Detecting A Bullet

[EasyGoing1]

Hello,

Just a general question hoping someone can point me in the right direction.

If I wanted to detect, electronically, whether or not a bullet hit a target. What kind of circuit or electrical / analog set up could I use?

I want to use an Arduino to read whether or not someone hit a target within the overall boundaries of the target at fairly close range (3 to 7 feet, maybe 10). Someone mentioned to me something about using a large loop of copper wire that generates a weak magnetic field that the bullet would interrupt while passing through the loop but I have no idea what kind of concept this is nor where to find any details on it. I've googled around, so now I'm asking in here.

Thank you,

[stj]

Re: Detecting A Bullet

put a microphone or peizo sensor on the target frame.

[eccerr0r]

Re: Detecting A Bullet

I'd ask for more details else I'd go for an expensive optical solution because I don't like the idea of my electronics creations getting shot at...

[EasyGoing1]

Re: Detecting A Bullet

What about this whole idea of some kind of wire that is wrapped multiple times in a circle then the bullet creating a magnetic ripple in the wire and detecting that? Anyone know how something like that works or what principle it is?

[eccerr0r]

Re: Detecting A Bullet

Okay, to be explicit for more details...

Do you care where (X,Y) in the region the bullet goes through? (i.e., are you using this as target practice?)

How big of an area?

Can you tolerate using a magnetic field that may alter the trajectory as the bullet goes through?

Will there be some other backing (target paper) or will the area need to be free space?

[EasyGoing1]

Re: Detecting A Bullet

Okay, to be explicit for more details...

Do you care where (X,Y) in the region the bullet goes through? (i.e., are you using this as target practice?)

How big of an area?

Can you tolerate using a magnetic field that may alter the trajectory as the bullet goes through?

Will there be some other backing (target paper) or will the area need to be free space?

Excellent questions!

So the targets will be fairly close to the shooter ... the furthest distance may be 10 to 15 feet away at the most.

They are humanoid style targets, so the area we need to detect would be approximately a circle (or maybe oval if it doesn't prove to be too difficult) that would essentially cover the chest region of a 6'2" human male.

Altering the trajectory of the bullet within reason is perfectly acceptable, I only care that the bullet gets through whatever I'm going to use as the sensor. Doesn't matter if it hits dead center or the edge. I have no interest in where it hits, only that it hits a defined region of the target.

The targets will have some backing, but there is concern that as they get used, bullets CAN - on rare occasion - pass through the same point which seems to eliminate the option of using any kind of a concussion phenomena to detect the bullet as has been suggested in this post already (no audio option where a bullet might hit a board or something). The desire is definitely to employ some kind of an electromagnetic field that could be disrupted by the bullet and detected.

Also, we aren't interested in the speed of the bullet (in case anyone was wondering about that). We only care about whether or not it passed through the sensor region.

Thank you very much,

[eccerr0r]

Re: Detecting A Bullet

I'd going back to using a very fast (and expensive) camera system to detect it.

Normally you'd use some sort of metal detector system like what you'd use going through the airport but these things would be sensitive to both "inside" and "outside" of the loop. Plus if someone hits the loop, it's time to get a new one.

[eccerr0r]

Re: Detecting A Bullet

If this is indoors under consistent lighting, the other option is perhaps light pipe/fresnel lens and trying to detect minute light differences much like PIR, except it must be active light detection to reduce scan costs. Have two of these, one in X and Y direction; and if both X and Y detect something, then we have the bullet passing through that region. I think this can be made small enough that it might be reasonable enough to protect it from stray shots. Then again if you can move the sensors far away from the target hit box region then protection might not be an issue.

Light pipe will be necessary with coherent light (probably low power laser) to not affect the other sensor.

This is not going to be an easy sense but much cheaper than the camera system, and it would be a square/rectangular sense region. If you do a 3x3 matrix or more of these, then you can have even better resolution of where the shot hit.

