Chronograph for bullets

[(*steve*)]

ITime to get from 5 V to 0.8 is gonna be long compared to the accuracy that I want, 1us or less.

You're thinking TTL. These circuits wont work with TTL.

The rising edge is all you're interested in. The falling edge just needs to be fast enough to get to a valid logic 0 before the next pulse from the other sensor comes along. We're talking about 1 RC delay here, and with 10nF and 100k, that's 1ms. If you like, reduce those to 10k and 10pF and it will be 0.1uS. The speed of the rising edge is determined by the capacitance of your sensor and the inductance of your resistors.

If you use two SR flip flops a-la Rleo6965's circuit, then the fall time is no longer an issue. -- however you'll need to wire a slightly smarter reset button to reset one SR flip flop, and set the other (the sensors will set and reset them respectively.

 

[CDRIVE]

I want an 'instantaneous' measurement, which I am approximating as 1cm for now, and seriously considering 0.25 cm mostly b/c of the ease of detector application to front and back of the target.

Another issue with placing the sensors close to each other is that the fragments of the sensor might touch and you may get triggering anomalies.

Actually, it's more probable than improbable.

 

[wannabegeek]

Thank everyone for all the meaningful responses....I'll be thinking about all this over the next few days at least...

I would contend it is true everywhere. I'd also contend that the delta-v caused by having to penetrate one of your sensors is likely to be significantly more than the delta-v over a couple of feet in air.

I'm going to work out the math and show you why I think that's wrong and maybe prove myself wrong...I've got some numerical solvers for projectiles I wrote a while back stashed away.

I've done analytic solutions before for dropped objects as well. The velocity for dropping an object with air resistance, modeling the resistance as linear with velocity, was a hyperbolic solution. So perhaps for a short distance, the average velocity will be fine.

My detectors will hopefully be made from very thin copper trace printed onto a thin plastic decal. I think the effect on velocity will be negligible. But If I'm making the bullet pass through the 0.25 cm of cardboard, maybe Steve is right about the effect on delta v.

I re ran some tests last night. I had some bad clock connections and went back to using my function generator for testing until I can borrow a scope. So now my original detector design works better. But, if I jiggle the wires around, I get triggers, so the SR FF designs are still in.

Thanks to Rleo6965 for his input too, I like his solution.

As for fragments, yeah I thought of that and can only address that more when I get a good working trace-target detector made. I might be ordering some conductive copper spray paint soon and tediously preparing a template.

Gonna think this all through more based on your observations...thanks again,
wbg

 

[CDRIVE]

Wouldn't aluminum foil be cheaper? Make a cardboard or paper (insulator) sandwich. When the bullet pierces the sandwich it will close a circuit between the two aluminum sheets.

Chris

 

[wannabegeek]

Sorry Steve I misunderstood your post about a linear fit to the data between two detectors.
I agree that is true, if and only if the distance between the sensors is small. This falls apart when we want to know the velocity far from the origin of the reading, that is when we need more than just the first term from a Taylor Series expansion.

Note the plot I have attached. Scrutinize it...I am rusty at this. later during the week I'll have to go through the math and theory that describes the set up for the air resistance model that I am describing, and I'll post a summary with decent explanation ( hopefully ). The plot is a numerical solution to the equations below using Euler's method.

EDIT : the x ticks are off on the plot...gotta go for now tho...
EDIT2: So the xticks are correct, it's just that I forgot I'm completely guessing about the value of the drag coeff k, the purpose of this plot is to show that yes, there is a strong linear looking dependence on the velocity vs. distance, until a point where it falls apart....

d^2x/dt^2 = g - k/m * sqrt(v^2 * v_x )
d^2y/dt^2 = -g -k/m * sqrt(v^2*v_y)

k is a coeff of friction/drag.

Wouldn't aluminum foil be cheaper? Make a cardboard or paper (insulator) sandwich. When the bullet pierces the sandwich it will close a circuit between the two aluminum sheets.

Yeah that 's a good idea...I played around with that at one time in my head...gonna mock it up and test it. Was thinking to use plastic wrap rather than paper.

 

[CDRIVE]

Yeah that 's a good idea...I played around with that at one time in my head...gonna mock it up and test it. Was thinking to use plastic wrap rather than paper.

If this concept works reliably it will also have the advantage of being reusable. All it would take is a sharp hollow hole punch and a backer board to cut away any ragged holes made by the bullets. Your circuit could have a continuity function to insure that any shorts have been cut away. I figure you could put a bunch of holes in it before needing replacement.

 

[wannabegeek]

I had played around with the idea of making a capacitor out of the 'sandwhich', using an external inductor. V_in would be ac and at the f_resonant. When a hole is made, the capacitance would change, causing v_out to be lower. Using external C 's and L's, I might be able to make the Q factor tight enough to get an easy to detect change in v_out.

