Small Coilgun Project with Trigger Circuit / Electromagnetic Projectile Accelerators / Forums

Forums

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

« Previous topic | Next topic »

Small Coilgun Project with Trigger Circuit

previous 1 2 3

Move Thread

LAN_403

Barry

Thu Jan 31 2008, 02:11AM

Registered Member #90 Joined: Thu Feb 09 2006, 02:44PM
Location: Seattle, Washington
Posts: 301

Random comments which I hope you'll take in the intended spirit of being helpful...

The current schematic is hard to read, it would really help us if you re-draw the schematic with a bus bar (straight line) for the +3v supply, this will make the common connection to the firing switch obvious, and show how the timing sections are independent. Keep things flowing left-to-right as much as possible. Inputs on the left, outputs on the right. Avoid crossovers, too.

Please show the charging circuit connections. The schematic I see still has both ends of the capacitor above 0v (ground), making it unnecessarily difficult to charge.

I still recommend replacing the open-loop timing with position detectors, even a simple wire brush contactor will work better than an open-loop design. Otherwise your SCR and NPN transistors timing will depend on ambient temperature. But if you can't bear to dump the timers then ...

I recommend replacing the RC-BJT switchgear with 555 timer chips for reliability and controllability, and they can source up to 35 to 200 mA (depending on model chosen) which can drive the SCR gate directly, if you have your heart set on 2n3055 then read on...

If I read your schematic correctly, then for transistor biasing purposes I think each stage looks like this, right?

Here, the 10K base resistor serves no purpose and will probably limit the base current too much, so I think you can replace it with a 0-ohm resistor. You can get 0-ohm resistors from Radio Shack for only a dollar, lol.

Now the timing depends on the RC time constant of the 10uF cap and the 0-100K variable resistor. To estimate the minimum resistance, let's assume the transistor switches in 1 RC time, and the lowest time needed is 1 ms. Solve for R(min) = (1 ms) / (10 uF) = 100 ohms.

To avoid burning out the base junction, replace the variable resistor with a 100 ohm resistor in series with an ordinary potentiometer. Otherwise if you accidentally turn it to zero, the 2n3055 will release its factory-installed blue smoke.

What's the maximum expected resistance? Use the maximum time, let's just say it's 100 ms. Now solve for R(max) = (100 ms) / (10 uF) = 10K. So we can expect a 10K pot to be all you need, instead of the 100K pot shown.

Will the SCR turn on with the 10K pot set to maximum? Let's look at the base current and the transistor gain, and see if that yields enough collector current for this SCR. The base current
Ib = (3v - 0.7v) / (10K) = 0.23 mA. Stay with me here, we'll get there. The 2n3055 is a power transistor with low gain, perhaps a beta of 25 is typical. The collector current is the base current times the gain, or Ic = (0.23 mA) x (25) = 5.75 mA. Hopefully this is enough to turn on the SCR, but it's pretty low and might not be reliable. If it's a problem, increase the timing capacitor so you can get by with a lower-resistance timing resistor.

Modular building is usually good, but you really must keep extremely tight connections in the high-current loop and make sure those wires are short. A low resistance in this path is extremely critical to the results. Even a few milliohms will make a difference. I highly recommend soldering everything in the high-current path.

Cheers, now go spend some quality time with that schematic for us, will ya?
Barry

Shaun

Thu Jan 31 2008, 02:19AM

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

What are you charging the capacitors with?

Backyard Skunkworks

Fri Feb 01 2008, 04:15AM

Registered Member #1262 Joined: Fri Jan 25 2008, 05:22AM
Location: Maryland, USA
Posts: 451

Thanks a ton for all the advice Barry+Shaun! I figure I'll go ahead and replace the 10k resistors with 100s like Barry said, the time delay will never be set for very long so things should work for firing the SCRs, if not I'll just use some less beffy, higher gain transistors to switch them. I'd say its about time to start building!

I've gotten pretty good at tuning 100k variables, within about 50 ohms! (using a gear reduction system) So I might manage to get away with the same componets here, otherwise I'll just get some 10ks. Seeing that such tiny amounts of resistance matter, I think I'll go ahead and try oil cooling again, my experiment got about 15-20% less resistance at -20C then room temp.

For charging I'll be using a varible bench power supply, set to 50VDC. I'll just manually touch the leads to the caps for charging.

BTW @Barry, There really are 0-ohm resistors costing $1.09

Don't believe it, look at mouser! (there really 0.003 ohm for high power applications)
http://www.mouser.com/Search/ProductDetail.aspx?qs=eOoTQizOqk%252bytC7GgOH0kA%3d%3d
Unless your buying these for their intended high power application, your just plain gullible

Barry

Sat Feb 02 2008, 02:27PM

Registered Member #90 Joined: Thu Feb 09 2006, 02:44PM
Location: Seattle, Washington
Posts: 301

Better read that again, BS ... I said to replace the 10K with a wire, and to replace the 100K pot with a series combination of 100 ohms and a 10K pot. But that comes after some more important suggestions.

Forget about chilling the coil. Your open-loop fixed-timing design will be so touchy and the results so variable that you'll never get around to optimizing the coil resistance. I've been there done that. The forum comments have been unanimous -- an open-loop design doesn't work with three stages, much less five. You need some type of position detector in a multistage coilgun design.

Barry

Backyard Skunkworks

Sun Feb 03 2008, 12:30AM

Registered Member #1262 Joined: Fri Jan 25 2008, 05:22AM
Location: Maryland, USA
Posts: 451

I forgot to mention that I'm also replacing the 2N3055s with some less beefy higher gain (100hFE min) transistors, as far as the open loop thing it is only four stages now so I figure I can get it working OK. As soon as I have it working then I might start tweaking or doing modifications to get it running more smoothly.

EDIT:
These transistors are KSC2328A, as usual Mouser had them on the cheap! They switch on at 1.0VDC so if my timing circuitry is running at 3VDC and the main timing resistor is 100k ohm, a 10uF timing cap will charge to 1v in about 500ms at full resistance. Lowering the resistance will give a tunable time delay of under 100ms and given how high the gain is on these transistors they wont have any trouble switching my SCRs. I'm now building the thing and will report back soon with results.

I know open circuit coilguns can be a real hassle but I figure it'll still be fun. I'll still keep things semi-modular so I can fire the caps single stage.

Backyard Skunkworks

Thu Feb 14 2008, 04:28AM

Registered Member #1262 Joined: Fri Jan 25 2008, 05:22AM
Location: Maryland, USA
Posts: 451

Well I just fired my coilgun single stage configuration at 30VDC (33J, about 40% power), I used a bundle of 50 1N4007 diodes for 1.5kA max spikes in anti-paralell (like that'll ever happen, but I do hate explosions and dead silicons). For a coil I used 18 gauge non-magnet wire with pretty thin insulation in a 3" long coil. Ammo was a 105mm long section of coat hanger wire. Results were pretty good, I got the wire to leave the barrel at roughly 4 m/s (eyeballing it) and it made a nice dent in my box three feet away. I can't wait to fire this thing at full power and with multiple stages...

BTW I've opted for mechanical timing, the projectile acts a contact in a swtich by sliding across hair thin copper wires inside the barrel, these will trigger the SCRs. I think errosion of the copper shouldn't be that bad as it's unlikely to get above 15 m/s.

Edit: The gun fired at a couple joules ballistic, but it was pretty lame. I ended up rebuilding it many months later and I'll have it up on my site...

previous 1 2 3

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