[EasyGoing1]

Re: Detecting A Bullet

Normally you'd use some sort of metal detector system like what you'd use going through the airport but these things would be sensitive to both "inside" and "outside" of the loop. Plus if someone hits the loop, it's time to get a new one.

I thought about that. I'll talk to the guy who wants this done about protecting the sensor and see what his thoughts are on it, but as far as detecting a bullet that passes outside the target zone, I think that's not so relevant because what we're looking for is the shooters response time to an event ... a light turns red, the counter starts, they draw their firearm and shoot. The sensor records the bullet then the time delta is recorded.

Do you have any more information on the 'metal detection' method you're thinking about? I kinda need to do this on the cheap and was thinking if I needed to just get some transformer wire and create a loop using many wraps of wire, that would make the most sense in terms of cost ... what I'm unsure of, is how the loop should be constructed (wire type and gauge, number of 'wraps' etc.) and also the circuit required to both generate the field and detect any changes in the field. I'm assuming that I need something based on a 555 perhaps, and some opamps to amplify the signal while reducing noise etc. but for all I know, there may be an IC out there somewhere that performs these functions better than I could with a constructed circuit ... just looking for as much education on this as I can get my hands on.

I've already got the Arduino software written using buttons as mock bullet detection sensors, and it's working pretty well. It detects response time at millisecond accuracy.

Thank you again,

[eccerr0r]

Re: Detecting A Bullet

See the problem is that if the bullet passes outside of the loop, it could also get detected as inside as it can't really tell. So unless you have a really big target zone and don't care about accuracy, then possible (but I suspect it defeats the purpose of such detector).

And no, 555 will not work. Because the bullet is traveling so fast, you need to be in the MHz region at least, need to do some math for feasibility here...

I still say light. Bullets may be fast, but do not get close to the speed of light.

---

Thinking out loud here. I don't have any documentation of metal detection handy, this is all done in theory, at least if you haven't seen any documentation about it, it might not actually be feasible unless tried. The back of envelope calculations here say that a 1000m/s bullet that's about 4cm long or so traveling near a metal detector will pass through it in 10µS or so. The detection wavelength needs to be faster than this so likely at least 100KHz or so...

I'd be using a double loop "air transformer" and the bullet would temporarily become a metal core. We'll be detecting the permeability change of air versus a metal object. However the object is small, so we'll need to pump up the power such that it can be detected... this would require quite a bit of power that could possibly generate RF noise...

---

I still say do it via light. Light doesn't care about getting hit by bullets.

[EasyGoing1]

Re: Detecting A Bullet

See the problem is that if the bullet passes outside of the loop, it could also get detected as inside as it can't really tell.

This isn't really a problem. It's easy to see whether or not the shooter hit the target. If they didn't, then the results of the detector will be ignored. What cannot be done visually by a human, is determine how quickly the shooter responded to an event that is generated by the microcontroller, so all I need, is the shooters response time to an event ... I don't need the technology to confirm that they actually hit the target.

The back of envelope calculations here say that a 1000m/s bullet that's about 4cm long or so traveling near a metal detector will pass through it in 10µS or so. The detection wavelength needs to be faster than this so likely at least 100KHz or so...

I'd be using a double loop "air transformer" and the bullet would temporarily become a metal core. We'll be detecting the permeability change of air versus a metal object. However the object is small, so we'll need to pump up the power such that it can be detected... this would require quite a bit of power that could possibly generate RF noise...

After my last reply to you, I looked up some articles on metal detection, and I think we might be looking to implement a Colpitts Oscillator - which seems to be the most popular method of metal detection. From what I can tell, using a multi-wire loop as big as maybe 20 inches in diameter or larger shouldn't be a problem - as you said - with enough power. I'm sure a 12 volt power supply with a couple of amps capability would be more than enough. And 100kHz seems to be the common frequency chosen for these circuits.

I still say do it via light. Light doesn't care about getting hit by bullets.

