounting system for the coil itself? I suppose the three upright posts are sufficient for energy transfer but would there be any benefit at all to something like a solid piece of copper? The upright posts are the hottest component on my board and will absolutely burn skin.

    My suspicion is that the posts are acting as heatsinks for the PCB that is heating up (assuming your water cooling is active). I think that once you optimize the coil you will see temps decrease all around and it won't be a concern. 8 seconds seems like a very long run time with the supply you have. When I originally tried the included coil I could not get the Tempilaq to melt on 223. As soon as I switched coils it went to 3.4 seconds. In 7-8 seconds the neck/shoulder is glowing red hot. At 12 seconds it is lighting up the room lol.

    By the end of the week I will have this completely figured out but I think very rapid heating of the neck/shoulder is the ticket. The longer it takes the more conduction down the case body will happen which is what we do not want. My target will be to get to about 1000F quick enough at the neck to prevent any over temp of the case body a few mm below the shoulder. Brass cools incredibly fast so I don't see this being a problem.

    Also don't be afraid to vary the placement of the brass in the coil to find the sweet spot. The magnetic field center may or may not align with the geometric center in the vertical axis. A few mm could make a huge difference in strength.
     
    Last edited:
    T

    TSloper

    Private
    Minuteman
    Oct 1, 2019
    76
    142
    Is there anyone with a seating force gauge that would be willing to test seating force of some 223 samples? New brass, fired (not annealed), fired (annealed)? I would supply the cases ready to go and bullets. This would be an indicator that the annealing is having a positive effect.
     
    B

    Bradv86

    Sergeant of the Hide
    Full Member
    Minuteman
    May 18, 2019
    437
    285
    TX
    I have noted huge differences in the speed different headstamp cases heat up. For example, R-P .260 Remington brass is glowing red hot in the same time that Lapua .260 brass melts the 750 tempilaq on the neck alone. This may be mostly a function of the neck wall thickness, but I imagine the brass is not exactly identical in composition either. For this reason I am creating a log book of my anneal times by brass headstamps.

    I agree that 8 seconds is a long time, and it concerns me about heat creeping further down the case than intended. Larger cases heat much more efficiently than the smaller ones as well. .300 Wby cases are also red hot in the same time that the .308 FGMM brass takes to melt the tempilaq. All of my blind messing around has given me the conclusion that my coil is simply improperly sized and inefficient. Now with the right direction I expect to drop the annealing time and heat output down significantly.

    From melting lead out of copper jacketed bullets to red hot screwdrivers to overheating cases before I got my tempilaq this is a great project to play with.
     
    T

    TSloper

    Private
    Minuteman
    Oct 1, 2019
    76
    142
    I have noted huge differences in the speed different headstamp cases heat up. For example, R-P .260 Remington brass is glowing red hot in the same time that Lapua .260 brass melts the 750 tempilaq on the neck alone. This may be mostly a function of the neck wall thickness, but I imagine the brass is not exactly identical in composition either. For this reason I am creating a log book of my anneal times by brass headstamps.

    I agree that 8 seconds is a long time, and it concerns me about heat creeping further down the case than intended. Larger cases heat much more efficiently than the smaller ones as well. .300 Wby cases are also red hot in the same time that the .308 FGMM brass takes to melt the tempilaq. All of my blind messing around has given me the conclusion that my coil is simply improperly sized and inefficient. Now with the right direction I expect to drop the annealing time and heat output down significantly.

    From melting lead out of copper jacketed bullets to red hot screwdrivers to overheating cases before I got my tempilaq this is a great project to play with.

    I will be doing the same... logging each headstamp and time once I get things dialed in. AMP did essentially the same thing before Aztec came along. This is a fun project for sure. Not knocking the drill/torch approach but I do not see how you can achieve consistency with that method. The oxidation marks tell very little about what happened in the metal. The automated flame approaches I see as viable because you can maintain consistency. Induction heating seems to be the best approach as you can heat much quicker and control the duration of the heating very tightly.

    Tim
     

    mcfred

    Sergeant
    Full Member
    Minuteman
    Feb 17, 2011
    1,314
    594
    SW USA
    I've got a DIY induction annealer similar to the OP's. Case wall thickness and ambient temperatures affect how fast a case will come up to temperature. The 24v/15a power supply has a potentiometer left of the DC power bus that will get you a few more volts. I have one that has a range of 19v-30v). I turn it down for .223 and smaller and up for WSM and larger.

    Mixed headstamps and Summer vs Winter garage temperatures will have you chasing your tail.
     
