DIY induction guns? (warning:long)

[Michelle_]

My impression was that the ring down waveforms are necessary to generate an offset three phase signal. This is why I thought they were using thyristors.

As paper shows, they are using thyristors only in trigger circuit.  Main switches are ignitrons (gas discharge tubes) that can handle much higher dI/dt.  Ignitrons are also slow to turn off (like thyristors), forcing the ring-down waveform.  Probably not optimum from a theoretical perspective, but necessary due to lack of other suitable switching technology.

Might be a good design to scale given their actual achieved velocities are in the range of your interest (100 to 250m/s).  Finding suitable switches will be a key difficulty, however.  You could try TRIAC arrays.  First stage only achieved 100m/s if I'm reading correctly.  First stage scales to under 600V so allow using a parallel-only array.  Might have better luck with that.  Requires simultaneous triggering of perhaps 100 devices per phase.  IGBTs might best option.  Depends on the money you are willing to spend.  Mechanical switching works for single stage and phase (disk launcher).  Triggered spark gaps (similar to ignitrons) might work if impedance is scaled higher (higher voltage and inductance and lower capacitance relative to simple scaled design).  Actually, that's probably what I'd try first if were designing this.  Though a simple single-coil pushing design might achieve the same velocity for less work.

Thanks, cost isn't a major concern; I would prefer to use the best possible switching method. Unfortunately at this point there's a big gap in my knowledge still. I'll look into the options you proposed and maybe do some asking around. Is there anyy reason not to use an ignitron?

I think your pusher design is good and the spark gap idea is good. In fact I found something you might be interested in:

https://www.rapp-instruments.de/index6.htm

Click on induktion gun 2. This seems more like what you're describing. Admittedly maybe I should look more into this as an alternative but I would still really like to build the polyphase style launcher, if I can learn enough to pull it off.

I'm assuming the polyphase gun is more efficient, which I don't know for a fact.

[davekni]

Is there anyy reason not to use an ignitron?

I don't know enough about ignitrons to say.  If you can find ones for your needed voltage and current after scaling, probably fine.  I don't think ignitrons are common devices around today, but don't really know.

I think your pusher design is good and the spark gap idea is good. In fact I found something you might be interested in:

https://www.rapp-instruments.de/index6.htm

Click on induktion gun 2. This seems more like what you're describing. Admittedly maybe I should look more into this as an alternative but I would still really like to build the polyphase style launcher, if I can learn enough to pull it off.

Yes, from a quick look it is probably a 6-stage pusher design.  (I learned some German 50 years ago.  Was never fluent, and lost most of what I knew.  Not taking time to run it through Google translate.)

I'm assuming the polyphase gun is more efficient, which I don't know for a fact.

I think it is for projectiles with high (normal) aspect ratio (longer than diameter).  Also most useful with multi-stage launchers where projectile current builds up, though I haven't thought through this bit in detail.

[Michelle_]

Is it safe to use the projectile to trigger spark gaps?

Anyhow their first pusher gun apparently accelerated a 7.8g to 100m/s using only 600J (6.5% efficiency)* they used an odd projectile and these results may be skewed

2nd pusher something like 4500J and 300m/s (my estimated) 4% efficiency? assuming 2mm wall projectile

I might be willing to build a pusher just to see what they are capable of, I'm just worried it's a waste of time and I could leapfrog a pusher by doubling down on the polyphase.

[Michelle_]

So for barrels I’m looking at

Polycarbonate
Carbon fiber
Garolite G10 (can I use this?)

I’m also thinking about risk analysis and this is almost too dangerous imo, especially since thinking about holding it next to your face.

I don’t know what to do about this besides scale down to a small version which is still dangerous but I might have a shot at making shielding, maybe out of Kevlar composite, I don’t know yet.

For now I’m considering walking back some of my goals and starting with a maximum projectile mass of .25g-.5g and simulating a single coil discharge, as recommended.

Once I have some kind of specification I will investigate more specific switching options.

Ideally in my mind right now I would have a list of switching options and the range of frequency, current, etc… they are capable of.

I am more optimistic about mechanical and design factors that can increase efficiency based on a safer, more compact, and off the shelf electrical circuit.

Based on my understanding thyristors would be ideal besides their dIdT limitations, however if I can design around these limitations and still use them as a starting point I would be interested in doing so, even if this precludes building the polyphase drive.

Ignitions are hard to find, old, and very large; and I don’t intend to switch anywhere near the current as the experimental unit using them.

Apparently a Thyratron is a thing and could potentially be what I need. I just found out about these, a lot of them are old sketchy looking tubes but you can still get new or recently made ones, sometimes cheap on eBay.

SiC Thyristors apparently are a thing but it doesn’t sound like something you can go and buy.

Besides that I’m also looking at how an IGBT could work.

