180J Half-Bridge Coilgun Complete / Electromagnetic Projectile Accelerators / Forums

Forums

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

« Previous topic | Next topic »

180J Half-Bridge Coilgun Complete

1 2 next

Move Thread

LAN_403

CyrusTheGreat

Mon Jun 02 2008, 01:13PM

Registered Member #537 Joined: Sun Feb 18 2007, 07:22PM
Location:
Posts: 10

Hello all, long time lurker, first time poster here. This is also my first coilgun and my first electronics project, so don't bite my head off, please.

I've been waiting until I got this whole thing up and running and done with the science fair circuit before I posted this here. (It took 3rd place in Electrical and Mechanical Engineering at the Intel ISEF) But now I've finished it with a ton of help from the archives on this forum, and the rest of the coilgun-related sites on the internet.

Anyway. This is a single stage diagonal half bridge coilgun. The efficiency is 4.66%. It runs a 1880 microfarad capacitor bank charged to 437 volts. My peak kinetic energy was around 3J, while my peak velocity (not coinciding with my peak kinetic energy b/c it was a different projectile) was 51 ft/s.

This is my circuit diagram

Many thanks to aahz for creating an awesome boost converter.

I'm using a ir2101 high voltage gate driver to drive two toshiba MG400Q1US41 brick IGBT's. I got really freaking tired of replacing my IGBT's (I was using discretes that didn't like to current share), so I got 4 bricks for 50 bucks. I highly reccomend looking on ebay to get some cheap ones.

Through my research, I noticed that changing the coil size (# of windings) didn't have a whole lot of effect on the peak efficiency (has anybody else noticed this), i.e. when I've searched through all the possible combinations for pulse time and starting position. So I created an equation to model the force exerted on the projectile by the coilgun.

2f29facadbfa08f883ff335uc9

Where:
F = the force on the projectile
cursive l = the length of the coil
C = capacitance of the bank
t = the time since the coil has been turned on
x = the distance of midpoint of the projectile from the midpoint of the coil
V = the initial voltage of the capacitors
L = the inductance of the coil
R = the resistance of the coil
mu sub r = the relative permeablity of the projectile

Notice that the only place the term L shows up is in the decay constant and the angular frequency. If anybody wants to see how I got the equation, I'll post the derivation. I also noticed that my peak efficiency (4.66%) occurred with the longest coil (an accident in winding). The coil was approx 1/4 in. longer and the efficiency was 0.3% higher. (the other coils were 1.25 in. long) If anybody wants to see my data and/or the research paper that I wrote, I'd be happy to send it to you.

This equation is really not very useful for finding anything besides the rough proportionalities of the starting parameters on the projectile. It cannot be solved for kinetic energy or efficiency, because it's a 2nd order differential equation that is not one of the types that can be solved using calculus. It's probabbly possible to solve it by summing up forces and obtaining a rough estimate of hte integral, but it's not going to be very precise.

I swept 4 coil sizes (540 turns, 400 turns, 200 turns, and 50 turns) pulse times from 0 to 10 ms and all the projectile positions where the projectile got enough speed to trip my chronograph. I've got some nifty 3D graphs to show all this if you guys want to see them.

I also tested a couple of different projectiles, they were all junk except for the solid steel. I also tried running this thing a PLIM, and ran into the same problem as Braam Daniels, you just can't get the current high enough without exploding something to get the projecties to move significantly. Not to say that they didn't move, I got a coil of wire shorted back on itself to move about 4 cm. I also tried solid and hollow aluminum projectiles which were even more junk than the coil of wire.

If I think of anything else, I'll post it.

my email is: the coil gunner atsign gmail dotsign com (remove the spaces and change the at and dot signs)

Criticism is welcome,
Duncan Townsend

Bjørn

Mon Jun 02 2008, 05:33PM

Registered Member #27 Joined: Fri Feb 03 2006, 02:20AM
Location: Hyperborea
Posts: 2058

The efficiency is 4.66%. It runs a 1880 microfarad capacitor bank charged to 437 volts. My peak kinetic energy was around 3J

These numbers does not add up. 3 joules would give less than 4.66% efficiency.

CyrusTheGreat

Mon Jun 02 2008, 06:16PM

Registered Member #537 Joined: Sun Feb 18 2007, 07:22PM
Location:
Posts: 10

Yeah, with a half-bridge, all the energy in the capacitors isn't used up, so the kinetic energy of the projectile divided by the difference in potential energy in the capacitors is 4.66%. With SCR drive coilguns you don't have to account for the difference because all the energy in the capacitors is dumped into the coil. You have to take into account the energy that the diagonal half bridge returns to the capacitors in the efficiency, otherwise you're not getting the whole picture.

flannelhead

Mon Jun 02 2008, 06:25PM

Registered Member #952 Joined: Mon Aug 13 2007, 11:07AM
Location: Finland
Posts: 388

I'm just curious, but what exactly is the advantage of using a half bridge instead of a SCR?

CyrusTheGreat

Mon Jun 02 2008, 07:08PM

Registered Member #537 Joined: Sun Feb 18 2007, 07:22PM
Location:
Posts: 10

The advantage to the diagonal half bridge is that it's a turn-off design, so you don't have to be too precise with how you construct your coils, you can shut off the current through them before it drops to zero, unlike an SCR. The main reason for using one, though, is that as the magnetic flux collapses at the end of firing, with an SCR design the energy either returns to the capacitors with opposite polarity (unusable) or is wasted as heat. With a diagonal half-bridge, the magnetic flux collapse biases the freewheeling diodes and recharges the capacitors some. So you don't get higher kinetic energy per shot (well maybe you do because the coil can be turned off to minimize suckback), but you get less energy wasted and a much lower charge time because the capacitors are already partially charged.

EDIT: I probably don't have to say this here, but just to make sure I have my bases covered: I did not design the diagonal half bridge. I think that Yohan did. Jonathan Filippi also built a really nice half bridge coilgun.

OZZY

Mon Jun 02 2008, 09:03PM

Registered Member #511 Joined: Sat Feb 10 2007, 11:36AM
Location: Somerset UK
Posts: 55

Hi Cyrus

Thats a very nice coilgun, it`s much more advanced than most science fair coilguns. Also 4.66% is much more efficient than most single stage guns. I would be interested to read your research paper.

Your bridge looks very unbalanced, you have huge IGBT modules and tiny diodes, this tells me something is wrong. Did you destroy lots of smaller IGBTs and then just fit bigger ones? Did you ever find out what was causing the IGBT failures? Some things I have noticed from your circuit diagram.

1. 1N5408 are standard recovery diodes and probably not fast enough, they will give you a high voltage spike accross the IGBT at turn off.
2. You have no gate drive resistors to provide damping.
3. you have no over voltage protection on the gates.
4. You have no snubbers on the IGBTs (if everything else is working properly you should not need them).
5. The long wires between components will add stray inductance and may cause voltage spikes during switching transients.
6. The Toshiba modules may contain some of these protective circuits so they have survived where discretes did not.

Take a look at this page mosfet switching

Ozzy

CyrusTheGreat

Mon Jun 02 2008, 11:22PM

Registered Member #537 Joined: Sun Feb 18 2007, 07:22PM
Location:
Posts: 10

Yeah, I've destroyed a lot of IGBTs. I had 3 smaller IGBTs in parallel to get a higher current carrying capacity, but they kept blowing, I think because they had a negative thermal coefficient, so one of them would get thermal runaway, then the others would blow because they couldn't handle the current. I got tired of trying to get the smaller IGBTs to work, so I bought the beefiest ones that I could find.

What speed diode do you recommend that I use? I wasn't aware that the difference between standard and high speed diodes would be significant enough to make a difference. The IGBTs that I'm currently using seem to be tolerant of these voltage spikes, though. They're rated for 1200 volts.

I didn't think that I needed a gate drive resistor. Could you please explain what purpose they serve? I just thought that you wanted to turn the IGBTs on and off as fast a possible, thus no gate resistor.

I was hoping that using the half bridge design would mean that I didn't need snubbers. The modules have an anti-parallel diode, so that may help them to avoid blowing from reverse voltage, but as far as I know that's all they have.

I was worried about the long wires, but I didn't think that they would matter that much considering that I'm driving a coil that has a much higher inductance. I guess that the next revision will have either shorter or flat wires to limit the stray inductance.

As I said, this is my first electronics project, so I probably missed a lot of the finer points of circuit design. Just so you know, I'm not a complete newbie, I've been working on this thing for three years.

I'll email you the research paper as soon as I get the chance

I just thought I'd add a few more thoughts on the equation that I created. It may not be useful for calculating the exact efficiencies of a given design, but it does let you know how to optimize your design if you're starting from scratch. The commonly held knowledge that the pulse time should be pi*sqrt(L*C) holds true. Although the windings on the coil don't matter that much (for efficiency, if you want kinetic energy, it is crucial), the coil geometry is rather important. The L/R constant should be maximized as much as is possible because it becomes the decay constant for the force on the projectile. Another interesting bit is that the length of the coil can increase the efficiency greatly: the longer the coil the more efficiency. Lengthening the coil should also increase the kinetic energy, if the starting distance of the midpoint of the projectile to the midpoint of the coil is kept the same. The x^3 term in the denominator also shows why the starting position is so crucial.

Thanks,
Duncan Townsend

Shaun

Tue Jun 03 2008, 12:12AM

Registered Member #690 Joined: Tue May 08 2007, 03:47AM
Location: New Jersey, USA
Posts: 616

The first thing I would say is absolutely tighten things up in the high current section; use shorter, thicker wire. Other than that its pretty awesome. You've obviously put a lot of thought into this; and theres far more math behind it that most coilguns ive seen.

sofi

Tue Jun 03 2008, 02:12PM

Registered Member #1464 Joined: Sat May 03 2008, 10:56AM
Location:
Posts: 9

Great, at last one who build IGBT halfbridge stage.
I'm looking for your solution of charging bootstrap capacitor. If I understand this, before you push fire button S2, you must push S4 to charge bootstrap cap through low side IGBT. I have the same problem and I put 15kohm resistor between pin 6 of IR2101 and ground. Now the capacitor is always charged and during firing only small current (390V/15kohm=26mA) flow through resistor.

But in your case I must recommend another driver - IR2110. It has ten times higher output current (short circuit max. 2,5A) then IR2101 (max. 0,2A) and use fast rectifier diode like UF5408 (peak current 150A, 50ns).

Also building driving electronic on "contact field" (or how it is called in english) is not good idea if you working with high voltage (in addition pulsed HV).

rp181

Tue Jun 03 2008, 05:10PM

Registered Member #1062 Joined: Tue Oct 16 2007, 02:01AM
Location:
Posts: 1529

do you mean breadboard (the white thing)?

1 2 next

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