    Reactions: D_TROS

    D_TROS

    Flag-Sword-Cross
    Full Member
    Minuteman
  • Aug 19, 2010
    2,408
    2,486
    North Denver, CO
    tagging in. def interested in this project as my broke ace isn't going to get a $1500 AMP anytime soon...


    DT
     
    Reactions: WM5L
    T

    TSloper

    Private
    Minuteman
    Oct 1, 2019
    76
    142
    I took some more accurate measurements at 24.05 V and 28.20 V. The voltage and current measurements were taken at the ZVS board terminals. 4 V turned out to be pretty significant. i used the same lot of brass trimmed to the same length (1.750") for these tests. Until I receive my 900F and 1000F Tempilaq I consider the 750F Tempilaq change 4mm below the shoulder as my optimum baseline.

    @24.05V

    3.4 s (baseline) -> 11.40 A or 274 W

    @28.20 V

    2.8 s (new baseline) -> 12.75 A or 360W
    3.0 s -> 13.00 A or 367 W
    3.4 s -> 13.50 A or 381 W

    As you can see a 4 V increase dropped the time from 3.4 s to 2.8 s. If I find I need a little more power with the additional Tempilaq temps I will move to the 36 V/15 A version of this supply. Current will rise as the brass heats up. My current figures here are peak measured during the cycle. If you are going to replicate this setup I would probably advise just going with a 36 V/ 20A supply right out of the gate.

    Tim
     
    Last edited:
    A

    Archwarder

    Private
    Minuteman
    May 26, 2019
    12
    15
    AMP stopped doing anything temperature related years ago when they introduced hardness testing as their means of validating their results. The large variations in times different cases take to heat up using a fixed power output with time control only make one single target temperature impossible to use as an annealing target.
     
    T

    TSloper

    Private
    Minuteman
    Oct 1, 2019
    76
    142
    AMP stopped doing anything temperature related years ago when they introduced hardness testing as their means of validating their results. The large variations in times different cases take to heat up using a fixed power output with time control only make one single target temperature impossible to use as an annealing target.

    Yes AMP is using hardness testing to validate their algorithms but at the end they are striving to achieve a temperature vs time profile for a given lot of brass. They can only control temperature ramp rate, peak temperature, and time as it pertains to what the brass experiences. I checked today with a lab and they wanted $1K to test some samples. For that money I would just buy an AMP unit with Aztec mode

    At our lowly hobby level without lab testing means the best we can do is use tools like Tempilaq to fine tune the duration and peak temps at various locations on the brass and then use an arbor press with force guage and range results to infer effectiveness. AMP has clearly put quality science behind what they have done. I'm an engineer so I fully respect everything they have done but most don't have $1200 - $1300 to blow on a machine that honestly will get under utilized. Home induction setups like myself and others have implemented will work it just takes a bit more attention to the finer details to product the desired results. If AMP could find a way to get the cost closer to the $600-$700 range I think they would knock it out of the park.
     
    Reactions: goneballistic
    T

    TSloper

    Private
    Minuteman
    Oct 1, 2019
    76
    142
    The 900F and 1000F Tempilaq arrived today. Working with 750F alone does not give a complete picture as to what is happening at the neck and shoulder area on a setup like this.

    Below are pictures of my LC 18 223 brass trimmed to 1.750". This is a factory new piece so it still shows the factory annealing marks. The green is 750F, white is 900F, and pink is 1000F. Dialing in at 2.9 seconds (28.2V) you will see that the neck and shoulder are hitting 1000F extending about 1mm down from the shoulder base, 900F is extending about 4mm below the shoulder, and 750F is extending about 6mm below the shoulder. The bottom of the shoulder is changing the 1000F Tempilaq just as the timer stops.

    Correlating to the AMP data published on the web I would consider 2.9 seconds to be about perfect. At 2.8 seconds the very bottom of the shoulder (.5 mm or so) does not change at 1000F.

    Regardless of whatever method you are using you have to ensure the neck/shoulder is getting up to around the 1000F mark or you will not anneal sufficiently. Any glow is going to be too hot. This also has to be balanced with not letting temps close to 1000F extend down the case body. 750F - 900F momentarily is not going to be hot enough to have any appreciable change in hardness if you trust what AMP has published (I personally do). I hope this test helps the rest of you out.

