Coilgun projectiles question: Magnetic Saturation? / Electromagnetic Projectile Accelerators / Forums

Forums

4hv.org :: Forums :: Electromagnetic Projectile Accelerators

« Previous topic | Next topic »

Coilgun projectiles question: Magnetic Saturation?

Move Thread

LAN_403

ShieldExperiments

Thu Dec 07 2017, 01:29AM

Registered Member #61689 Joined: Mon Jul 10 2017, 02:46PM
Location:
Posts: 14

I've noticed that most of the coilguns that I see use mild steel projectiles of around 4~8mm diameter. Given that coilguns are most efficient at higher velocities, (thus the great benefit of multistage designs), would it not make sense to use smaller projectiles of the same proportions, such as 1mm diameter? With lower mass, much greater muzzle velocities might become attainable (due to efficiency improvements from using less massive projectiles), resulting in equal or greater kinetic energy. This lead me to wonder if smaller projectiles - practically inductor cores - might become magnetically saturated and less efficient in such strong magnetic fields?

I ask because I am in the process of building a 50J input range 'hybrid' coilgun, using a Li-Ion pack for a small primary coil, an 'injector' if you will, before a ~40 joule BJT switched capacitive stage (40V, .05Farad ish), and intend to use a 0.4 gram, 0.1mm diameter projectile.

Also, does anyone have any suggestions on firing timings? I'm doing this as an arduino enabled junk box build, with efficiency as a primary goal. I plan on using a time defined pulse for the 'injector' stage, then a delay, then a timed pulse from the capacitive stage. I know that photogates are better than predefined timings, so please don't track me down and blast me with projectiles from your superior designs for this heresy

ShieldExperiments

the_anomaly

Thu Dec 07 2017, 02:59AM

Registered Member #19 Joined: Thu Feb 02 2006, 03:19PM
Location: Jacksonville, FL
Posts: 168

I think if you were to measure the speed of the projectile from its initial coil without firing any other coils, you could get a ballpark time to fire the second coil and so on. I have not done this but I would work incrementally. Is it just two stages?

I would also experiment with the pulse duration; I've seen a lot of other threads where the builder kept the pulse to a certain time to avoid having the projectile pass the center of the coil while the magnetic field is still on. I would also look at how the system is damped, which I think might affect pulse duration.

DerAlbi

Thu Dec 07 2017, 10:04PM

Registered Member #2906 Joined: Sun Jun 06 2010, 02:20AM
Location: Dresden, Germany
Posts: 727

would it not make sense to use smaller projectiles of the same proportions, such as 1mm diameter?

Unfortunately physics will be in your way.
First, current in a coilgun rises at least squared to velocity. This implies a sweet spot for thicker projectiles which will be slow enough to require manageable currents while still having serious projectile energy at lower speeds. (I found myself maxing out at 120m/s (theory, not tested))
Also the pulling force of a coil is related to the ratio of "inductance with the projectile in it" over "inductance without projectile in it" (for me this is a factor of >2.5). Your coil must be smaller in order to significantly change inductance in presence of a projectile. A smaller coil can however only hold smaller amounts of energy therefore it can convert less into kinetic energy. So you actually cant transfer as much energy as if you had a bigger projectile -> your speed gain will be disappointing (equal to a fatter projectile? not sure).
Another issue is that small projectiles will behave strange since it is possible that the effective Âµ_r of the material gets lower since Âµ_r is not only a material property but also a geometry property: solid iron is much more magnetic than the small grains of iron powder made from the same iron. Shooting needles is therefore not as easy as it might seems. They are magnetically less active. At least they can be in theory. Maybe its not noticeably at 1mm, but maybe at 0.1mm. Can have no experience in this, so its on you to test it. (Talking about needles, with iron powder i have experience)

a ~40 joule BJT switched capacitive stage

Not sure if you mean what you wrote but please use mosfets at this voltage.

Please understand that BJT is different to a mosfet. Anyway, mosfets are the way to go.

I know that photogates are better than predefined timings

It depends. Every shot is different due to different preconditions like pre-magnetization of the projectile. A predefined sequence will require you to compromise on efficiency but you will get very consistent results. The idea is that the traveling magnetic field will be slightly slower than it could be - so if the projectile gets too fast (due to favorable initial conditions) it will experience suck-back and therefore stay in place and exactly follow your timing.
If you build a projectile tracking system this works too as long as it will create sufficient suck-back which is the only way to get consistent shots.
When i had my old prototype tuned perfectly its reliability was horrible.
Of course you can implement some serious algorithm to implement a control loop in you microcontroller, i would certainly be interested!

ShieldExperiments

Fri Dec 08 2017, 12:36AM

Registered Member #61689 Joined: Mon Jul 10 2017, 02:46PM
Location:
Posts: 14

@ the anomaly
I'll try to run some simulations to figure out where it would be on damping, and also what my expected inductance is with vs without projectile in the coil. Getting the optimum pulse duration figured out should go a long ways towards getting an efficient system. I'll look through the parts bins, and might even implement a half bridge for the battery driven 'injector' coil. I'll have to do some reading and find out what others have done to tune timing [i]

@DerAlbi
Thanks for the explanation of the physics of current vs velocity, not yet having a solid understanding of calculus, this makes the equations relating current, EMF, and velocity which I have seen make much more sense. As you suggest, once I get this working I may try making several experimental guide tubes and coils of different diameter but same current*turns product to test the significance of the potential lower magnetic activity.
As for my switching elements, I'll have to rummage around a bit more; I had wanted to use MOSFETs for their [very slight] resistive on state behavior over the foreward voltage of a BJT, but I'm not sure that the MOSFETs that I have are suitable, [NTP18N06, 60vDS, 45A peak 10 microseconds].
Any advice on how to implement a half bridge once I get it working with low side switching only?