It would be fun to try at least. Gating would then depend on a phase detection circuit. i read about a cool simple phase detector in "The Art of Electronics" using an XOR gate.

For now I'd like to practice use of the SR FF, there's a lot there for me to learn.

cheers,
wbg

 

[CDRIVE]

If you decide to go the uC rout it's going to look very much like this. You appear to be very bright and I suspect you would have no trouble at all learning Picaxe Basic.

I know exactly what you're thinking... "OMG! That looks soooo simple!" Well, yes it does and yes it is. One caveat though. Not with the sensor spacing you want.

Chris

 

[(*steve*)]

that is when we need more than just the first term from a Taylor Series expansion.

Do some analysis of your graph.

Determine the maximum error between a straight line and the curve for a distance of 5 metres.

If that's not small enough consider 1m.

If that's not small enough, fit a curve to it.

The disadvantage of sensors widely spread is that you need to shoot through both of them, but it doesn't look like you're talking 400 yard range, so your target should be able to be pretty small.

 

[Rleo6965]

Please see drawing. Suggestion only.

Drawing of Bullet Sensor Speed for Start Sensor and Stop Sensor.

Initial position of Aluminum frame will cause Normally Open contact of Limit Switch to be Closed. Same position of initial status of Start and Stop sensor switch in my proposed circuit.

When bullet hit the changeable pad. Aluminum frame moves backward and release the limit switch. Causing to start the counting.

You can use wire mesh pad for changeable material to avoid false tripping of air pressure from gun powder explosion.

 

[wannabegeek]

Hi All....I'm glad everyone is still interested in this project !

@ Steve, will do

@ Rleo
Like the idea of the trap and your FF trigger circuit. One of the restrictions on the design h of the detector is that the range master must allow it. It might be hard to let them allow me to put a metal box on the ground behind my target, so I don't think I can do anything really different with the target.

This need to conform to the the typical shooting range makes things more difficult. That's why a short spacing on my original detector idea is appealing...

A short spacing is good partly b/c the effectiveness of barrier materials can be checked and because the range master will likely allow it Plus it's hard...

@ CDRIVE
I 'm super interested in the uC and the Picaxe. I suppose that the baud rate is the limiting factor. I know that National Instruments makes a uC PC interface card for event timing and they use an external crystal to the uC for the additional sampling clock resolution. If I could debounce and do a little signal processing with software ( assembler even ) that would rock. A mathematical 'analog' filter is not that hard to program and would also be good practice for me considering I want to be in the signal processing field.

First, I am determined to learn how to use FF and the digital components !!
I have some pics of my bread board set up...it's messier than I want....

TODO: List
For practice, I want to debounce my counter-display reset switch with RC debouncer.
Get NOR gates and make FF and test both Steve and Rleo's gate designs.
Fix my clock, it's only outputing 40 mV amplitude at correct freq of 10 Mhz.

I might need to try a Colpits oscillator, I don't have much experience with any oscillator.
The one I'm using is just an RC circuit with a crystal and a NOT gate for feed back. I'll post a link to the page of this thread with the PDF and the circuit.

I'm temped to try sticking a biased JFET amp on my existing clock and see if I can get it swing higher. I also need to practice basic amps....lots to practice.

I should be at work !

cheers,
wbg

 

[BobK]

I don't understand why you want to debounce. This adds an unpredictable delay to the signal. For the most accuracy, you want to trigger as soon as possible after the connection is broken and ignore any bounce.

Essentially you have 3 states:

1. Ready - both wires intact, counter is zero
2. Count - the first wire is broken the second is not, counter is running
3. Done - the second wire is broken, counter is stopped

In my opinion you want the state transitions to occur as fast as possible, not after debouncing.

Bob

 

[CDRIVE]

I don't understand why you want to debounce. This adds an unpredictable delay to the signal. For the most accuracy, you want to trigger as soon as possible after the connection is broken and ignore any bounce.

Essentially you have 3 states:

1. Ready - both wires intact, counter is zero
2. Count - the first wire is broken the second is not, counter is running
3. Done - the second wire is broken, counter is stopped

In my opinion you want the state transitions to occur as fast as possible, not after debouncing.

Bob

Can't seem to shake the debounce issue. Why? I don't know. That was the whole point of using the SRFF!

 

[wannabegeek]

Hi Bob...

I am going to use the SR FF b/c I don't know for a fact what is going to happen when a bullet tears through a conductor. Since bounce can occur on the micro second scale, I think it's safest to avoid it.

