Ring launcher optimum oscillator frequency / Electromagnetic Projectile Accelerators / Forums

Forums

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

« Previous topic | Next topic »

Ring launcher optimum oscillator frequency

previous 1 2 3 next

Move Thread

LAN_403

Uspring

Sun Feb 17 2019, 05:33PM

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

There is a LTSpice simulator for a ring launcher in this thread. It might help to find an optimum L.

klugesmith

Sun Feb 17 2019, 07:10PM

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

Re. blast cage: why not a simple box filled with sand? Quarter-shrinking generates tiny high-velocity fragments, without much "blast" energy, IMHO. You could surround the box with bricks or concrete blocks for initial trials. Then vacuum out the sand to see how far the metal fragments went. If the sand is dry enough to flow, it's dry enough not to interfere with the electricity.

Bottom of "production" box could be funnel-shaped. After a shot, remove bottom cap. Sand and metal bits drain into your sifting screen, thence into open sand-storage pail. Ready to pour into loaded box for next shot.

Whether for shrinking quarters or launching disks, I think skin effect in copper wire is a secondary consideration. For thin wire, you get to use its full conductivity all the way to the middle. Very thick wire needs to be derated some for skin effect, but the factor is small unless wire is unnecessarily thick for other reasons.

1. The damping effect of wire resistance falls off when total circuit R is much lower than Z = sqrt(L/C) of tank circuit. I figure Z is about 0.2 ohms for the 8500 Hz example you gave, and for my can crusher. Quarter shrinking time (ref. Hackerbot highspeed video) is comparable to one cycle of undamped oscillation.

2. The wire needs to be sized to withstand I-squared-T value of discharge pulse, without getting hot enough to damage its insulation. Go work it out. I'm guessing 8 AWG is plenty, even for quarter shrinking, but you can check with Bert or Uspring. Hackerbot narration is very wrong when it says "this brilliant light is our work coil vaporizing". Torn to bits, yes. But the whole initial energy is too small to even melt more than a fraction of 1% of the work coil mass.

Here's how a quick & dirty spreadsheet responded to cutting and pasting:

var unit R_CC1 M_QS1
C uF 52 100
L uH 2 3.5
C F 0.000052 0.0001
L H 0.000002 0.0000035
Z ohm 0.196116135 0.187082869
T s 1.0198E-05 1.87083E-05
F Hz 15606 8507

MRMILSTAR

Sun Feb 17 2019, 08:07PM

Registered Member #62119 Joined: Sun Feb 04 2018, 04:59AM
Location: Cedar Rapids, Iowa
Posts: 136

The term "vaporized" is almost always used incorrectly. The work coil isn't turned to vapor, except for maybe a negligible part, it is disintegrated.

I have a lot of copper bus bar left over from other projects so that is what I plan to use even though it takes work to bend it around. I just have to keep things spaced far enough apart to avoid flash over.

klugesmith

Sun Feb 17 2019, 11:37PM

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

We use lots of custom sheet copper busbars at work, for single-digit voltages.

The inductance of a busbar pair is minimized by running the metal strips face to face, separated only by a thin insulating sheet.
Unfortunately that maximizes the force tending to tear them apart from each other, in a quarter-shrinker application.

You can compute the magnitude of that Lorentz force when you have a reasonable guess about peak current.

For better accuracy, get set up to measure the inductance of assembled busbar pair. Repeat after increasing the separation. From dL/dx you can get the force for a given current, just like people do here to for coilguns.