    Tim
     

    Attachments

    • IMG_0187.jpg
      312.3 KB · Views: 626
    • IMG_0188.jpg
      304.5 KB · Views: 610
    • IMG_0189.jpg
      304.5 KB · Views: 588
    • IMG_0190.jpg
      280.6 KB · Views: 600
    • IMG_0191.jpg
      304.4 KB · Views: 611
    • IMG_0192.jpg
      276.4 KB · Views: 620
    Last edited:
    B

    Bradv86

    Sergeant of the Hide
    Full Member
    Minuteman
    May 18, 2019
    437
    285
    TX
    Nice results.

    It would be interesting to run the same brass in the AMP and see how the 3 tempilaq values compare to what you have there.
     
    Reactions: Snuby642
    T

    TSloper

    Private
    Minuteman
    Oct 1, 2019
    76
    142
    Nice results.

    It would be interesting to run the same brass in the AMP and see how the 3 tempilaq values compare to what you have there.

    I agree! They have far more to account for in accommodating various case sizes vs coil size. Their test data showed 3.4 seconds to reach 1000F on a piece of Norma 223 brass. From some tear downs I saw their coil diameter is larger (as expected) so the brass to coil gap is larger on 223. They do have more output power capability so that is a variable too. I’m happy with where my setup has ended up and would expect favorable results with where it sits for my 223 brass anyways.
     
    Last edited:

    Snuby642

    Two Star General
    Full Member
    Minuteman
  • Feb 11, 2017
    8,299
    10,736
    I worked around EE's and a lot of posers in the fighter jet industry for a long time

    I had to make things work.

    Refreshing to read posts from someone that knows thier trade.

    @TSloper

    Thanks
     
    T

    TSloper

    Private
    Minuteman
    Oct 1, 2019
    76
    142
    I worked around EE's and a lot of posers in the fighter jet industry for a long time

    I had to make things work.

    Refreshing to read posts from someone that knows thier trade.

    @TSloper

    Thanks

    I appreciate that. I am no expert but I do find this project very interesting and fun. Thank you!
     
    Reactions: laserlite
    T

    TSloper

    Private
    Minuteman
    Oct 1, 2019
    76
    142
    So some learnings from just messing around with the setup and some brass...

    1. The solid wire coil is great for controlling distance and inductance. If you can keep your case rate to 4-5 per minute you should have no problems thermally. In full disclosure I am going to mount a 120mm 24V fan tomorrow in front of the ZVS board so airflow is blowing from the capacitor side to the aluminum heatsink side. This will help dissipate heat that conducts from the coil through the standoffs into the PCB around the capacitors. The MOSFETs stay cool as is. It will also help dissipate heat from the coil itself as it passes over the straight sections. I did not encounter any significantly high temps during my 25 piece run but keeping heat minimized is always good. A tubular coil is great if you need to run volume as fast as possible but it will be harder to turn the optimum coil dimensions and will increase the need for more power supply. If I can achieve a good tubular coil with water cooling capability I will post the process.

    2. In complete darkness, the neck will have a very subtle glow about 0.2 seconds before the timer shuts off. 0.1 seconds less still exhibits the momentary glow but the 1000F Tempilaq does not change state that last 0.5 mm or so. I'm going to leave the cycle time at 2.9 seconds and achieve the complete Tempilaq change down to the shoulder base. I am not at all worried about the case body temperature at this cycle time.

    3. I was curious to see how over temping the brass would cause it to behave. I took 3 pieces and ran at 1 cycle, 2 cycles back to back, and 3 cycles back to back (glowing a very bright yellow). I put them in my small vice with the case head on one jaw and the mouth on the other. The 1 cycle piece did not deform in any way with significant pressure. It was holding up as well as a non-annealed piece of brass. The 2 cycle piece formed a ridge right at the base of the shoulder with moderate pressure. The 3 cycle piece folded completely at the neck/shoulder area very quickly. Over annealing is absolutely a dangerous condition. I have to reiterate that using only 750F Tempilaq is not sufficient to know where your neck and shoulder are actually getting to. If you are going to anneal spend the $ and get the proper Tempilaq (McMaster Carr is a good source) target temps of 750F and 1000F. You won't achieve proper annealing if you aren't reaching 1000F.

    I plan on hitting the range tomorrow and seeing how this brass behaves at 200m.

    Tim
     
    B

    Bradv86

    Sergeant of the Hide
    Full Member
    Minuteman
    May 18, 2019
    437
    285
    TX
    I found that making a coil using the sand/salt fill method did not allow for all of the filler to evacuate at this tight of a bend radius.