I agree, but cost constraints would seem to limit this option unless I can use some kind of inexpensive LED solution that would cover a 20 to 25" radius ... off the top of my head, would be possible to use a couple of infrared LEDs with infrared receivers, then somehow send a known timing signal out of the IRLEDs then detect maybe a change in magnitude as the bullet passes through? Might even be able to use the Arduino as both the signal generator and the signal monitor at a high enough frequency to detect the change.

What are your thoughts on it?

Thank you again,

[eccerr0r]

Re: Detecting A Bullet

This isn't really a problem. It's easy to see whether or not the shooter hit the target. If they didn't, then the results of the detector will be ignored. What cannot be done visually by a human, is determine how quickly the shooter responded to an event that is generated by the microcontroller, so all I need, is the shooters response time to an event ... I don't need the technology to confirm that they actually hit the target.

If accuracy is not a concern, then the original solution by stj should work (though it need not be on the target)? If we don't care if we get a bullseye, then the loud BANG! noise generated by the explosion of the gunpowder can easily be detected via sound?

[EasyGoing1]

Re: Detecting A Bullet

If accuracy is not a concern, then the original solution by stj should work (though it need not be on the target)? If we don't care if we get a bullseye, then the loud BANG! noise generated by the explosion of the gunpowder can easily be detected via sound?

That's something I had not considered ... I thought sti was suggesting placing microphones onto the wood that the target might be mounted on to listen for the bullet hitting the target ... listening for the loud bang might be an idea ... but two things come to mind when I think about this a little more:

1) A bullet travels faster than sound ... so in theory, the metal detector would be more responsive and more accurate when determining the shooters response time to a light turning on as opposed to just waiting for the audio from the gun ...

2) We were at the range yesterday and I was wearing active audio ear protection which allows you to hear people talking until the ambient audio reaches a high enough decibel, then it kicks in some circuit that reverses the audio wave and cancels it out before it hits your eardrum ... $20 at Harbor Freight Tools ... they work amazing ... but the protective circuit kicked in a lot because of other nearby shooters ... and I'm thinking that such a solution would be difficult if not impossible to limit the detection to the shooter that I am interested in... and not some other shooter who is in the same range as me or the one next to me.

I suppose an argument could be made that if the audio equipment being used was sensitive enough or the circuit was discerning enough, then we MIGHT be able to tune the sensitivity so that only a nearby gun would provide enough amplitude and be much higher than the amplitude of near by shooters ... perhaps with some kind of a can or a tube that is pointed at the shooter that we're interested in that also has baffles built in that would block a significant portion of audio from nearby shooters ... hmmmm ... might have to experiment with that ... as audio would be much easier ...

But would it be fast enough ... need to run the calcs that show how quickly a sound wave would hit the microphone at ... maybe 20 yards max distance...

ok, some quick and dirty calcs show that an audio wave traveling at 343 m/s converts to roughly 375 yards per second ... just looking at it, I'm thinking it will be fast enough, but if I divide 375 by 1000, I get .375 yards per millisecond ... which when divided into 20, it gives me about 53.3 milliseconds of delay time between when the shooter pulled the trigger and the Arduino hears it ... that's an acceptable margin of error, because I could easily subtract it back out of the calculated time, and be even more accurate than if I didn't ... also with my tests using buttons instead of guns, I'm getting response times in the hundreds of milliseconds, so I'm sure that even 53 milliseconds would be low enough to call it good enough.

Another thought ... sound travels at the same speed all the time, no matter what caliber bullet / grain you're shooting ... so from that perspective, it would even be more consistent when looking for someone's response times where detecting the presence of a bullet might indicate a slower or faster response time depending on the bullet... so from that point of view, dealing with known and consistent variables like the speed of sound ... might end up yielding a better solution in the long run..

But the metal detector idea sounds funner ... but admittedly more difficult to do ... AND, I COULD tell the software to not listen for any sound until the correct light is lit at which point, id only be opening up the monitoring for less than two seconds so my chances of getting false positives from nearby shooters would then be limited to a significantly smaller time window GREATLY reducing my margin of error ....

I've got some of those really cheap and very common LM317 (i think opamps ... maybe if I could find a good schematic to start messing with I might be able to come up with something...

but I'll need a mic ...

