Chronograph for bullets

[CDRIVE]

An RS FlipFlop has no clock input. A JK FlipFlop does. As I see it, you want to measure the elapsed time between contact openings which an RS or JK will do. One of my freq counters has an elapsed time function. There's a jack on the rear that when logic low starts the counter. When the same input is made high the counter stops. The elapsed time is then displayed. A FlipFlop is a perfect interface. Basically, you want to do the same.

You want to use the FlipFlop as a gate a fast (pulse train) clock. You want to measure how many clock pulses get through per each gate period and convert it to velocity.

 

[wannabegeek]

I don't think I understand how the RS FF works....I'm not reading the table correctly...

Could you spell it out a little more ...

Using a SR NOR FF
Detector 1 ---> connected to S
Detector 2 ----> S also ?

Then for output I use Q only ?

So then the sequence would be:
Reset occurs manually, by setting R=1 and S=0 and then R = 0 again
d1 and d2 are fresh and at 0 0
d1 throws open ( and stays open ), output Q goes HI
d2 throws open ( and stays open). output Q goes LO

This SR FF is connected in series with AND gate and clock source, thus
gating the counter....

 

[(*steve*)]

The problem is that you're "spreading my project questions around different threads".

I remember this project. In another thread I'm sure detailed how you would use an SR flip flop.

The comment about not needing to debounce is correct.

Find your original thread and see if I gave useful advice. If this were the same thread I'd look myself.

 

[CDRIVE]

Steve, I found it.

https://www.electronicspoint.com/use-make-counter-decoder-1mhz-events-t249053p2.html#post1474610

Yeah, scattering this all over the forum is really counterproductive. I was going to draw a schematic. Not sure about that any longer.

Chris

 

[(*steve*)]

I note that I suggested an XOR gate could be used to indicate when the sensors were in different states.

Was this the project to time bullets?

I can't see how the sensors would bounce.

Even if it's being simulated with switches, I would just ignore any bounces. Run the timer while the signal is in the appropriate state.

 

[wannabegeek]

I'll figure out a way to gather all my threads back into the original.

I did re-read Steve's original post but was confused by my experimental results.
I'll try it a few more times to make sure it doesn't work as is.

There's also a problem with communicating b/c this very new to me and people often assume that I understand some standard connections or something else. Inaddition I have referenced where I was getting my information and I have consistantly used that language. So calling a flip-flop the same as a latch and saying that a latch uses a clock but not a flip flop is becoming insane...

The idea that a tearing trace could bounce seems just as likely as a momentary button bouncing since bounce/noise can and does occur in the micro second range.

I did give a URL to a cool article about bounce where I referenced some ideas, page 4,

"The two from the el cheapo game joystick were nothing more than gold contacts plated onto a PCB; a rubber cover, when depressed, dropped some sort of conductive elastomer onto the board. Interestingly, the analog result was a slow ramp from zero to five volts, with no noise, wiping or other uncertainty. Not a trace of bounce. And yet… the logic channel showed a msec or so of wild oscillations! What’s going on?
With TTL logic, signals in the range of 0.8 to 2.0 volts are illegal. Anything goes, and
everything did. Tie this seemingly bounce-free input to your CPU and prepare to deal
with tons of oscillation – virtual bounces."

I do appreciate your time and consideration,
wbg

 

[(*steve*)]

I'll figure out a way to gather all my threads back into the original.

send me links to all of them and I'll merge them together.

I did re-read Steve's original post but was confused by my experimental results.
I'll try it a few more times to make sure it doesn't work as is.

sure. If the experimental results seem weird, tell us. We may have some insight.

There's also a problem with communicating b/c this very new to me and people often assume that I understand some standard connections or something else. Inaddition I have referenced where I was getting my information and I have consistantly used that language. So calling a flip-flop the same as a latch and saying that a latch uses a clock but not a flip flop is becoming insane...

Let me give you a car analogy...

So calling a vehicle a truck and saying that a truck uses diesel but not all vehicles do is becoming insane...

Flip-flop is a term that covers a whole range of devices that share certain behaviour.

An SR-flip-flop is a type of flip-flop. (Other types are JK, D, and T). Flip-flops can be used to perform many functions, one of those is to act as a latch. All of the types of flip-flop (other than the T flip-flop) can be used to latch a signal. Some flip-flops (but NOT the SR-flip-flop) have a clock input. You DON'T want one with a clock input.