Also, this is an out-of-doors device with all kinds of unforseen stuff happening...if someone spends all day hand loading some special cartridges for a field test and then the measurement fails...well I don't want to be around when that happens...

Now as for the timing...

The the worst case delay for cmos gates is something like 90ns or less...
The fastest event I'm willing to design for is a bullet traveling at 1600 fps = 600 m/s
over the distance of 1 cm = 0.01 m.

if v = delta(r)/delta(t) then t = 0.01 m / 600 m/s = 16us
The max descrepancy will be around [ (16us + 200ns + 2ns)/ 16 us - 1 ] ~= 1 %

Not bad for a worst case... the 2 ns is the uncertainty in the clock gating since the projectile can be off by one pulse on both detects. 200 ns is worst case for the AND gate and two SR FF rated around 20 ns and 90 ns.

please tell me if I'm makgin any huge mistakes here...this is how I've been thinking about it for a while and it's easy to fool one's self...

wbg

PS The debounce I mention in a post this morning is just for the counter-display reset, not the gate rest....will edit that post now.

I don't understand why you want to debounce. This adds an unpredictable delay to the signal. For the most accuracy, you want to trigger as soon as possible after the connection is broken and ignore any bounce.

Essentially you have 3 states:

1. Ready - both wires intact, counter is zero
2. Count - the first wire is broken the second is not, counter is running
3. Done - the second wire is broken, counter is stopped

In my opinion you want the state transitions to occur as fast as possible, not after debouncing.

Bob

 

[Rleo6965]

I think we don't have problem with mechanical and cmos propagation delay. As long that we use same Limit Switch for 2 sensor.

I'm not very good in explaining. Hope illustration below will help.

When bullet hit Start Switch. The Counter will not start counting because of mechanical delay of transition of contacts and the cmos propagation delay. After the first delay Counter will start counting. When bullet hit stop sensor. Counter still continue counting until the same Stop mechanical and cmos delay was over and the Counter stop counting.

See my illustration of combination of delay on "start mech/cmos delay" and "stop mech/cmos delay " they are same. Therefore Length of time BULLET HIT SWITCH was still the same with COUNTER.

 

[wannabegeek]

I did a test tonight using two pieces of aluminum foil separated by a layer of plastic wrap.

I shot holes through this sandwhich with my .22 Cal air rifle and it did not short until the fourth shot... didn't think it would but I'm typically wrong ...

I was also curious if I could get a resonant series LRC circuit to change it's output by putting holes in the home made capacitor....didn't get that lucky...I think that idea is a
waste of time, but was fun to try...plus I spend more time with my new function generator and learned my way around it better.

Still need to do Steve's analysis and order a quad NOR gate to make SR FF.
Looking forward to having my gate being robust and moving forward.

I might not work on this again until next week since I have some regular life stuff to do...

wbg

 

[CDRIVE]

How did you determine that the foils didn't short until the fourth shot? Did you use a latch? An Ohmmeter is no test at all.

Edit: That suggestion wasn't supposed to short the foils. The metallic projectile was meant to be the momentary shorting element.

 

[wannabegeek]

How did you determine that the foils didn't short until the fourth shot? Did you use a latch? An Ohmmeter is no test at all.

Edit: That suggestion wasn't supposed to short the foils. The metallic projectile was meant to be the momentary shorting element.

hmmm I didn't use a latch...so I guess I didn't measure it right.

Ordering parts right now before I leave town so I can build the two different SR FF gates when I return.

Then I can test your idea properly...

EDIT: I need to be careful to order the right NOR gates. Rleo6965 shows 4001B and 4012B ( AND ) . I am tempted to just order those....are 74HC chips in general OK ? for this....

EDIT2: Here's what I was gonna get: Part no. 1996719
http://www.jameco.com/webapp/wcs/st...&freeText=NOR gate 74HC&search_type=jamecoall

 

[CDRIVE]

hmmm I didn't use a latch...so I guess I didn't measure it right.

I'm not certain how reliable the foil method will be but you're correct when you say that you didn't measure it right.

 

[Rleo6965]

EDIT: I need to be careful to order the right NOR gates. Rleo6965 shows 4001B and 4012B ( AND ) . I am tempted to just order those....are 74HC chips in general OK ? for this....

You can use 74HC series CMOS IC in replace with CDXX series. But take note that they are not pin compatible.
Replace CD4001B with 74HC02
Replace CD4012B dual 4 input NAND 74HC20 or with 74HC10 Triple 3 input NAND gate. We only use 3 input pin in our diagram. Other input pin was tied to +VCC.

Take care of inserting CMOS IC in your circuit. Be sure to measure that there's no voltage remaining in +VCC or + DC Supply. It will damage CMOS IC.