DerAlbi

Mon Feb 18 2019, 01:11AM

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

Ha, Uspring, i searched for my old work too, but since the forum is damaged, the whole thing is gone. Its a shame. Cant find it on my computer anymore.
But he would need Measurements of a fixed geometry already. From that he could however perfectly design the wire-thickness and with it the number of turns for his coil.
Again, its a shame the data is gone. (if its gone)

Signification

Mon Feb 18 2019, 01:45AM

Registered Member #54278 Joined: Sat Jan 17 2015, 04:42AM
Location: Amite, La.
Posts: 367

This reminds me of a magnetic firing circuit I started to play with years ago. It is a bit similar to yours, but fires a non-ferrous HD platter in a VERY aerodynamic way--like a Frisbee. The firing coils are connected in and mounted on two stiff boards with the projectile sandwiched in between. The high current pulsed coils tend to greatly attract each other when fired (an annoying disadvantage). The disc is in between them, but slightly offset toward the firing direction. This lets the pulsed field lines from the coils (when triggered) 'wrap' or 'connect' slightly together behind the disc. When fired (pulsed) the field lines go around the disc (due to Eddy) and then the disc is projected outward like a Frisbee at great velocity as the lines -reconnect-. I think it is called a RE-CONNECTION pulse. To add to the aerodynamic efficiency, I was trying to think of a way to give the disc 'spin' as it emerged. Be careful, as this could fire at lethal manner. I have even seen the non-aerodynamic (THUMPER) types embed the disc deeply into wood.

Uspring

Mon Feb 18 2019, 03:19PM

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

Oops, didn't see, that the attached files are inaccessible now. I saved a copy, though and can post it as ASCII text, which you can probably copy/paste, save as .asc file and then open with LTSpice. So if anyone is interested, I'll comply.

Does anybody know, if archive.org can be convinced to yield 4hv forum files?

Wrt to the optimal oscillation frequency: I believe it's best to match the first half cycle duration to the "takeoff" time of the disk. For Kizmos launcher, that time can be established from the simulation to be about 0.5ms, i.e. 1kHz. The force on the disk depends only on the energy stored in the coil, not its inductance. A too large inductance will take too long to charge up, so that the disk will have flown away before max current is reached. A too small inductance will cause unnecessary losses in the caps ESR and ESL.

DerAlbi

Tue Feb 19 2019, 10:24AM

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

Well.. I would be interested for sure

Thanks for your data backup.

Uspring

Tue Feb 19 2019, 11:21AM

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

I think it's a lovely little program. Since the LTSpice schematics (.asc) are simple ASCII texts, one should be able to import the text below into LTSpice. I had to modify it slightly, since the forum software doesn't like smaller and larger symbols.

