Capacitors for coilguns considered harmful? / Electromagnetic Projectile Accelerators / Forums

Forums

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

« Previous topic | Next topic »

Capacitors for coilguns considered harmful?

Move Thread

LAN_403

BigBad

Wed Apr 15 2015, 10:54AM

Registered Member #2529 Joined: Thu Dec 10 2009, 02:43AM
Location:
Posts: 600

One of the limits on performance is the power you can get from the capacitors, and big capacitors are relatively expensive.

I was recently thinking about using mains power as a way to power coilguns.

Obviously, the current would be limited, I thought, but not necessarily trivially small.

Consider a 13 amp socket. It can supply ~3.1 kW of power at ~240 volts with a peak voltage of ~340 V or so.

Of course, if you take more than 13 amps, the fuse will blow.

But wait! How long does the fuse take to blow, 13 amps is the average . If you take more for a short time, and less for a short time, it won't blow.

I looked it up. It turns out that you can go about 5 times higher current, briefly, without blowing the fuse, and much longer still if a slower fuse is used. And this is not dangerous, it turns out that mains wiring can easily take that current, briefly; the resistance of the mains wires is very low, it just can't take it for multiple seconds. But high power coilguns fire for only about a hundredth of a second!

So a 13 amp socket can give you 80 amps (which if you think about it is 25 times the power in the resistance of the coils/magnetic force - ~80KW at the coilgun). And a 35 Amp cooker circuit... 175 amps (500kW at the coilgun!!!). Both these numbers are pretty huge.

So it looks like for static firings, that you would never want to use capacitors. I mean, you could certainly have capacitors in the circuit for buffering, but they needn't be really big ones.

There is a caveat or two obviously.

1) Don't remove the fuse, use a slow one if you must, but fuses are really, really, really important
2) if you're on single phase supply, you will need to time your coilgun to the 50 hertz waveform so it fires when there's voltage, otherwise you'll get no current(!) Detecting the mains phase is not particularly difficult. Firing the gun at zero crossing, and running it initially on capacitor for the early stages and then using the mains later on is probably a good idea, coilguns use more power/energy at high speeds.

I would say be careful out there, but you're all using high voltage caps anyway, so I know you must be.

aarpcard

Wed Apr 15 2015, 02:46PM

Registered Member #2848 Joined: Tue May 04 2010, 05:19AM
Location:
Posts: 44

If I'm understanding what you're proposing correctly, I don't think there'll be enough current.

Coil guns operate by having a huge current pass through the firing coils thus inducing a large magnetic flux. All the coil guns I've built have had a current pulse peak of around 1-2kA per coil resulting in about 1 to 2+MW per coil. You really need the high current to get a strong B field. (Also maybe I'm confused, but a 175amp pulse at 340V mains voltage would be about 60kW . . . not 500kW)

175A at 340V will work, but it won't yield the results you can get from decent capacitors.

The only way I could see this being useful would be if you used mains power to fire a coil gun with many many many firing stages - that in itself would be pretty neat though.

hen918

Wed Apr 15 2015, 05:26PM

Registered Member #11591 Joined: Wed Mar 20 2013, 08:20PM
Location: UK
Posts: 556

aarpcard: Based on BigBad's other thread, I'd say he is going for a longer coil length and slower acceleration with less current required.

Orrrrr, you could just throw some large capacitors in there and a large diode or four (plus a current limiting resistor) and you would have your standard coilgun!

klugesmith

Thu Apr 16 2015, 02:02AM

Registered Member #2099 Joined: Wed Apr 29 2009, 12:22AM
Location: Los Altos, California
Posts: 1716

Coil guns operate by having a huge current pass through the firing coils thus inducing a large magnetic flux. All the coil guns I've built have had a current pulse peak of around 1-2kA per coil resulting in about 1 to 2+MW per coil. You really need the high current to get a strong B field.

It's not the amperes, it's the ampere-turns. Anything you could do with 1000 amps, you could do with 100 amps in 10 times as many turns (using thinner wire to fit the same space). The re-wound coil would need 10x the voltage. But then 1/100 of the capacitance would store the same energy and give the same pulse timing as the original.

If you can choose the wire size, you can trade voltage for current to match your power source. Peak power remains the same for a given peak ampere-turn requirement, unless you a) use more copper, b) use metal with lower resistivity, or c) change the coil shape to reduce the average turn length.

That said, people routinely get megawatt power in pulses from electrolytic capacitor banks. It happens in mains wiring under short-circuit conditions, but that's hard on the protection devices.

By the way, in RLC circuits, the product of peak current and initial voltage is much higher than the peak power. But the change in stored energy, divided by discharge time, accurately gives the average power.

aarpcard

Tue Apr 21 2015, 02:41PM

Registered Member #2848 Joined: Tue May 04 2010, 05:19AM
Location:
Posts: 44

. . . but with 10x the turns, with smaller diameter wire, your resistance will skyrocket lengthening your pulse width. As such you'd need longer coils and longer projectiles which usually aren't very efficient.

klugesmith

Wed Apr 22 2015, 06:29AM

Registered Member #2099 Joined: Wed Apr 29 2009, 12:22AM
Location: Los Altos, California
Posts: 1716

. . . but with 10x the turns, with smaller diameter wire, your resistance will skyrocket lengthening your pulse width.

I think you missed my point. The higher R and L are compensated by using smaller C, as I said above, when rewinding coil to match a different power source. Coil shape and size, stored energy, timing, and power are constant.

Here is a numerical example figured with Barry's Inductor Sim

I chose dimensions to minimize roundoff discrepancies in the integer turns per layer and layer count.
Coil ID = 25 mm, OD = 65 mm, Length = 30 mm.
Wound with 12 AWG & 22 AWG wire (10 AWG steps changes cross-sectional area by factor of 10).
n_turns = 126 & 1276 (n/n = 10.13).
L = 0.418 & 42.911 mH (L/L = 102.7)
R = 0.093 & 9.587 ohms (R/R = 103.1)
wire length = 17.81 & 180.39 m (l/l = 10.13)
wire mass = 0.53 & 0.53 kg (m/m = 1.00)

Now as for the discharge dynamics, let's use Barry's RLC discharge sim

R and L were set as close as possible to inductor sim values.
C = 20000 & 200 uF (C/C = 0.01)
V = 100 & 1000 V (V/V = 10)
E = 100 joules in both cases
I_peak = 454 & 45 amps (I/I = 0.1)
Time to I==0 is 9.52 & 9.68 ms (t/t = 1.02)
Peak ampere turns = 57204 & 57420 (I/I = 1.00)

I always wanted to do that exercise. Dale, didn't you ask for something like that?

Product of peak voltage and peak current is about 45.2 kW in both cases. The peak power into the coils is a bit more than half that (because of damping between 1/8 and 1/4 cycle) -- say 28 kW.
From the perspective of original post, consider a 60 amp (RMS) resistive load on 240 volt (RMS) AC circuit. Average power is 14.4 kW, peak power is 28.8 kW. If energized for an isolated half-cycle, I bet that would not trigger overcurrent protection in a typical appliance circuit.

But wait, there's more. Suppose we have the same supply voltage options (100 & 1000 V) and want to increase stored energy to 1000 joules. That means increasing C by a factor of 10, to 0.2 F & 2000 uF.

Used with the coils of 12 & 22 AWG wire, the timing would be underdamped & extend beyond 50 ms.
We can preserve the original pulse timing by rewinding with thicker wire (7 & 17 AWG).
n_turns = 40 & 400.
L = 0.042 & 4.2 mH. Anybody seeing a pattern here?
R = 0.0095 & 0.95 ohms
wire mass is still 0.53 kg.
Back in RLC sim (for 0.2F-100V & 2000uF-1000V)
Peak I = 4476 & 448 A.
Time to I==0 is 9.68 ms for both cases.
Vpk*Ipk = 450 kW. Peak power into the coil is about 280 kW.
Peak ampere-turns is 179000 for both (only 3.16 times greater than the 100 joule cases).
1000 joules will heat any of the 4 coils by about 5 degrees C.

Argh, time to call it a night. Too bad the thread title won't make it easy to find this a year from now.

BigBad

Wed Apr 22 2015, 09:48AM

Registered Member #2529 Joined: Thu Dec 10 2009, 02:43AM
Location:
Posts: 600

Consider that, in the end, you ideally want to have a certain number of amp turns placed ahead of the projectile, always in the same place relative to it, in a 3D cylindrical blob that's running ahead of it.

So the total field energy is constant- the electrons looping around in the volume don't care much at all whether they're in 10 square turns or a ten thousand turns

But what does matter is the voltage of the coil to the drive circuitry.

I think using different wire only confuses matters.

Suppose you keep the same wire, but break the coil in two longitudinally, and connect them in parallel, using the same wire.

What you've just done is, halved the voltage; or equivalently, doubled the current for the same voltage. It's exactly like feeding your coil through a 2:1 step up transformer.

So if you wind short, flat disks, you never have to rewind, just connect them up differently; like if your active coil consists of 6 disks in series, later stages can be two parallel sets of 3 in series, and then 3 parallel sets of 2 in series, and then finally all 6 in parallel.

The point is the energised section should always be the same shape, and the wire should always be carrying the same volumetric current, and the drive circuitry voltage should always be the same.

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