So, a flip-flop is a latch, and some latches use a clock. Not all latches (in fact very few) are SR-flip-flops. Also you can make a latch that isn't a flip-flop -- but now we're getting weird.

The idea that a tearing trace could bounce seems just as likely as a momentary button bouncing since bounce/noise can and does occur in the micro second range.

But the question is, is it significant? And if you use the flip-flop you can pretty safely ignore all but the shortest (under a couple of ns) bounces.

I did give a URL to a cool article about bounce where I referenced some ideas, page 4,

I'll go take a look.

"The two from the el cheapo game joystick were nothing more than gold contacts plated onto a PCB; a rubber cover, when depressed, dropped some sort of conductive elastomer onto the board. Interestingly, the analog result was a slow ramp from zero to five volts, with no noise, wiping or other uncertainty. Not a trace of bounce. And yet… the logic channel showed a msec or so of wild oscillations! What’s going on?

the slow transition through the forbidden zone between 0V (or close to it) and 5V (or close to it) where the output of logic is not defined.

With TTL logic, signals in the range of 0.8 to 2.0 volts are illegal. Anything goes, and everything did. Tie this seemingly bounce-free input to your CPU and prepare to deal with tons of oscillation – virtual bounces."

There you go. Just what I said. You need to control that transition. If you have slow signals you can use a Schmitt trigger, if you have multiple transitions (where there should be just one) then you can use a latch (but the right sort! which one do you think? SR is the answer).

Clocked flip-flops (and latches) are used when you want to synchronise transition times with a clock signal, or to sample the input at a particular time. In your case you want to do neither. The inputs are completely asynchronous and you don't want to delay them.

 

[(*steve*)]

Did you mean the pdf you linked to in this post?

I've seen that before. I may have linked to it for someone else.

You something the SR debouncer on page 11. However that one may not work for you without some modification (because your detector is essentially a SPST switch. At the very least you'll require a momentary switch to reset the trigger.

I'll try to draw something later. Maybe someone else will beat me to it.

 

[wannabegeek]

Here is how I understand the SR flip flop to work in my case.

To clarify, my case is asynchronous detection where both dectors start with 0 V and are blown open to 5 V and stay until manualy reset.

R = LO for an experiment run, which I didn't notate in the schematic.

This is the truth table I've been thinking of from wikipedia.http://en.wikipedia.org/wiki/SR_Flip_Flop#Set-Reset_flip-flops_.28SR_flip-flops.29

SR Flip Flop operation
Characteristic table Excitation table
S R Qnext Action Q Qnext S R
0 0 Q hold state 0 0 0 x
0 1 0 reset 0 1 1 0
1 0 1 set 1 0 0 1
1 1 X not allowed 1 1 X 0

cheers,
wbg

 

[CDRIVE]

There are a number of things wrong with your circuit.

(1) If you're using TTL the value of those resistors on the input of the SRFF are waaaay too high. I wouldn't use values that high even if it was CMOS.

(2) Detector1 represents the first sensor that the projectile strikes. This gets connected to the Set (S) input of the SRFF.

(3) Detector2 represents the second sensor that the projectile strikes. This gets connected to the Reset (R) input of the SRFF. I explained this in an earlier post.

(4) I see no purpose for your clock input to the AND gate as you show it. It would only make sense if the AND output was connected to an 'Events Counter', not a Frequency Counter. A freq counter requires a steady state signal, not short frequency bursts,

(5) If you're freq counter has a Stop Watch function, also called a 'Timer' function, you can pick the signal directly off of the SRFF output. The signal from the SRFF gates the timer on and off.

Post the model counter you have and some pix if you can,

BTW, a scope at the output of the SRFF can be used too!

 

[wannabegeek]

There are a number of things wrong with your circuit.

(1) If you're using TTL the value of those resistors on the input of the SRFF are waaaay too high. I wouldn't use values that high even if it was CMOS.

yeah I just kinda guessed since I was at 'work' , I should have just left them as R.
I think the 74LS wants 16uA or something as a min current. A tricky thing here, is that the target trace, hasn't been worked out yet. If I end up using a spray paint type solution, carbon or copper, then the target trace might have a fairly high resistance....then the voltage divider could go above 0.8 V. That's for the near future.