DerAlbi

Fri Dec 08 2017, 08:05AM

Registered Member #2906 Joined: Sun Jun 06 2010, 02:20AM
Location: Dresden, Germany
Posts: 727

As you suggest, once I get this working I may try making several experimental guide tubes and coils of different diameter but same current*turns

I can spare you the trouble.. given your 40J source efficiency is actually important for you so you will find that your optimium coil shape will be: "Outer diameter" = 3x "Inner diameter". Its called a Brooks Coil.
The problem will be, since you go for smaller diameters, that your tolerances must really be tight. Its a big difference is you have a 1mm air gap to the projectile with 10mm projectile diameter versus 1mm projectile diameter. I am not sure how you will manage the mechanical side of things. The projectile must fill all the area it can inside the coil.
Also your coil shape is basically constant. You change inductance by choosing the right wire diamter [therefore the amount of turns that fit in your winding volume]. The faster the projectile the lower the inductance must become in order to have current built up fast enough. I am not sure how to handle that in small coils.

but I'm not sure that the MOSFETs that I have are suitable

Get rid of the idea that you build a viable half bridge with no investment. You need good mosfets that are suited / selected for the job and you need good gate-driving and a solid mechanical solution for your circuit [PCB]. Everything else turns your electromagnetic launcher into an explosion driven device. There is no quick&dirty solution for a half bridge - for that there is the SCR design

You have only low voltage to work with. You want to have high speeds, so you have to transfer "a lot" of energy in short time which leads to high pulse power. This pulse power must come from somewhere and basic formulas like P = V * I will apply. Since your voltage issnt particulary high you will need *reeally* high currents for every stage you consider your projectile to be "fast".

Just imagine if you go for 10% efficiency, which leads to 4J kinetic energy. A needle might weigh about 0.2g. [i]. Anyway, 4J in such a small mass yields 200m/s using E=m/2*v^2.
Now lets say you have 10 coils to deliver that energy you need to push 0.4J per coil. Your coil will be maybe 2cm long? So at 200m/s your active coil time will be 100Âµs [using 200m/s = 2cm / time]. Delivering 0.4J in 100us yields a peak power of 4kW since Power = Energy / Time. => which is 4kA, at 10V (not 40V because the capacitor will be discharged at the end! ) Sounds doable for now. However this is only the real-part of the power. A coilgun is a coil driven device which needs much more energy in the coil than the coil it self can deliver as kinetic energy. [for 2.5J energy transfer i have >20J in the coil, so 10x more energy]. For 10x more energy you need around 3x [sqrt(10)] more current. We are talking >10kA now. 10kA at 10V also mean your overall ESR will need to be lower than 1mOhm. And there is the point where i would say "not a chance!"

These are just ball park figures. The mathmatical path i took to give you some insight will give you a hint where to cut expectations. I hope this helps.
This is not intented to discourage you, dont get me wrong

Plug in your own numbers and do your thing, but having realistic expectations is the only way to avoid premature disappointment which would lead to a premature project end in which case you can spare yourself the cost by wildly swinging you pencil instead of your soldering iron

the_anomaly

Fri Dec 08 2017, 02:26PM

Registered Member #19 Joined: Thu Feb 02 2006, 03:19PM
Location: Jacksonville, FL
Posts: 168

Any advice on how to implement a half bridge once I get it working with low side switching only?

The high side will just need a high side driver. I suggest using a half bridge driver IC. They are not too expensive and allow logic inputs so you can control it directly from your arduino. If you have the ability to work with SMT components, that opens a lot of doors. If not you can usually find something in a dip package. Just a quick search on Digikey and I found this bridge driver:

Of course you would have to review the specs to see if they fit your design. Probably the most important thing is to determine if its output current is adequate to switch your chosen FETs and if its rise/fall time are fast enough. For that I would look at the total gate capacitance of the FET and then calculate the gate V rise/fall time. So choose FETs first then choose a gate driver to match. You can build your own gate driver too but I think it adds a layer of complexity.

DerAlbi

Fri Dec 08 2017, 03:01PM

Registered Member #2906 Joined: Sun Jun 06 2010, 02:20AM
Location: Dresden, Germany
Posts: 727

I suggest using a half bridge driver IC

Not a good suggestion

Half bridge driver ICs supply the high-side via a bootstrap capacitor which needs an active low-side to switch first in order to function. This is not a problem in a true halfbridge with periodic switching, but it has its limits. You cant switch the high side first for example.
In a coilgun with a halfbridge, the topology is different and its actaully called an "asymmetric halfbridge" which is a full H-Bridge but it has diodes as one diagonal pair and Mosfets as the other diagonal pair.
One could implement a full bridge by building 2 half bridges but before a shot care must be taken in order to precharge the bootstrap capacitor supply. Its doable - and given the low voltage approach maybe not even a bad idea given the voltage drop across diodes at several kA might be higher than across some paralleled mosfets. hmmh.
And thanks for pointing out this trap by falling into it, it gave me a bunch of ideas how to save cost and reduce complexity ^_^

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob