Inductive Coilgun Trigger Success! / Electromagnetic Projectile Accelerators / Forums

Forums

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

« Previous topic | Next topic »

Inductive Coilgun Trigger Success!

Move Thread

LAN_403

Dave Marshall

Sat Nov 07 2009, 05:10AM

Registered Member #16 Joined: Thu Feb 02 2006, 02:22PM
Location: New Wilmington, PA
Posts: 554

I've been kicking around the concept of an inductive coilgun trigger since my last major coilgun build in 2002/2003. In that time, I've seen a few attempts, but all have failed hard, primarily because of a serious vulnerability to induced current on the sensor inductor from the propulsion coil causing false triggering.

Early attempts at detecting the magnetic field of the projectile (like a moving transformer core), changing the frequency of an LC oscillator, and mimicking a metal detector circuit all met with exactly that problem. They worked fine, until you fired the coilgun and the sensor inductor wrapped around the barrel detected every pulse as a projectile.

I was pointed towards the metal detector method by Dave Billington several years ago, but there were a few problems with it.

It turns out that we weren't so far off with the metal detector thing. We just had it backwards. A traditional metal detector typically uses an RF oscillator, which is passed through an LC tank circuit acting as a band reject filter. When no metal is present, the filter resonates at the frequency of the oscillator, attenuating it severely. When metal is introduced near the sensor inductor in the head of the metal detector, it pulls the inductance up, changing the resonant frequency of the filter, and moving the reject-band, allowing RF current to flow to a detector circuit that starts beeping. You can see the problem here though. The detector isn't sophisticated enough to tell the difference between the RF signal, and a transient spike induced by a huge electromagnetic pulse nearby. A voltage spike is a voltage spike.

However, if you switched out the band reject filter for a sharp edged band PASS filter, it can work. In my circuit (schematic below) the 10MHz oscillator provides a stable RF input to a series LC tank circuit acting as a band-pass filter. With no projectile present, the filter is tuned to resonate at the oscillator's frequency, and passes the RF to the rectifier/detector circuit unimpeded. When a projectile enters the tank inductor (wrapped around the barrel), the coil's inductance spikes, causing the tank's resonant frequency to dive significantly. This causes dramatic attenuation of the RF signal, which is detected by the voltage comparator on the output of the rectifier (in this case, the rectifier circuit is identical to an Envelope Detector used in some radio receivers). This drop in voltage (from 3v to .8v in my unoptimized circuit) triggers a change in state of the comparator, triggering the next event. By using the comparator any transient spike that might appear on the sensor coil is simply ignored, as it only serves to push the input even further above the reference voltage. Only when the RF drops off and the rectified signal falls below the reference voltage does anything occur. The only possible problem could be if one of those spikes managed to occur at precisely the right time, and last precisely long enough, masking the projectile's presence in the coil. I have yet to manage to replicate this circumstance, but I would be very surprised if it couldn't be avoided with some careful tuning of the reference voltage and tank circuit.

I have to thank Chris Russell for his help bringing this idea to fruition. He came up with the envelope detector from scratch, not knowing that a design was already out there, and hit the nail on the head. He also helped me troubleshoot a bunch of really strange problems with it during prototyping.

A schematic is attached below, but I must warn you, its probably not usable yet for most of you. The only 10MHz oscillator I have access to right now is an odd one, and does not source current, it only sinks it. This required basically inverting the whole circuit. The comparator is also very poorly suited to this application, having an open collector output and a terrible response time. New ultra-fast comparators are on the way now, as are new oscillators, at which point I'll brew up a no kidding schematic that anybody can use. I'm also hoping to ditch the hard to find and expensive variable capacitor and replace it with a varicap diode, or better yet, a ceramic disc.

Some pros to this approach:

No drilling or slotting your barrel. It will work with any non-magnetic barrel material, including brass/aluminum with a little tinkering.
Crazy fast. The math suggests that without breaking the bank, this design can track a 1" projectile at Mach 3 with a 1" long sensor coil.
Versatile. With careful design and a little isolation, the sensor coil can actually be placed under the propulsion coil.
Very sensitive. Any remotely ferrous material will set it off if its properly adjusted.

Some cons:

A little more complex than a photo diode trigger, and has a few somewhat exotic parts (that can still be sampled!)
Higher per-stage cost. If every stage has its own oscillator, it appears it will cost about $4-$6 to buy all the components.
Requires a little more tinkering. Each tank circuit has to be carefully tuned to resonance upon installation.

This trigger has opened up the possibility of a huge new project for me, and I'll be undertaking my first major coilgun build in 8 years starting this week. Some highlights include Neodymium projectiles, a carbon fiber barrel, inductive triggering, 100% custom machined components, and a completely new method of propelling the projectile that promises significant efficiency gains. Keep an eye out for the project thread!

-Dave

aonomus

Sat Nov 07 2009, 06:17AM

Registered Member #1497 Joined: Thu May 22 2008, 05:24AM
Location: Toronto, Ontario, Canada
Posts: 801

I'm still eagerly awaiting more results, but it seems promising so far! This will be a nice development over optical triggers, particularly in the realm of sensitivity and requirement for shielding/protection.

One question though is whether each stage will require its own independent 10MHz source? I can see the attenuation triggering other stages if its all running off one 10MHz source...

This is the perfect way to celebrate your birthday though too!

Dave Marshall

Sat Nov 07 2009, 01:56PM

Registered Member #16 Joined: Thu Feb 02 2006, 02:22PM
Location: New Wilmington, PA
Posts: 554

I think it should be quite possible to use a single amplified RF source with a few low value 5% carbon resistors as a power divider of sorts to drive a number of stages.

The fact that the resistors in the circuit, as well as the comparator input are all very high impedance means that it barely sinks any current at all regardless if the state of the tank circuit, but some small fluctuations could be expected.

-Dave

rp181

Sat Nov 07 2009, 05:17PM

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

Great work!
It seems like one of those "Duh" things that no one thought of.
I was wondering, could the same method be used to detect when the pulse begins sloping off, as to cut the power, and restore the energy? Basically a combined sensor to function as stage triggers in a mulistage design, and as a trigger for a half bridge. I know that the current slope can be detected with current sensors, but I think combining them would provide very good integration.
Looking forward to the final product.

PYLEMET-mk2

Fri Nov 13 2009, 08:26AM

Registered Member #1513 Joined: Sat May 31 2008, 12:08PM
Location: Russia Klin
Posts: 48

Dave Marshall, why don't you use power coils(which accelerate a projectile) as sensing? If we have such construction of coils
931115

second coil maybe will "recognize" projectile when it's time to accelerate it

Slid

Wed Nov 18 2009, 01:13AM

Registered Member #853 Joined: Thu Jun 21 2007, 03:08PM
Location:
Posts: 14

This might be stupid, but why can't one just use a counter of some sort like 4017 and let it count each pulse it gets. Then the pulse codes can be arranged to fire at the correct count, just like fire at output 2, 4, 6 and so on..?

Dave Marshall

Wed Nov 18 2009, 04:37AM

Registered Member #16 Joined: Thu Feb 02 2006, 02:22PM
Location: New Wilmington, PA
Posts: 554

Slid, I attempted this with some of my early designs. It worked to some extent, but inductive kicks are unpredictable in nature. Damping, proximity to the inducing field, coil geometry, etc all made it difficult to know for sure how its going to act from shot to shot. I went with this design because of its very high degree of repeatability and predictable operation.

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