(2) Detector1 represents the first sensor that the projectile strikes. This gets connected to the Set (S) input of the SRFF.

(3) Detector2 represents the second sensor that the projectile strikes. This gets connected to the Reset (R) input of the SRFF. I explained this in an earlier post.

Yeah this is not explained well to me, b/c that doesn't make sense with the truth table as I see it... 1 1 is a restricted input....my detectors stay on they don't pulse...I'll always have a 1 1 on second detect...never a 0 1 unless there's a clever use of Q bar that I'm not seeing.

(4) I see no purpose for your clock input to the AND gate as you show it. It would only make sense if the AND output was connected to an 'Events Counter', not a Frequency Counter. A freq counter requires a steady state signal, not short frequency bursts,

Yeah I should have labeled it to 'Counter'...but you're splitting hairs here with terminology,
This circuit is a period timer...it counts clock signals between events and returns the period.

(5) If you're freq counter has a Stop Watch function, also called a 'Timer' function, you can pick the signal directly off of the SRFF output. The signal from the SRFF gates the timer on and off.

I'm not using a universal bench counter as I've written before....This is for field use, in-situ, I am building everything. The gating is done by letting the clock run and connecting the signal to the Counter chip clock input.

Post the model counter you have and some pix if you can,

BTW, a scope at the output of the SRFF can be used too!

My scope is 1972 Techtronix N210 2Mhz analog POS that I found. It takes considerable patience to use...I'm considering buying a Rigol 55 Mhz scope but other life expenses have gotten in the way. But that's not really a great scope for seeing 10 MHz.

wbg

 

[(*steve*)]

Thanks for the reminder

The following circuits use CMOS. They won't work if you use TTL.

Here is a debouncer.

Use a 40106 or similar. Pressing the switch will make the output go high, release and it goes low. You shouldn't have problems with bounce.

Here is a simple SR flip flop. The pull-down resistors should be in the order of 100k.

I have shown a switch to reset it, note that it does not need to be debounced.

Following is a way to wire up your sensors:

The sensor holds the left side of the capacitor low. When it breaks, the output sees a brief pulse (a rising edge then a decay). This can be used to first set, the reset the SR flip-flop.

You can replace your sensor with a normally closed switch and press it to simulate the sensor breaking.

If you connect all of these, then (after pressing the reset button) the output will be low until the first sensor is broken. It will then go high and remain high until the second sensor goes low (where it will remain until you replace the sensors.

 

[CDRIVE]

Perhaps it was stated earlier and I missed it but this is the first I'm reading that the target panels are kaput once the bullet has impacted them. I'm exhausted just thinking about trucking back and forth down & up range to replace them.

What happened to the photoelectric panels that was discussed at the beginning?

Also, did you understand what I described if your counter has a Timer (Stop Watch) mode? Edit: See next post

****************************************************************************
Steve, good edge triggering fix for the FF.

Chris

 

[CDRIVE]

I would have edited my last post with this question but sometimes the edits get missed .

Regarding your counter. Since you don't have a multifunction counter and your going to build it, why not build it as a stop watch to begin with?

Chris

 

[wannabegeek]

I'm super grateful for your drawings...

The following circuits use CMOS. They won't work if you use TTL.

I tried the same debounce topology with TTL....I guess that's why it didn't work....I'll have to think about why that is ( too little current to operate the transistors or something ).

About the sensor adaption with 10nF caps to 100k resistors...interesting idea...
I like the rising edge..but the decay to bring S = 0 , ready for next event troubles me.
Recall that I'm planning on events on the order of 16us. RC = 1 s with your values, correct ?

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

Can I use 10 pF Caps and 1 K resistor ? Assuming very low ripple bench power for testing and battery operated in the field.

I'm exhausted just thinking about trucking back and forth down & up range to replace them.

The idea is to put 4 target-traces on a target and use a register to hold the four times, then collect them all when done. That's after I get the basics working. If people really find this useful, a micro controller version would be in order.

What happened to the photoelectric panels that was discussed at the beginning?