    Filling the tubing with water and crimping the ends over provided enough internal support to get 1/4" tubing to a 1" ID. I believe that 1/8" tubing done in the same way will likely not have much trouble getting to 0.6" ID when wrapped around a straight mandrel. If you happen to be able to get ahold of "coil rod" or a junked ballscrew of the proper size you could potentially have a very nice form tool to wrap it with. Perhaps a trapezoidal threaded rod of the appropriate size would also give you a nice guide to wrap with while having the tubing filled with water for internal support.

    Being that the project is budget based, a hardwood dowel rod could likely be made into a form tool as well with a little bit of time and patience.

    I won't be home to work on rebuilding my setup for probably a few weeks so I am eager to see what you come up with.
     
    N

    northern50

    Null User
    Full Member
    Minuteman
    Mar 22, 2006
    343
    142
    WI
    Looks good Tim!

    I used the same board and power supply. Chinesium parts from big rock candy mountain. Lol.

    Mine has been running for about 3 years with zero issues, I installed a prox to sense when a case enters the holder to start the cycle, and rotating trap door to drop the case once the cycle completes.

    As Tim said, coil size is key to proper operation, I messed with a couple coils before running the current version.
     

    Attachments

    • B4C19930-1D11-40B5-9EEF-2344F25ADD07.png
      3.1 MB · Views: 1,323
    T

    TSloper

    Private
    Minuteman
    Oct 1, 2019
    76
    142
    Looks good Tim!

    I used the same board and power supply. Chinesium parts from big rock candy mountain. Lol.

    Mine has been running for about 3 years with zero issues, I installed a prox to sense when a case enters the holder to start the cycle, and rotating trap door to drop the case once the cycle completes.

    As Tim said, coil size is key to proper operation, I messed with a couple coils before running the current version.

    Very nice work!
     
    T

    TSloper

    Private
    Minuteman
    Oct 1, 2019
    76
    142
    Some days I wish I didn't have an engineering mindset. LOL So at 2.9s (28V) the 1000F Tempilaq indicates I'm heating the neck/shoulder to at least 1000F uniformly. That's great but it did begin to glow for about 0.2 seconds (complete darkness) at the end of the cycle. I was unable to hit 1000F uniformly without glowing at the end. The glow at the level I am seeing is indicative of at least 1022F per brass incadescence charts. Without additional Tempilaq levels I have no idea exactly where it is getting to at that last moment. It could be closer to 1100F.

    I went back to 24V and it was coming up short on reaching 1000F quickly enough to avoid extending the 750F too far down the body. I then dialed in the supply to 26.05V (measured at the ZVS board inputs). 26.05V @ 12.5A (326W) at 3.5 seconds yields uniform 1000F at the neck/shoulder and it does NOT reach the point of glow in complete darkness. I view this as a worthy improvement over the 28V / 2.9s configuration. I also mounted a 24V 120mm fan. It wasn't absolutely necessary but why not? The moral of the story here is that, along with coil dimensions, time adjustments may not be the only other knob you need to turn. Supply voltage adjustments may also be necessary for a specific lot or caliber of brass.

    Here are the pics of the 26V run... (G = 750F, W = 900F, P = 1000F). Annealing marks are the factory anneal. This was a new piece of LC 18 brass trimmed to length.
     

    Attachments

    • IMG_0196.jpg
      308.2 KB · Views: 271
    • IMG_0197.jpg
      331.7 KB · Views: 245
    • IMG_0198.jpg
      313.4 KB · Views: 221
    • IMG_0199.jpg
      324.6 KB · Views: 223
    • IMG_0200.jpg
      286.2 KB · Views: 236
    • IMG_0201.jpg
      377.2 KB · Views: 258
    Last edited:
    T

    TSloper

    Private
    Minuteman
    Oct 1, 2019
    76
    142
    So I finally got around to making an equivalent coil out of 1/8” copper tubing. It is dimensionally and coil count identical to the 8 AWG coil. I obtained the copper from O’Reilly Auto. It is very soft and I had zero issues turning it. No kinks or flattening. I just took my time.

    I’m going to solder some 1/8” to 1/4” adapters then use the appropriate compression fittings ok the 1/4” side for connecting to the pump. Initially it will be a quick and dirty pump setup but once I’m satisfied it is working ok I’m going to order a fan/heat exchanger unit from Amazon ($18) and hook it up to a small water reservoir... all mounted to the board.


    The angle of the photo my make it appear they are different dimensions but they are the same.

    Tim
     

    Attachments

    • 067EECC0-CBDC-4E6B-8AF4-D02F258ADEE1.jpeg
      311.9 KB · Views: 668
    • 91A2A0D0-76D8-4C3A-AD39-416E755792E9.jpeg
      277.5 KB · Views: 644
    induction case annealing machine