(as a side note) Going to the range was a birthday gift to me - I got to shoot a brand new AR-15 at targets 120 yards down with iron sights - and I hit the paper plate with over 90% accuracy ... not bad for a guy who hasn't shot a gun in over 20 years. That AR-15 had ZERO recoil on the rifle ... that and I got to shoot a couple of hundred rounds of 9mm rounds while being trained by a certified NRA instructor (the friend of mine who took me shooting and who wants this project done) ...

Sorry for the long post today ... .did a lot of thinking out loud with my fingers it seems

Thank you again for taking the time to discuss this with me,

[eccerr0r]

Re: Detecting A Bullet

Will the shooter be in a stall? You can place it closer to the shooter than the target, or even have the sound detector in the pocket of the shooter and possibly wirelessly transmit the data as light is faster than sound.

The microphone wouldn't even matter much I think, you can use any old audio range speaker, piezo element, microphone, or whatnot. The gunpowder explosion causes so much noise whose amplitude is probably high enough to be picked up by nearly anything pointed at the right direction. But you will have to experiment to be sure.

[EasyGoing1]

Re: Detecting A Bullet

Will the shooter be in a stall? You can place it closer to the shooter than the target, or even have the sound detector in the pocket of the shooter and possibly wirelessly transmit the data as light is faster than sound.

No, the shooter will be standing outdoors ... desert kinda terrain. The wireless option MIGHT be dooable ... would have to explore that a little more. One thing I like about that idea is that if we used bluetooth, and there are multiple shooters - each one could get their own mic and I could track which shooter is doing what.

The microphone wouldn't even matter much I think, you can use any old audio range speaker, piezo element, microphone, or whatnot. The gunpowder explosion causes so much noise whose amplitude is probably high enough to be picked up by nearly anything pointed at the right direction. But you will have to experiment to be sure.

I just ordered these

and I saw these when I was browsing ... do you know anything about these motion sensors? You think they might be too slow or unable to detect something as small as a bullet?

[eccerr0r]

Re: Detecting A Bullet

Yeah the sound sensor probably will work fine. The PIR sensor is probably just not sensitive enough to detect a small bullet (they were designed to detect humans or parts of humans), but speed might be artificially reduced to filter out noise since humans usually don't move as fast as bullets. So you may still need to design your own.

Note that using bluetooth you will have to add more delay as BT also adds latency. Should be in the millisecond range however.

Gunpowder explode -> microphone sensor = speed of sound in air
microphone sensor -> initial processor = instantaneous
initial processor -> bluetooth = packetization latency+processing (~1-2 milliseconds)
bluetooth -> main processor = bluetooth and depacketization latency (~1-2 milliseconds)

[petehall347]

Re: Detecting A Bullet

ir beam might be quick enough ..may need a lot of beams though

[eccerr0r]

Re: Detecting A Bullet

IR (light) beam IS fast enough, but as said earlier, devil is in the details, hence abandoning it for sound. Need an intricate system of light pipes to make light work with a reasonable number of sensors (which cannot be scanned in an array, they need to be detected all in parallel).

[Andyy]

Re: Detecting A Bullet

I remember an episode of Mythbusters http://www.youtube.com/watch?v=2qY0DA7DG9s where they used a simple apparatus consisting of two pieces of foil separated by a thin insulator (like kitchen cling wrap film). When the bullet pierced the foil and insulator, the foil would make contact with the other foil, closing a circuit.

I'm not sure if it's adaptable to your scenario though.

[theojt]

Re: Detecting A Bullet

Not sure if it would work for you but TargetImpactSignal.com makes sensor that triggers a photo flash to indicate target hits. The cheapest one is about $99. Their sensor (could be piezo by the description) is $50. Theirs is designed to go on the back of a metal plate target but could possibly be adjusted for appropriate sensitivity to poster board/cardboard. I'd have a chat with them at least to validate the feasibility. I'm not affiliated, just read about them.

http://targetimpactsignal.com/how-it-works/