Measuring speed of pool breakshot

[misterT]

The goal is to find out the speed of a cueball in a pool breakshot using electret microphone and measuring the time between the cue hitting the cueball and the impact to the rack. The distance is fairly constant (50") and the sounds are quite similar if the break-cue has a fenolic tip. Time between the two impacts are expected to be 50ms to 100ms.

This would be the general design:
Electret microphone -> Amplifier -> Comparator -> Microcontroller -> Display

I would expect to get a series of small pulses out of the comparator when the cue hits the white ball.. detect the first one and start a timer.. discard all the following pulses for ~40ms. Then wait for the next train of pulses from the impact to the rack and stop the timer. I can handle the digital part easily, but I am rusty with analog design.

Is 5V enough to power a small electret microphone? I would like to power the circuit with 4 AA batteries (4x1.2V).

Is this a good amplifier (below) to use and can I use LM358 instead of TL071?
Are the capacitor/resistor values suitable for this application?
What kind of frequency response does this amplifier have?

What pitfalls I need to avoid and how to get the best accuracy? I would like to get reliable response to the impact sounds, very small (or consistent) delay on the signal, and a simple design.

 

[Sceadwian]

Just a recommendation, use a decent video camera. 30FPS film at a modest 640x480 resolution will capture close to world record breakspeeds. Virtually any cell phone now days can manage this, and the better ones can do HD. Use something like a PS3 eyecam that can capture at 60-100fps and you'll get fine grained results.

 

[misterT]

Just a recommendation, use a decent video camera. 30FPS film at a modest 640x480 resolution will capture close to world record breakspeeds. Virtually any cell phone now days can manage this, and the better ones can do HD. Use something like a PS3 eyecam that can capture at 60-100fps and you'll get fine grained results.

That is a good technique, but I want a small device that gives immediate results without the need to analyze video images or sound waves by hand. There is even a phone app that does measure the speed from the sounds, but it does not work on my phone. I would use the device during my own practice and just for fun during our weekly tournaments.. so it should be easy to use and give the results immediately. The measurements should be comparable, but they do not need to be dead-accurate.

 

[Mr RB]

Sound won't give you the ball speed, because the first sound occurs at the strike speed and strike distance and then the ball position and speed varies after that before it's released from the cue tip. Also you don't know the exact impact point on the other balls, which changes your distance depending which ball they aim for, or where it happens to hit.

If you want this accurate I would use 2 lasers pointed across the table in a place where the cue will not reach, ie well after the strike. That will give actual ball speed.

If you don't want to use break lasers (so you don't need to install stuff in the cushion) you could point the two lasers down a bit and mount them above the cusion, and use reflection as the ball detect not beam-break.

 

[WTP Pepper]

Because of the application, you need to consider the response of the amplifier as well as its bandwidth.
A pool ball noise will not only have higt frequency components, but will have fast attack.
I am concerned over the charge time of C2 in the stability of the amp cct with what could be upto 470K charging it because of the time constant.

You could try a high frequency amp, bias the input of a monotstable slightly off centre and use the amp to fire the monostable. Could draw, but LT Spice isn't working at the moment. The monostable would prevent re-firing due to high frequency crack of the balls on each other.

 

[misterT]

Sound won't give you the ball speed, because the first sound occurs at the strike speed and strike distance and then the ball position and speed varies after that before it's released from the cue tip.

The cue is in contact with the cue-ball only ~1 millisecond. Here is high-speed video of a power break: https://www.billiards.colostate.edu/high_speed_videos/new/HSVA-59.htm

I don't think the ball slows down very much during the 50" distance before it hits the rack.

Also you don't know the exact impact point on the other balls, which changes your distance depending which ball they aim for, or where it happens to hit.

Good players try to make their break shots as identical as possible. Practically all breaks are shot from the headstring and aimed to the closest ball in the rack (the ball on the footspot). The goal is to hit that ball full on the face. 9-ball and 10-ball rules even require the player to hit the 1-ball on the footspot. 90% of the time the distance is either ~50 or ~56 inches. I can have a potentiometer to select the distance if I like, but the error is within my tolerance.

Because of the application, you need to consider the response of the amplifier as well as its bandwidth.
A pool ball noise will not only have higt frequency components, but will have fast attack.
I am concerned over the charge time of C2 in the stability of the amp cct with what could be upto 470K charging it because of the time constant.

