Re: Coil design assist
Reid, Sat Apr 11 2009, 09:36PM

ps. the excel file is a .xlsx, the 2007 format. sorry. here's an '03 version
]coilgun_data_plot.zip[/file]

Re: Coil design assist
Barry, Sun Apr 12 2009, 02:40PM

Well I didn't see any immediate replies, so I'll try my hand at discussing it a little...

But first let me hasten to say that coil optimization for a given capacitor bank and projectile is not simple or well understood. This whole problem is vaguely like an impedance-matching exercise where we try to match the magnetic pulse waveform to the projectile dynamics for maximum efficiency.

I noticed you didn't have measurements for your 20mm and larger projectiles. Did they not move at all, or are those measurements simply not done yet? My experience has been that best efficiency occurs when the projectile length is about equal to coil length (33mm in your case).

Coil resistance ... This one is easy. Resistance dissipates energy through heating effects. We want this kept low enough to at least get an underdamped response from the circuit. All the work I've done tells me that lower is always better, and even a few milliohms makes a significant difference. Keep wires short and fat, connections soldered, and screw terminals tight. I expect that heavier wire than 22-gauge in the coil will work better, but don't have a formula to say how heavy.

Coil inductance ... For coilguns based on SCR control, the waveform is a sinusoid with frequency determined by the LC time constant. Use my LC Time Simulator to figure out how much. Note this is independent of the circuit's resistance, thank the FSM for one simplifying factor.

Voltage ... By adjusting the charging voltage you can get a feeling for how your coilgun is tuned. Next time, try even lower voltages than 200 so you can identify the peak. You might find the efficiency got even better below 200v, and maybe you can brag about better than 2.7% efficiency even though the velocity are lower. A curve like yours (diminishing speeds at higher voltage) usually means a bad case of the DSE (dreaded suck-back effect) which means the projectile is moving faster than the waveform permits. Either slow down the projectile (make it heavier) or speed up the waveform (lower inductance).

Capacitance ... We want to maximize total stored energy so more capacitance (or voltage) is better. Shrug. Simply buy all the joules you can afford, within the limits of the desired LC time constant and the ability to charge and switch high voltage.

So much for the background. What should you do next? If it were me I'd try a coil with heavier wire, probably 16-gauge, and enough turns for an inductance about half of what you have now. Or perhaps even less inductance, given you're interested in short projectiles of 10-20 mm. Use that Inductor Simulator again to help design it.

And then if I saw a graph of decreasing speed as voltage goes up past 200v, I'd start removing turns to decrease inductance further. In fact, you could try that right now with your existing 22-gauge wire coil and I think you'd see higher speeds right away.

Cheers, Barry
Our POP server was kidnapped by a weasel.

Re: Coil design assist
Reid, Sun Apr 12 2009, 05:02PM

thanks for the reply, Barry.
Last night I was playing with my efficiency numbers, as well as some math, here's what I did

1j= kg* (m^2/sec^2)

j= .001* (91^2/1^2)
j=8.281

now, joules in a capacitor bank = .5*F*v^2, efficiency of entire assembly assumed to be 3%
276.03J=.5x*300^2
x=6134uF

now, If I'm going to throw my 1g, 15mm projectile, I'll use a 30mm coil, internal dia. of 6mm, wire of 22 gauge
I want my projectile's center of magnetism (it's center of mass) to move to the COIL's center of magnetism in the time that the capacitor bank discharges.

91m/1sec = .15m/xsec
x = .0164sec

so my capacitor bank needs to discharge in .0164 sec. I use the RLC simulator on coilgun.info (props, Barry), input my voltage, capacitance, and now I'm stuck.
I've played with the simulator, tuning my inductance and my resistance to keep my inductor pulse time at .016 sec, but I need to isolate one variable that I can use to control the coil sim.

What I really want to do now, I need to figure out the optimization point where I can trim inductance and resistance to get my desired discharge time. Is there a math way to do this, or do I have to trial and error it with the sims?

ps. the missing values in the spreadsheets are experimental results not taken yet.

Re: Coil design assist
hotcrazyfruit, Sun Apr 12 2009, 06:38PM

from my experience, you want the coil to 'turn off' just as the tip of the projectile reaches the centre of the magnetic field. to prevent ANY suckback at all. also, if you want to move on to higher energies and higher speeds, you need to GREATLY reduce the number of layers of your coil. the max i ussually go is 4 layers, and 2 layers for 1.3 kj. this lowers inductance, meaning faster discharges.
Nick,

Re: Coil design assist
Barry, Mon Apr 13 2009, 02:40AM

Nice approach, I wish more people used some of these fundamental analytical estimates.
May I suggest a 15mm projectile is best served by a 15mm coil. Otherwise you'll have an extra 15mm of unused coil sitting around with wasted resistance that's too far from the projectile to contribute any force.

Start by assuming R=0 and then plan on over-building wire size and keeping the discharge path short and tight. Every milliohm works against you to waste energy.

For all practical purposes, the discharge time is independent of resistance R. It just only depends on L and C. There is a simple LC formula for this and I also made it into a Java applet for LC Time Simulation to remove chances of an off-by-two error. The main idea here is that a coilgun's discharge time is only interesting during the first half-cycle of the sine wave.

Also note what hotcrazyfruit/Nick said about using 3 or maximum 4 layers. There's something about coilguns such that they favor coils with few layers. The outer layers add a great deal more inductance and resistance than inner layers, and are farther from the iron projectile and therefore less effective. Also to get low resistance, coilguns work better with heavier rather than thinner wire.

Cheers, Barry
Barry denies everything. Actually no I don't.

Re: Coil design assist
Reid, Mon Apr 13 2009, 10:30PM

Okay, so before I start crunching the numbers, what would a better timing estimate be for my coil?

15mm ling, 6mm internal dia, 15mm projectile.

Ps. Thanks for the math compliment.
My goal with this project is to ENGINEER, not just build, a coilgun similar to the one shown in this video.
I am doing this a a research project, with a view to learning the math, and analysis process, so I don't want to go galavanting off, building myself a sniper rifle. Besides, I've done some efficiency numbers, and I recognize that'd take a tabletop covered in capacitors.

Thanks for your continued help,
Reid

Re: Coil design assist
KLH, Wed Apr 15 2009, 08:14AM


Adding more layers greatly increases the interwinding and interturn capacitance. This would (probably) cause adverse effects, but I don't know if these are important in a coilgun.

Re: Coil design assist
MachineShop, Mon May 04 2009, 05:19AM

i understand the aim is to decrease resistance as much as possable but i was wondering can this be done by using small gauge on the outside and larger gauge on the inside?like twards the middle of the coil?