[davekni]

Is it safe to use the projectile to trigger spark gaps?

That's an interesting idea that might function well, though probably needs a ceramic launch tube to withstand spark heat.  Can't think of any new safety concerns, as I presume the entire launch assembly needs to be enclosed in something to prevent user contact.  Oh, another possible issue is metal vapor deposits on ceramic tube eventually creating a conductive path.  Spark gaps usually avoid having insulation immediately adjacent spark path.
Perhaps projectile could trigger a small spark that in turn triggers the real power-switching spark gap, perhaps through a trigger transformer.

Garolite G10 (can I use this?)

Looks fine to me.

I’m also thinking about risk analysis and this is almost too dangerous imo, especially since thinking about holding it next to your face.

I see three key risks:
1) High voltage shock hazard
2) Explosion of coils due to magnetic force (as happens every shot of my quarter shrinker)
3) Explosion due to unintended short circuit of charged caps.
For 1, proper insulated housing should mitigate this risk well.
For 2, I don't have enough mechanical skills or knowledge to say just what would be safe.
For 3, hearing loss might be the biggest risk based on personal experience.  I had some hearing loss for months (and likely never completely recovered) due to accidentally shorting a large electrolytic cap I was testing in front of me.  Beyond that, any shielding that can handle (2) is likely sufficient.

I don’t know what to do about this besides scale down to a small version which is still dangerous but I might have a shot at making shielding, maybe out of Kevlar composite, I don’t know yet.

Not my area of expertise, but I'd think you want a housing optimized for impact strength such as polycarbonate.  Kevlar is strong, but not stretchy, so can't absorb as much impact energy.  Would be a good material around coils to handle expansion force.  There you do not want any stretch as stretch would allow coil to deform.  UHMWPE is strong and I think a bit better than Kevlar for impact strength.  Not sure.

For now I’m considering walking back some of my goals and starting with a maximum projectile mass of .25g-.5g and simulating a single coil discharge, as recommended.

Tiny is hard.  Efficiency doesn't scale well even if gaps could be scaled down.  Larger coil to projectile gap (proportionately) makes efficiency even worse.

Especially for your smaller (but I'd stay above 1g if I were you) design especially, I think IGBTs are the optimum choice.  I looked up a few parts for you using DigiKey search.  For best price on any of these, check oemstrade.com.
IKY75N120CS6,  1200V 300A peak.  Cheapest option for initial experiments.
IXGK75N170,      1700V 580A peak.
IXBF55N300,       3000V 600A peak.
IXBX50N360HV   3600V 420A peak.
FZ400R65KE3NOSA1 6500V 800A, very expensive option to use only if needing more than 3600V.
All except the final one are in TO247 variant packages, ~5mm thick.  Stacking an array of these in parallel is relatively easy presuming you have ability to design ECBs.

Finally, I've been thinking more about that "500m/s" linear motor paper that really achieved 150-250m/s.  My intuition now says that is not at all an efficient design worth copying.  Projectile is only 4 poles long, not even one full cycle of 6 poles, and it has only two stages.  I think linear motors will be more efficient only when projectile is at least one cycle long (6 poles of 3 phase) or much more and with more stages, at least 5 or 6.  Until that level, a pusher is likely more efficient.  Linear motors take time to start as magnetic field must penetrate projectile before thrust is generated.  Before that force is all radially inward (as in the projectile crushing issue in above paper).  Optimum starting requires low initial frequency of first stage that then ramps up in frequency as projectile moves.  On the other hand, a pusher provides maximum force initially as it relies on projectile blocking magnetic field.  For a pusher, frequency must be high enough that field reverses before too much has penetrated projectile.

[Michelle_]

Is it safe to use the projectile to trigger spark gaps?

That's an interesting idea that might function well, though probably needs a ceramic launch tube to withstand spark heat.  Can't think of any new safety concerns, as I presume the entire launch assembly needs to be enclosed in something to prevent user contact.  Oh, another possible issue is metal vapor deposits on ceramic tube eventually creating a conductive path.  Spark gaps usually avoid having insulation immediately adjacent spark path.
Perhaps projectile could trigger a small spark that in turn triggers the real power-switching spark gap, perhaps through a trigger transformer.

Garolite G10 (can I use this?)

Looks fine to me.

I’m also thinking about risk analysis and this is almost too dangerous imo, especially since thinking about holding it next to your face.

I see three key risks:
1) High voltage shock hazard
2) Explosion of coils due to magnetic force (as happens every shot of my quarter shrinker)
3) Explosion due to unintended short circuit of charged caps.
For 1, proper insulated housing should mitigate this risk well.
For 2, I don't have enough mechanical skills or knowledge to say just what would be safe.
For 3, hearing loss might be the biggest risk based on personal experience.  I had some hearing loss for months (and likely never completely recovered) due to accidentally shorting a large electrolytic cap I was testing in front of me.  Beyond that, any shielding that can handle (2) is likely sufficient.