You could try a high frequency amp, bias the input of a monotstable slightly off centre and use the amp to fire the monostable. Could draw, but LT Spice isn't working at the moment. The monostable would prevent re-firing due to high frequency crack of the balls on each other.

Thanks, I think I know what you mean. I'll build a prototype and see how it performs.. I'll be back with results and possibly more specific questions if there are problems.

 

[Mr RB]

Good points MisterT. I still see it as an imperfect system. People may put the ball on a different place on the start line, and I know some people prefer to hit one of the second balls not the lead ball, especially more tactical players. Those two factors might give you 4" difference or more in your 50" expected distance.

As for the cue being in contact with the ball 1mS there's still some unknowns in the actual releasepoint of the ball from the tip, and at the speed of sound 1mS is a foot of distance and you are measuring a 50" distance.

It sounds like you want a sound based system, which should work ok but would just not be as accurate as a dual laser system. The dual laser could still be made small and simple like a cellphone size, and sat on the edge of the table for the break. If the lasers are 4" apart and you use a typical micro timer resolution of 10Mhz I think you would get a much more accurate reading of ball speed.

 

[misterT]

It sounds like you want a sound based system, which should work ok but would just not be as accurate as a dual laser system.

If a player breaks from the middle of the headstring (start line), the distance is ~50". From the side of the headstring the distance is ~55" (because the ball travels diagonally). That is 5% error if I assume the distance is ~52.5". Laser system will have exactly the same error because of the geometry (ball traveling straight or diagonally). The ball is almost never hit from far behind the headstring.

I know some people prefer to hit one of the second balls not the lead ball, especially more tactical players.

That is very rare and can happen only in the game of 8-ball. We play mostly 9 and 10-ball.

As for the cue being in contact with the ball 1mS there's still some unknowns in the actual releasepoint of the ball from the tip, and at the speed of sound 1mS is a foot of distance and you are measuring a 50" distance.

No, there are no unknowns And I don't care how far the sound has traveled during the 1ms. It has no effect to the measurement of cue-ball speed. I only care about time difference between two sounds.

Anyway.. I like brainstorming/depating these kind of things, but at the end I want a device that fits in my pocket, is easy to use and can be used from a distance without the need to install anything to the table or the need to disturb the players. If I can get 5% accuracy and 2% precision out of this I am quite happy.

I (almost) have a working circuit at the moment. It is "clapper switch circuit" with a basic microphone amplifier and a 555 wired as a monostable with biased input (as WTP Pepper suggested). Still need to write the code etc. The clapper seems to react well to clapping I have good controll of the sensitivity.. no idea about any possible delays in the signal.

 

[Nigel Goodwin]

Assuming you have access to a digital storage scope it would be childs play to build the simple preamp and record the output from the preamp - you could even do it with a PC sound card and Audacity software. From that you could experiment with various filters to try and produce reliable logic outputs to measure using a micro.

 

[misterT]

Assuming you have access to a digital storage scope it would be childs play to build the simple preamp and record the output from the preamp - you could even do it with a PC sound card and Audacity software. From that you could experiment with various filters to try and produce reliable logic outputs to measure using a micro.

Do you think I could record a video/audio of a break-shot and analyze the soundtrack? I believe the soundtrack could be used as an input in LTspice circuit. Or, should the sample come from my preamp circuit (and the electret mic I'm using)?

 

[jpanhalt]

You may find Sound Ruler useful. It is freeware: http://soundruler.sourceforge.net/main/

I haven't used it in a couple of years, so I can't speak to any changes that have been made in it. I used it to analyze popping corn and found it quite intuitive and helpful.

John

 

[Mr RB]

...
I (almost) have a working circuit at the moment. It is "clapper switch circuit" with a basic microphone amplifier and a 555 wired as a monostable with biased input (as WTP Pepper suggested). Still need to write the code etc. The clapper seems to react well to clapping I have good controll of the sensitivity.. no idea about any possible delays in the signal.

A clap detect circuit should work ok, just beware of any recovery time if you use capacitor coupling or capacitive filters etc, you want to make sure all the circuit voltages recover quickly after the first firing and are exactly the same when the second firing occurs.

Have you done much with sonic ranging? You do know if you have the sensor not an equal distance from the two sound sources you will get an additional error at 1mS error per foot of distance difference?