High power EM projectile launcher / Projects / Forums

Forums

4hv.org :: Forums :: Projects

« Previous topic | Next topic »

High power EM projectile launcher

first previous 4 5 6 7 next

Move Thread

LAN_403

hen918

Sat Jan 23 2016, 11:30AM

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

You can improve coupling my milling or sanding off as much of the top insulation on the coil as possible. Obviously, if you take off too much you'll weaken the structure and it might break up.

Kizmo

Sat Jan 23 2016, 11:52AM

Registered Member #599 Joined: Thu Mar 22 2007, 07:40PM
Location: Northern Finland, Rovaniemi
Posts: 624

hen918 wrote ...

You can improve coupling my milling or sanding off as much of the top insulation on the coil as possible. Obviously, if you take off too much you'll weaken the structure and it might break up.

It might be possible but even now there is not a whole lot on top of the copper, about 3mm. Winding itself is 35mm tall.

Uspring

Sat Jan 23 2016, 05:43PM

Registered Member #3988 Joined: Thu Jul 07 2011, 03:25PM
Location:
Posts: 711

Nice measurement Kizmo

There is an expression for the force on the disk using your experimental values. The force is:

F = dL/dx * E/L,

dL/dx is the derivative of L wrt position of the disk, E the energy in the coil and L its inductance. If, e.g. we assume, that all 8700J get into the coil right at the beginning at position 1, we get:

F = 1uH/4.2mm * 8700J / 5uH = 400000N or about 40 tons.

At position 4, same energy assumed:

F = 0.262uH/4.2mm * 8700J / 6.5uH = 80000N or about 8 tons.

These values are ridiculously large. There are several reasons for that. Firstly, the cap energy takes some time to go into the coil. My guess is, that you have an oscillation frequency of about 300 Hz, This time will be at a quarter period, i.e. 0.8 ms. An e.g. 200g disk with such a force on it would have left the first 4mm within 40us. The assumption of all energy in the coil in the beginning is unrealistic. Secondly there is some resistance in the circuit, which reduces the current.

Mostly I think, it's still another effect. The disk is not a perfect conductor. Its DC resistance limits the current, which will reduce repulsion. The relevant quantity is here

2*pi*f*Ldisk/Rdisk

i.e. the ratio of its AC resistance to its DC resistance. If this ratio is much larger than 1, the disk DC resistance is negligible, i.e. it is a perfect conductor for our purposes. If it gets smaller than one, the current in it will drop and the repulsion also. Your k measurement was made at much larger frequencies than your real shots, so the disk will have in practice a much smaller resistance ratio, making it much less effective in terms of repulsion. I believe, these measurements should be made at the real frequencies. Dunno how to do that, since that would require your probably electrolytic cap bank be run with AC.

It's probably possible to work out the dynamics of the acceleration, if the time dependence of the current in the coil is known. The most important parameter is its DC resistance together with the cap ESR.

DerAlbi

Sat Jan 23 2016, 06:44PM

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

Where do you get those forumlas

Uspring, with the help of this formula one could easily just make an LTSpice circuit simulating the whole shot include efficiency and everything.
What i do currently not know is this: what inductance to use? Kizmo calculated the Inductance based on the resonant frequency. Based on 1uF +/-some%
So the L is wrong basically. The only relieable data we have is the coupling factor (its independend on C). Still i need a usefull L. Kizmo said before its 8uH, now its just 7.5uH..
Let me do some work. the excat value does not matter right now. Whats more annoying is the Curve-Fitting of k

DerAlbi

Sat Jan 23 2016, 07:45PM

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

Wow. Its does indeed predict everything perfectly.
Sry i have only a LT-Spice simulation, pictures are useless. But we can play with the parameters within the simulation. It directly spits out Kizmos 90m/s end velocity (assuming 5mOhm overall ESR. The K-Factor function does not fit perfectly. There is a small missmatch right at the beginning where the missmatch actually hurts the most -.-).

I have commented to easier use.
]ringlauncher.zip[/file] [Edit: increased numerical stability]

I allways wanted to do such simulation for a coilgun. But in a multistage i didnt figure out how the variing coupling factors of the coils can be respected properly.
Compared to any FEM or whatever this is instantaneous

Kizmo

Sat Jan 23 2016, 09:26PM

Registered Member #599 Joined: Thu Mar 22 2007, 07:40PM
Location: Northern Finland, Rovaniemi
Posts: 624

Thank you very much about all this interesting stuff! This is why i love this forum so deeply :)

Anyways, I made a first "next step" test fire today.

Parameters:
- 192g disk
- 31.6mF 1400V capacitor
- Capacitor charged to 920Volts 13.373kJ

Disk velocity measured 120m/s

As you can see even the large steel piece i had on top of the improvised projectile stopper actually jumped up a bit. This is serious stuff right here :D

And im nowhere near done with this.

DerAlbi

Sat Jan 23 2016, 10:07PM

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

Confirmed

hen918

Sat Jan 23 2016, 10:43PM

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

Because the coupling decreases exponentially the further away the disk is from the coil, I've always thought that a shorter, higher current pulse (more voltage, less capacitance) would transfer more energy to the coil. Could you try at a higher voltage/current or wouldn't your switching device withstand it?

Thanks,
Henry

Kizmo

Sat Jan 23 2016, 10:45PM

Registered Member #599 Joined: Thu Mar 22 2007, 07:40PM
Location: Northern Finland, Rovaniemi
Posts: 624

I am moving towards using very large triggered spark gap switch so almost any voltage from 1kV and up will be doable. Any amount of current is doable.

Uspring

Sun Jan 24 2016, 05:37PM

Registered Member #3988 Joined: Thu Jul 07 2011, 03:25PM
Location:
Posts: 711

I had some fun playing around with your simulation. Cool and thank you, DerAlbi

I have no reference for the equations, so I derived them. If you're interested, I can PM or post a bit about this.

The max possible efficiency you can achieve is kÂ². That would be for zero resistance in the coil and full current in it before the disk would begin to move. Put in other words: Say you have an energy E in the coil and you move the disk from a position where kÂ²=0.25 to a position, where kÂ²=0.20. You will then gain a kinetic energy of 0.05*E (the difference of the kÂ²). For Kizmos setup kÂ² is about 0.4, but the actual efficiency is at about 8%. One reason for lower efficiency is, that the disk has moved to considerably lower kÂ² values before all the energy arrives in the coil.

The other reason is the ESR (and ESL).

first previous 4 5 6 7 next

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