I don’t know what to do about this besides scale down to a small version which is still dangerous but I might have a shot at making shielding, maybe out of Kevlar composite, I don’t know yet.

Not my area of expertise, but I'd think you want a housing optimized for impact strength such as polycarbonate.  Kevlar is strong, but not stretchy, so can't absorb as much impact energy.  Would be a good material around coils to handle expansion force.  There you do not want any stretch as stretch would allow coil to deform.  UHMWPE is strong and I think a bit better than Kevlar for impact strength.  Not sure.

For now I’m considering walking back some of my goals and starting with a maximum projectile mass of .25g-.5g and simulating a single coil discharge, as recommended.

Tiny is hard.  Efficiency doesn't scale well even if gaps could be scaled down.  Larger coil to projectile gap (proportionately) makes efficiency even worse.

Especially for your smaller (but I'd stay above 1g if I were you) design especially, I think IGBTs are the optimum choice.  I looked up a few parts for you using DigiKey search.  For best price on any of these, check oemstrade.com.
IKY75N120CS6,  1200V 300A peak.  Cheapest option for initial experiments.
IXGK75N170,      1700V 580A peak.
IXBF55N300,       3000V 600A peak.
IXBX50N360HV   3600V 420A peak.
FZ400R65KE3NOSA1 6500V 800A, very expensive option to use only if needing more than 3600V.
All except the final one are in TO247 variant packages, ~5mm thick.  Stacking an array of these in parallel is relatively easy presuming you have ability to design ECBs.

Finally, I've been thinking more about that "500m/s" linear motor paper that really achieved 150-250m/s.  My intuition now says that is not at all an efficient design worth copying.  Projectile is only 4 poles long, not even one full cycle of 6 poles, and it has only two stages.  I think linear motors will be more efficient only when projectile is at least one cycle long (6 poles of 3 phase) or much more and with more stages, at least 5 or 6.  Until that level, a pusher is likely more efficient.  Linear motors take time to start as magnetic field must penetrate projectile before thrust is generated.  Before that force is all radially inward (as in the projectile crushing issue in above paper).  Optimum starting requires low initial frequency of first stage that then ramps up in frequency as projectile moves.  On the other hand, a pusher provides maximum force initially as it relies on projectile blocking magnetic field.  For a pusher, frequency must be high enough that field reverses before too much has penetrated projectile.

Thanks for the risk ideas, that helps narrow things down. I didn't consider exploding coils. Yikes! I agree polycarbonate could make a good shielding material.

Thanks for the IGBT list and information. That's really helpful and is something I believe I can work with.

I have been doing some analysis on this 500m/s launcher. Apparently they based their design on a best practices from one of the older design papers that a lot of this stuff is based off of. The authors didn't directly state a lot of inportant information but I have derived the following:

Section 1 Total Energy Actual Max   18KJ@3.3KV
Section 2 Total energy Actual Max   36KJ@13.5KV
Total Gun Actual J   54,000
Designed Gun J   87,300
Maximum Projectile Speed Designed m/s   500
Maximum Projectile Speed Observed m/s   250
Projectile KE Designed J   16,750
Projectile KE Observed J   4187
Gun Efficiency Designed   19.18
Gun Efficiency Observed    7.754
Gun Efficiency Simulated (as built, not designed)   8.5

Stage 1 Max Velocity m/s   100
Stage 1 Max KE    670
Stage 1 Efficiency    4.46



They noted that their air gap was nearly double their initial design and apparently made some adjustments but their efficiency was ultimately limited by that and their projectile had a thinner wall than specified. So I'm not sure the design is fundamentally bad as it was based off of the best they had available I presume and theoretically would have been pretty efficient.

At any rate I had an idea about using a pusher for the first acceleration phase from V0 as to not waste an entire three phase stage on that acceleration. Then firing the pushed projectile into the polyphase bank.

Last thing right now is I found out about SiC MOSFETS which based on my ignorant preliminary research could work for me, but obviously I am behind the curve and could be off base. Supposedly you can use them in paralell without major current sharing concerns. These are a newer technology I believe but it appears one can actually find and purchase them.

Oh yeah and there's this gem which I wish I found a long time ago: "In the design of a linear induction launcher (LIL), the synchronous speed should ideally be higher than the desired projectile speed. This is because the slip, which is the difference between the synchronous speed and the actual speed of the projectile, represents the mechanism by which the launcher imparts kinetic energy to the projectile." the 500m/s launcher had a critical slip specified at 15% and 500m/s.

[davekni]

At any rate I had an idea about using a pusher for the first acceleration phase from V0 as to not waste an entire three phase stage on that acceleration. Then firing the pushed projectile into the polyphase bank.