I don't like photoelectric gates for various reasons. For one, light is emitted in a solid angle, and it's hard to know the solid angle for any two LEDS/detectors. Speed accuracy will be linearly dependent on distance uncertainty, but the distance uncertainty in a solid angle will be cos(delta{phi}), where theta is the reported emission angle for given current and phi = theta - 90 deg. Delta phi could be a real PITA to keep small, but how would I even know ? That's just ONE, of many uncontrollable variables with LED's and detectors. Not to mention the shear amount required, the power and the OP AMPs needed for a clean detection out of doors.

The existing chronographs appear pretty hinky IMO. I've only read about them though, and have not put them through any real tests.

A real advantage to this mechanical break design, is that one could setup three of them and get very useful energy calculations involving barrier materials.

Also, did you understand what I described if your counter has a Timer (Stop Watch) mode?

Yeah totally get it...I have an old universal counter...but I'm not using it anymore b/c I really need this to be at the range...and my counter isn't accurate or reliable enough for this project.

This project is simple and yet challenging b/c the events are fast...! 1us or less accuracy is not BS. Most embedded systems blogs talk about 100 ms precision....NI has a PC data collection board good up to 100 MHz they charge $500 for.

I'm so anxious to get the Rigol 55 Mhz scope...I could take some experimental data with my .22 cal air rifle. I'd like to know exactly what happens when the trace is broken. But, I don't really have a working target-trace yet ....just been using foil strips so far.

I'd love to have a copper trace decal made, that just sticks on the front and back of a cardboard mounted range target. Upping my timer crystal to 20 MHz would give me accuracy enough to use this short distance of about 0.25 cm.

 

[CDRIVE]

Upping my timer crystal to 20 MHz would give me accuracy enough to use this short distance of about 0.25 cm.

Maybe I'm misinterpreting this. The panels (1 & 2) aren't spaced 1 Ft apart for a direct conversion of ft/sec?

Re Debounce: I think Steve included that merely for reference. As stated, the SRFF doesn't need it.

 

[wannabegeek]

Maybe I'm misinterpreting this. The panels (1 & 2) aren't spaced 1 Ft apart for a direct conversion of ft/sec?

Yeah I 'm not sure why people like to use this as a standard, it's crude. People use this but it is only true near the muzzle and nothing happens there...

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.

I want the velocity of a bullet in the direction of the target, at impact, thus the speed:

v = dr/dt which is a differential ( ft/s is an approximation and not good enough for energy calculation). How does muzzle velocity help long range shooting...? It doesn't tell us anything about the performance of the cartridge down range, where it counts.

Air resistance is a major force and is likely to a linear or quadratic function of velocity, bullet shape, and other factors that we typically ignore since they are out of our control.

 

[CDRIVE]

I don't see the relationship of 1ft panel spacing to muzzle velocity vs down range velocity. The closer you make those panels the greater the demands on the electronics, mechanics and the greater the chance of errors.

 

[Rleo6965]

Try this simple circuit. Draws only 2ma suitable for battery operation. There's no capacitor circuit might cause inaccuracy of timing.

It's up to you to design mechanism for the start stop switch. If circuit favorable to you . I can also think of mechanism for the start stop if you want too. I'll try my best in my design that bullet will not hit the switch.

What the distance between the 2 sensor switch?

 

[(*steve*)]

Yeah I 'm not sure why people like to use this as a standard, it's crude. People use this but it is only true near the muzzle and nothing happens there...

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.

Having the sensors 10 ft apart will tell you the average velocity in that 10 feet. If you think something interesting is happening in that 10 feet, move your sensors up another 10 feet and measure there.

You will almost certainly get a very good model of the speed if you assume the decrease in speed is linear over the distance between the sensors, and therefore that the average speed was the actual speed when the bullet was mid way between the sensors. The exact position where the bullet had that velocity is probably a little closer to the weapon, but if the sensor hadn't slowed the bullet it may have been slightly further away. Unless you can quantify the effect on velocity of the sensor you're never going to know what your error is anyway.

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.

The closer the sensors, the greater the effect of positional error. If they move back and forth by 0.5mm, you will get a +/- 5% error in velocity due to uncertainty of the relative positions of the sensors. If they are 1m apart, that error is +/- 0.05%. A similar issue may arise if there is some deformation of the sensor before it breaks.