Version 4
SHEET 1 2036 772
WIRE 1568 -256 1504 -256
WIRE 1504 -224 1504 -256
WIRE 1248 -176 1184 -176
WIRE 1184 -144 1184 -176
WIRE 1504 -112 1504 -144
WIRE 1008 -96 960 -96
WIRE 960 -64 960 -96
WIRE 1184 -32 1184 -64
WIRE 960 48 960 16
WIRE -304 80 -352 80
WIRE 448 80 400 80
WIRE 720 80 528 80
WIRE 816 80 720 80
WIRE -32 96 -64 96
WIRE 0 96 -32 96
WIRE 400 112 400 80
WIRE 816 144 816 80
WIRE -352 176 -352 80
WIRE -64 176 -64 96
WIRE 720 192 720 160
WIRE -704 208 -704 96
WIRE -64 272 -64 256
WIRE 400 272 400 176
WIRE 720 272 720 256
WIRE 816 272 816 224
WIRE -352 288 -352 256
WIRE -704 304 -704 288
WIRE 1120 384 1056 384
WIRE 1056 416 1056 384
WIRE 1056 528 1056 496
FLAG 400 272 0
FLAG 816 272 0
FLAG -352 288 0
FLAG -304 80 Inductance
FLAG -704 96 K
FLAG -704 304 0
FLAG 960 48 0
FLAG 1008 -96 Force
FLAG 1184 -32 0
FLAG 1248 -176 Velocity
FLAG 1504 -112 0
FLAG 1568 -256 Distance
FLAG 720 272 0
FLAG 1056 528 0
FLAG 1120 384 Efficiency
FLAG -64 272 0
FLAG -32 96 InductanceChange
SYMBOL cap 384 112 R0
SYMATTR InstName C1
SYMATTR Value {C} ic={CapVoltage}
SYMBOL res 432 96 R270
WINDOW 0 32 56 VTop 2
WINDOW 3 0 56 VBottom 2
SYMATTR InstName R1
SYMATTR Value {ESR}
SYMBOL bv -352 160 R0
SYMATTR InstName B2
SYMATTR Value V=L0*(1-V(K)**2)
SYMBOL bv 816 128 R0
SYMATTR InstName B1
SYMATTR Value V=V(Inductance)*ddt(I(B1)) + I(B1)*V(InductanceChange)
SYMBOL bv -704 192 R0
WINDOW 3 -44 159 Left 2
SYMATTR Value V=0.63238*2.718281828**(V(Distance)*-0.0305072k)
SYMATTR InstName B3
SYMBOL bv 960 -80 R0
SYMATTR InstName B4
SYMATTR Value V=max(0.5*V(InductanceChange)/(V(Velocity)+1m) *I(B1)**2, 0)
SYMBOL bv 1184 -160 R0
SYMATTR InstName B5
SYMATTR Value V=max(idt(V(Force)/DiskMass), 0)
SYMBOL bv 1504 -240 R0
SYMATTR InstName B6
SYMATTR Value V=max(idt(V(Velocity)), 0)
SYMBOL diode 736 256 R180
WINDOW 0 24 64 Left 2
WINDOW 3 24 0 Left 2
SYMATTR InstName D1
SYMATTR Value DIdeal
SYMBOL res 736 64 M0
SYMATTR InstName R2
SYMATTR Value {ESR*2}
SYMBOL bv 1056 400 R0
SYMATTR InstName B7
SYMATTR Value V=DiskMass*V(Velocity)**2 / (C*CapVoltage**2)
SYMBOL bv -64 160 R0
SYMATTR InstName B8
SYMATTR Value V=ddt(V(Inductance))
TEXT -232 -288 Left 2 !.param C = 94m
TEXT -232 -256 Left 2 !.param CapVoltage = 450
TEXT -232 -224 Left 2 !.param ESR = 10m
TEXT -232 -192 Left 2 !.param L0 = 8u
TEXT 312 472 Left 2 !.tran 0 5m 0 0.1u uic
TEXT -232 -152 Left 2 !.param DiskMass = 0.2
TEXT 376 16 Left 2 !.model Dideal D(Ron=0.1m Roff=1G Vfwd=0.4)
TEXT -248 -368 Left 6 ;Adjust Parameters
TEXT 832 -376 Left 6 ;Interesting
TEXT 1280 -216 VLeft 6 ;==
TEXT -1640 176 Left 6 ;Curve-Fitting for K in \ndependence of the \n"V(Distance)" im m
TEXT -904 352 Left 6 ;==
TEXT 1448 408 Left 6 ;Interesting
TEXT 1288 408 Left 6 ;==
TEXT 128 -288 Left 2 ;Capacity
TEXT 128 -256 Left 2 ;CapBank voltage
TEXT 128 -224 Left 2 ;Overall ESR
TEXT 128 -192 Left 2 ;Inductance without any Disk
TEXT 128 -160 Left 2 ;in kg
TEXT -248 -56 Left 2 ;.step param ESR 10m 190m 10m
TEXT -248 -88 Left 3 ;Step Parameter to check influence
TEXT 1016 -128 Left 2 ;in N
TEXT 1280 -144 Left 2 ;in m/s
TEXT 1592 -224 Left 2 ;in m
TEXT 856 256 Left 2 ;Learned from Uspring.
TEXT 984 48 Left 2 ;Learned from Uspring.

DerAlbi

Tue Feb 19 2019, 11:30AM

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

Thank you. Putting it into a Text-File and renaming it to .asc works. The fact that the coupling factor is curve-fitted in the model is a bit user-unfriendly. Should have used a table lookup.

previous 1 2 3 next

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