Would get projectile moving faster initially.  However, does not provide for magnetic field penetration of projectile wall in the pattern used for LIM.  That might not matter in the simple two-stage version.  I'm not sure if that design matches phase well enough as projectile moves from first to second stage.  If not, initial magnetic field within projectile will not be useful to second stage.

Last thing right now is I found out about SiC MOSFETS

SiC FETs (MOS or junction) are used by high-end Tesla coil builders, especially for coils above ~300kHz.  FET pulsed current capability (including SiC) is much lower than for IGBTs, so not a good choice for your use.  BTW, the one SiC thyristor that theoretically can be purchased has fast turn-off, but not very high current (80A max) and is very expensive ($3000 each with minimum 10 order).

There are a few thyristors I've found that are designed for pulsed power, with a few that can be purchased:
Y500CNC250  50kA peak, 2500V, 11A/ns  ~$1200 from electronics distribution, but long lead time.
PT40QPx45    13kA peak, 4500V, 5A/ns    ~$100 on EBay
PT85QWx45  37kA peak, 4500V, 22A/ns   ~$700 on EBay

Oh yeah and there's this gem which I wish I found a long time ago: "In the design of a linear induction launcher (LIL), the synchronous speed should ideally be higher than the desired projectile speed. This is because the slip, which is the difference between the synchronous speed and the actual speed of the projectile, represents the mechanism by which the launcher imparts kinetic energy to the projectile." the 500m/s launcher had a critical slip specified at 15% and 500m/s.

This view works well once magnetic field penetrates projectile and presuming phase is matched from stage to stage.  At start speed is 0, so 15% of 0 is 0.  Starting is the complex part for LIM, but that's the entire length for short low-stage-count designs.  If each pole is it's own stage and frequency of each stage could change over time (instead of simple ring-down waveform), even short LIMs could theoretically be efficient (I think).  That way there could be smooth transitions from stage to stage.  Stages are shorter than projectile so could be optimized as projectile travels through stage.  At the higher velocity (later) end of a LIM, stages could be longer.  Duration within each stage is shorter at high speed.

In general, coil magnetic field wants to be some ideal phase ahead of projectile magnetic field.  Typically (for normal induction motors at least), maintaining that constant phase requires a constant slip speed, not a function of motor speed.  (Of course, most induction motors run at fixed speed, so it doesn't matter if slip is defined as a percentage or absolute speed.)  Perhaps that is why slip percentage is specified at exit velocity.  For high power LIMs, projectile will heat as it travels, increasing electrical resistance.  That will increase ideal slip for later stages.  Perhaps not a concern for your smaller designs.

[FPS]

For a 50mm diameter ~1mm thick aluminum disk, I've achieved 345m/s in a single stage.  However, coil was destroyed in the launch.  And speed was measured over the first 100mm of travel to minimize drag.

Why couldn't you slide the disk over a solid "barrel" made of ferrite or perhaps a metal rod cluster and use the accellerated disk as a form of sabot to drive a dart that is also slipped over the "barrel" in front of disk.

[Michelle_]

At any rate I had an idea about using a pusher for the first acceleration phase from V0 as to not waste an entire three phase stage on that acceleration. Then firing the pushed projectile into the polyphase bank.

Would get projectile moving faster initially.  However, does not provide for magnetic field penetration of projectile wall in the pattern used for LIM.  That might not matter in the simple two-stage version.  I'm not sure if that design matches phase well enough as projectile moves from first to second stage.  If not, initial magnetic field within projectile will not be useful to second stage.

Last thing right now is I found out about SiC MOSFETS

SiC FETs (MOS or junction) are used by high-end Tesla coil builders, especially for coils above ~300kHz.  FET pulsed current capability (including SiC) is much lower than for IGBTs, so not a good choice for your use.  BTW, the one SiC thyristor that theoretically can be purchased has fast turn-off, but not very high current (80A max) and is very expensive ($3000 each with minimum 10 order).

There are a few thyristors I've found that are designed for pulsed power, with a few that can be purchased:
Y500CNC250  50kA peak, 2500V, 11A/ns  ~$1200 from electronics distribution, but long lead time.
PT40QPx45    13kA peak, 4500V, 5A/ns    ~$100 on EBay
PT85QWx45  37kA peak, 4500V, 22A/ns   ~$700 on EBay

Oh yeah and there's this gem which I wish I found a long time ago: "In the design of a linear induction launcher (LIL), the synchronous speed should ideally be higher than the desired projectile speed. This is because the slip, which is the difference between the synchronous speed and the actual speed of the projectile, represents the mechanism by which the launcher imparts kinetic energy to the projectile." the 500m/s launcher had a critical slip specified at 15% and 500m/s.

This view works well once magnetic field penetrates projectile and presuming phase is matched from stage to stage.  At start speed is 0, so 15% of 0 is 0.  Starting is the complex part for LIM, but that's the entire length for short low-stage-count designs.  If each pole is it's own stage and frequency of each stage could change over time (instead of simple ring-down waveform), even short LIMs could theoretically be efficient (I think).  That way there could be smooth transitions from stage to stage.  Stages are shorter than projectile so could be optimized as projectile travels through stage.  At the higher velocity (later) end of a LIM, stages could be longer.  Duration within each stage is shorter at high speed.

In general, coil magnetic field wants to be some ideal phase ahead of projectile magnetic field.  Typically (for normal induction motors at least), maintaining that constant phase requires a constant slip speed, not a function of motor speed.  (Of course, most induction motors run at fixed speed, so it doesn't matter if slip is defined as a percentage or absolute speed.)  Perhaps that is why slip percentage is specified at exit velocity.  For high power LIMs, projectile will heat as it travels, increasing electrical resistance.  That will increase ideal slip for later stages.  Perhaps not a concern for your smaller designs.

Thanks for your thoughts and those part numbers for the thyristors!

Interesting update: I’ve come across A LOT more I formation:

40+ scientific papers including info on specific topics, design scaling, ranging up to present day and covering labs mainly in US, China, some in Israel

Internet resources outside of America: I have seen builds with hand held coil guns exceeding 100m/s. Let me tell you, the Russians are some crazy DIYers!

At this point I’m over saturated with information and need to let all of this sink in. The good news is I think I have more resources for estimating performance.

That said continuing to learn basics is key for me right now. I need to think about hands on practical skills to build anything safely before I try.

At this point I’m going to stick to researching for another 2-3 weeks then begin modeling and simulating whatever I’m going to build first.

I kinda want to update the thread with things I’ve learned but it’s gotten too far out of hand lol. If anyone reading this wants resources send me a message and I can help you find things that you probably won’t come across in a couple hours of googling.

There is 100x more information about inductive coil guns than I thought there was last week.

[davekni]

Glad you are finding lots of information.  Yes, I do find Russian posts or videos of very high power DIY projects, such as a 28kW RF Tesla coil making 1m+ high plasma flame running in a bedroom.  Fed by a radio transmitter tube.

Also, I realized I was making one key bad simplification in analyzing that "500m/s" paper.  I was calculating L/R time constant for the entire projectile in free space, not for each pole piece segment in close proximity to a copper coil.  Then I looked back at the paper and found I'd recalled a couple key parameters wrong too.  So I retract my thought about it being an inherently inefficient design.  Especially the initial theoretical design looks relatively good.  Once build realities changed coil gap and speed for their prototype, I'd still guess a couple other changes would be helpful to match.  Guessing a bit longer pole pitch to track gap and lower frequencies to track both pole pitch increase and speed reduction.  Might get a bit more speed out of real construction.

Have fun reading!  This has been an interesting conversation.  Intrigued me enough to think deeper than I'd planned.  Was a fun exercise in electromagnetics.  I'll be interested to learn what you end up designing.  In the mean time I'm focusing back on the projects I'm in the middle of.

[Michelle_]

Glad you are finding lots of information.  Yes, I do find Russian posts or videos of very high power DIY projects, such as a 28kW RF Tesla coil making 1m+ high plasma flame running in a bedroom.  Fed by a radio transmitter tube.

Also, I realized I was making one key bad simplification in analyzing that "500m/s" paper.  I was calculating L/R time constant for the entire projectile in free space, not for each pole piece segment in close proximity to a copper coil.  Then I looked back at the paper and found I'd recalled a couple key parameters wrong too.  So I retract my thought about it being an inherently inefficient design.  Especially the initial theoretical design looks relatively good.  Once build realities changed coil gap and speed for their prototype, I'd still guess a couple other changes would be helpful to match.  Guessing a bit longer pole pitch to track gap and lower frequencies to track both pole pitch increase and speed reduction.  Might get a bit more speed out of real construction.

Have fun reading!  This has been an interesting conversation.  Intrigued me enough to think deeper than I'd planned.  Was a fun exercise in electromagnetics.  I'll be interested to learn what you end up designing.  In the mean time I'm focusing back on the projects I'm in the middle of.

Yeah, what the Russians are up to is interesting especially with their old soviet stuff they can get, stuff from China. The message boards are wild.

Anyway thanks for all of your input and good luck on your project. I'll post when I come up with some kind of schematic and or simulations.

[davekni]

Responding to FPS' question

Why couldn't you slide the disk over a solid "barrel" made of ferrite or perhaps a metal rod cluster and use the accellerated disk as a form of sabot to drive a dart that is also slipped over the "barrel" in front of disk.

Certainly possible in some form.  However, requires mechanical strength (ie. cone shape) to take force from disk to smaller diameter dart.  Otherwise disk folds over from force.  Weight of entire assembly (disk, cone, and dart) must be accelerated, reducing efficiency.

[Michelle_]

Got a hantek 1200 series dso for only $400. Going to build a light trap and tinker with it before I start energizing caps.

I’ve decided to build a reluctance coil gun as well because I’ve accidentally learned how they work through researching the induction guns, and I feel like I could build a more reliable one to mess around with and try making portable.

Something I’ve found out about the induction guns which is obvious if you think about it, the radial reaction force puts a huge strain on the coils so they require mechanical reinforcement as to not fatigue and short. Sandia labs actually has a paper on this, it sounds possible to make them reliable but it will be extra work for sure.

[davekni]

Something I’ve found out about the induction guns which is obvious if you think about it, the radial reaction force puts a huge strain on the coils so they require mechanical reinforcement as to not fatigue and short. Sandia labs actually has a paper on this, it sounds possible to make them reliable but it will be extra work for sure.

Yes, that is why I mentioned coil explosion hazard.  For coin shrinking, every version I've seen just allows coil to explode.  A new coil is wound for each coin.
There is a company called Maxwell Magneform that makes reinforced coils for magnetic forming and welding.  Don't know how such is accomplished.

I’ve decided to build a reluctance coil gun as well because I’ve accidentally learned how they work through researching the induction guns, and I feel like I could build a more reliable one to mess around with and try making portable.

Certainly easier for experimenting.  Reluctance can be pushed beyond 2T projectile saturation, but efficiency drops faster as field increases.  For eddy current launchers, doubling field requires 4x input energy and provides 4x output energy.  For a reluctance launcher beyond saturation, 2x field still requires 4x input energy, but provides only 2x output energy.

[Michelle_]

Something I’ve found out about the induction guns which is obvious if you think about it, the radial reaction force puts a huge strain on the coils so they require mechanical reinforcement as to not fatigue and short. Sandia labs actually has a paper on this, it sounds possible to make them reliable but it will be extra work for sure.

Yes, that is why I mentioned coil explosion hazard.  For coin shrinking, every version I've seen just allows coil to explode.  A new coil is wound for each coin.
There is a company called Maxwell Magneform that makes reinforced coils for magnetic forming and welding.  Don't know how such is accomplished.

I’ve decided to build a reluctance coil gun as well because I’ve accidentally learned how they work through researching the induction guns, and I feel like I could build a more reliable one to mess around with and try making portable.

Certainly easier for experimenting.  Reluctance can be pushed beyond 2T projectile saturation, but efficiency drops faster as field increases.  For eddy current launchers, doubling field requires 4x input energy and provides 4x output energy.  For a reluctance launcher beyond saturation, 2x field still requires 4x input energy, but provides only 2x output energy.

I have one paper from sandia labs that suggests that they found the coils they built to work after some 200 firings but it took engineering to do so. They seemed optimistic.

I have a different Chinese paper that complains about coil durability and seems less optimistic and implies they think new materials could be needed. I would need to double check their engineering and intended load.

It’s possible imo that there might be a size and force limit related to possible materials and fabrication techniques of the coil. Luckily for me I’m not launching anything into orbit.

What are you saying about increasing the field beyond 2T? Do you mean specifically certain materials such as cobalt, or is there a way to utilize >1T in iron/steel that I don’t know about yet?

[davekni]

What are you saying about increasing the field beyond 2T? Do you mean specifically certain materials such as cobalt, or is there a way to utilize >1T in iron/steel that I don’t know about yet?

Typical low-carbon steel such as 1018 very-hard saturation is slightly above 2T, why I used that as an example.  Saturation describes the limit of polarization a material can achieve.  It is not a limit on magnetic field.  Beyond saturation, relative permeability approaches 1, same as air.  For example, an applied H field from a coil of 4e6 A/m will generate 5T B field in air.  In 1018 steel, that same 4e6 A/m H field will generate 7T B field, 2T higher.

[Michelle_]

What are you saying about increasing the field beyond 2T? Do you mean specifically certain materials such as cobalt, or is there a way to utilize >1T in iron/steel that I don’t know about yet?

Typical low-carbon steel such as 1018 very-hard saturation is slightly above 2T, why I used that as an example.  Saturation describes the limit of polarization a material can achieve.  It is not a limit on magnetic field.  Beyond saturation, relative permeability approaches 1, same as air.  For example, an applied H field from a coil of 4e6 A/m will generate 5T B field in air.  In 1018 steel, that same 4e6 A/m H field will generate 7T B field, 2T higher.

Interesting, I should look more into this.

I bought some parts to eventually prototype a single coil reluctance and induction launcher.

I found the perfect carbon fiber tubes, they are super thin wall (.5mm) and the inner diameter (4.97mm) is just over 3/16" (4.7625mm). It comes in lengths up to 3 feet so I could get as crazy as I want but for now I bought three 12" pieces.

This means all 3/16 stock is fair game. I got some nice thin 3/16 aluminum tubing, solid aluminum dowels, solid steel dowels as well as slightly oversized 3/16" dowels which fit NICE in the tube. They slide through like butter but are tight enough you can blow them out like a blowgun. This is about as good as it will get without custom machining. I think with some graphite in the barrel the friction will be quite low.

[Michelle_]

Still working on this.

Now that I've learned more I'm realizing Dave's wisdom (thanks!)

I think I'm going to stick to inductive pusher designs for a while instead of worrying about a linear motor. I agree now that for something the size of what I'm doing it makes a lot more sense. I think the scientists are adamant about the linear motor design because they plan on launching something to the moon instead of making a handheld device.

One thing I'm working on (which the pusher design will also make easier) is trying to compare an inductive and reluctance coilgun efficiency with the same amount of J going into the system.

Another thing I've come to realize is that thyristors are dumb (kinda) and for the scale of my designs IGBTs make way more sense and are a way better choice.

That said I did manage to buy some thyristors to play with, a set of three which are a nice size (2" diameter 1/2" thick) and rated at 700v and a gajillion A. I think these might work pretty OK for a reluctance coilgun or at least something to experiment with.

I also found out you have to clamp them which is annoying since I'll have to build a clamp that is also an electrical connection. I will also have to do math to calculate the clamping force! The horror...

I did manage to clamp one in a vice and figure out how it works (which took several hours of tinkering) and I have it switching on, and apparently charging a power bank at 15v is enough for it to latch.

I also got some IGBTs and figured out how those work which was quite a bit easier and they can turn on and off a single LED no problem.

One thing I'm wondering about that I might need to do an experiment with is: does a thyristor block negative voltage? The reason I'm wondering about this is because to generate an eddy current for the inductive coilgun I think the LC ringdown needs to be complete with the +/- AC waveform. For the reluctance coilgun this is not desirable.

On a related note electrolytic capacitors presumably won't work for the inductive coilgun since they won't like the negative voltage, but I think the IGBT will transmit that whereas the thyristor may not?

For now I ordered a few film capacitors that are made for pulse power however they are only rated at 45uf and 800VDC (or something like that) but since the joules are related to v squared I think the 4 of them I bought might have enough power for testing the first prototype against a reluctance coilgun which I bought some 480v electolytic capacitors for.

I've found bigger film capacitors but they are big and scary and I need to practice safely discharging smaller ones and generating voltages to begin with before I mess with those.

To that end, I got a fluke HV probe and a couple transformers to mess with now that I might have means to switch the power.

One thing I'm unsure of is if I need to disconnect the capacitors from the charging power source while they are discharging.

I still haven't simulated anything but now I have a better idea what I want to accomplish so I'll probably try that soon also now that I have an LCR meter.

[davekni]

Now that I've learned more I'm realizing Dave's wisdom (thanks!)

Thank you for the compliment!  But please don't hesitate to question or challenge my "wisdom".  I do make mistakes.

I think I'm going to stick to inductive pusher designs for a while instead of worrying about a linear motor.

Definitely simpler to start.  However, remember that pusher projectiles are more efficient when short, opposite of what you want for ballistics.

I will also have to do math to calculate the clamping force!

I haven't used clamped thyristors before.  However, AFAIK clamping force is primarily to improve heat transfer.  For your low duty cycle use, just enough force to make a reliable electrical connection is likely sufficient.

One thing I'm wondering about that I might need to do an experiment with is: does a thyristor block negative voltage?

Some analog circuit simulation (LTSpice or other) would help in thinking about current and voltage, especially through inductors.  Ideal inductor has current lagging voltage by 90 degrees.  Most thyristors (other than TRIACs) block reverse current.  In other words, they allow reverse voltage without conducting current.  However, be careful with high-speed thyristors intended for pulse discharge.  They often handle only small reverse voltage before being damaged.  IGBTs are similar in that respect.  They block only small reverse voltages.  Many IGBTs packages include an internal diode.  Internal diode prevents significant reverse voltage by conducting reverse current.  If you are using a pulse-discharge thyristor or an IGBT without internal diode, an external diode is needed to prevent device damage.

On a related note electrolytic capacitors presumably won't work for the inductive coilgun since they won't like the negative voltage

At least won't work efficiently.  A diode can be used across an electrolytic capacitor to prevent significant reverse voltage.  But then inductor current decays slowly.

but I think the IGBT will transmit that whereas the thyristor may not?

Here's where circuit simulation will help.  The same positive current direction that discharges cap will "discharge" it through zero to negative voltage.

a reluctance coilgun which I bought some 480v electolytic capacitors for.

My reluctance coilguns are all low power for kids to use.  (One down to 20V so hand wire contact is sufficient to fire.)  They all use electrolytic caps.  My coils have enough resistance to be critically damped, avoiding capacitor reverse voltage.  This is not efficient, however.  Nor is the above solution of diode across capacitor.  Ideally coil current would end when projectile is at center of coil.  Anything that slows current decay causes reverse force as projectile proceeds past coil center.

I've found bigger film capacitors but they are big and scary and I need to practice safely discharging smaller ones and generating voltages to begin with before I mess with those.

Good to be cautious.  Your 45uF 800V caps are in the energy range of defibrillators.  Even that can be lethal.  As energy increases (ie. coin shrinkers etc.), one mistake is the last one you will ever make.

Have fun!  And of course do so safely!

[Michelle_]

Now that I've learned more I'm realizing Dave's wisdom (thanks!)

Thank you for the compliment!  But please don't hesitate to question or challenge my "wisdom".  I do make mistakes.

I think I'm going to stick to inductive pusher designs for a while instead of worrying about a linear motor.

Definitely simpler to start.  However, remember that pusher projectiles are more efficient when short, opposite of what you want for ballistics.

I will also have to do math to calculate the clamping force!

I haven't used clamped thyristors before.  However, AFAIK clamping force is primarily to improve heat transfer.  For your low duty cycle use, just enough force to make a reliable electrical connection is likely sufficient.

One thing I'm wondering about that I might need to do an experiment with is: does a thyristor block negative voltage?

Some analog circuit simulation (LTSpice or other) would help in thinking about current and voltage, especially through inductors.  Ideal inductor has current lagging voltage by 90 degrees.  Most thyristors (other than TRIACs) block reverse current.  In other words, they allow reverse voltage without conducting current.  However, be careful with high-speed thyristors intended for pulse discharge.  They often handle only small reverse voltage before being damaged.  IGBTs are similar in that respect.  They block only small reverse voltages.  Many IGBTs packages include an internal diode.  Internal diode prevents significant reverse voltage by conducting reverse current.  If you are using a pulse-discharge thyristor or an IGBT without internal diode, an external diode is needed to prevent device damage.

On a related note electrolytic capacitors presumably won't work for the inductive coilgun since they won't like the negative voltage

At least won't work efficiently.  A diode can be used across an electrolytic capacitor to prevent significant reverse voltage.  But then inductor current decays slowly.

but I think the IGBT will transmit that whereas the thyristor may not?

Here's where circuit simulation will help.  The same positive current direction that discharges cap will "discharge" it through zero to negative voltage.

a reluctance coilgun which I bought some 480v electolytic capacitors for.

My reluctance coilguns are all low power for kids to use.  (One down to 20V so hand wire contact is sufficient to fire.)  They all use electrolytic caps.  My coils have enough resistance to be critically damped, avoiding capacitor reverse voltage.  This is not efficient, however.  Nor is the above solution of diode across capacitor.  Ideally coil current would end when projectile is at center of coil.  Anything that slows current decay causes reverse force as projectile proceeds past coil center.

I've found bigger film capacitors but they are big and scary and I need to practice safely discharging smaller ones and generating voltages to begin with before I mess with those.

Good to be cautious.  Your 45uF 800V caps are in the energy range of defibrillators.  Even that can be lethal.  As energy increases (ie. coin shrinkers etc.), one mistake is the last one you will ever make.

Have fun!  And of course do so safely!

Thanks for your thoughts as usual. I don't think I have an intuitive understanding yet of the relationship between current and voltage for these circuits. I'm watching LTspice tutorials now and hopefully that will be illustrative. I am aware of the general idea of back EMF and have questions in my mind about ringdown versus shutting off the current with the IGBT, but I think I need to simulate and see.

For now I think I'm going to shelf the thyristors and focus on understanding IGBTs unless there's a particular benefit to the thyristors that I'm missing. At this point I'm inclined to use a bunch of IGBTs for both the reluctance and inductive coilguns. I actually bought some of the (IKY75N120CS6 1200V 300A peak) IGBTs you linked a while back as they do seem very promising for my small applications. The peak current is lower than the thyristors by a lot so I'll have less margin for error, but on the other hand they are cheap and small and I can use as many as are required. I actually don't really have a concept of what peak currents I might even need yet.

A new idea I just thought of is using supercapacitors to generate a longer AC pulse and sending it through a ferrite core which a hollow projectile slides on the outside of. I'm not sure if anyone has tried this or if it's even a good idea or not though! I thought of this from seeing a ring launcher that was apparently switched to an AC main. Since supercapacitors are a newer invention I'm wondering if this concept is unexplored, but again I could be missing a key concept.