Arduino for pool

[iusedtoberich]

Does anyone have experience with Arduino open source projects? I just spent the last few hours reading/youtubing a lot about it (first time ever for me). Here is a good introductory video by one of its inventors if you've never heard of it: https://www.youtube.com/watch?time_continue=18&v=UoBUXOOdLXY

I want to use it to put a sensor inside of a ball. I'd like to measure things like how the ball rebounds off of the rail for different tables, how the spin changes during the rail rebound, etc. I think I would need velocity right before impact (which includes direction), velocity right after impact, and rotational speed right before and after impact. I think I'd like the ball mounted on a pendulum frame, so that the bed cloth does not influence the cushion measurements. The pendulum frame would also be easy to set up repeatable strikes of the ball into the rail. Although, maybe it could be a bare ball without any frame/pendulum mounting.

I don't know anything about programming, so I'd be starting from scratch from that standpoint. From the mechanical standpoint, my background is strong, and I can build a nice pendulum frame, etc.

I think what I would need is a chip that has a 3 axis accelerometer and a 3 axis gyroscope. But I'm not certain. Mr Jewet or Mr Alciatore would this be correct?

Anyway, has anyone fooled around with it? If so, what have you made?

 

[hang-the-9]

Does anyone have experience with Arduino open source projects? I just spent the last few hours reading/youtubing a lot about it (first time ever for me). Here is a good introductory video by one of its inventors if you've never heard of it: https://www.youtube.com/watch?time_continue=18&v=UoBUXOOdLXY

I want to use it to put a sensor inside of a ball. I'd like to measure things like how the ball rebounds off of the rail for different tables, how the spin changes during the rail rebound, etc. I think I would need velocity right before impact (which includes direction), velocity right after impact, and rotational speed right before and after impact. I think I'd like the ball mounted on a pendulum frame, so that the bed cloth does not influence the cushion measurements. The pendulum frame would also be easy to set up repeatable strikes of the ball into the rail. Although, maybe it could be a bare ball without any frame/pendulum mounting.

I don't know anything about programming, so I'd be starting from scratch from that standpoint. From the mechanical standpoint, my background is strong, and I can build a nice pendulum frame, etc.

I think what I would need is a chip that has a 3 axis accelerometer and a 3 axis gyroscope. But I'm not certain. Mr Jewet or Mr Alciatore would this be correct?

Anyway, has anyone fooled around with it? If so, what have you made?

The issue with this is that putting in the sensor will change how the ball travels vs a standard ball unless you can match the weight and balance perfectly. You would want the sensors to keep data as if it was a normal ball on a normal table, so bed cloth, rails, temps, etc.. should all be used, except the ball has to be balanced and weighed just like a real ball.

 

[iusedtoberich]

The issue with this is that putting in the sensor will change how the ball travels vs a standard ball unless you can match the weight and balance perfectly. You would want the sensors to keep data as if it was a normal ball on a normal table, so bed cloth, rails, temps, etc.. should all be used, except the ball has to be balanced and weighed just like a real ball.

I think that can be accomplished to a decent amount. The sensor is about the size of a dime. The ball can have a through hole through its core, then the sensor can be epoxied into place right in the middle, then filler material (made of the ball) can be used to plug both sides of the ball back to about normal.

 

[hang-the-9]

I think that can be accomplished to a decent amount. The sensor is about the size of a dime. The ball can have a through hole through its core, then the sensor can be epoxied into place right in the middle, then filler material (made of the ball) can be used to plug both sides of the ball back to about normal.

Remember balance, it should be a globe and evenly balanced or you will turn the ball into one of those trick ones that wobble which means results won't really be valid to see how the ball reacts. I would have thought Aramith or someone would have come up the idea already and did it, at least as a test or demo. Even a solid ball like the Cyclop cue ball had issues with rolling properly when the balance was off by a bit, remember people saying how it wobbled and rolled funny?

 

[DaveK]

Never used Arduino but have a little experience with Raspberry Pi / Python (and a degree in EE which included Basic/Fortran/Assembler/Machine language classes). You having no programming experience is a big drawback. Expect a lot of hours (dozens) learning the very basics, then a pile (hundreds ?) more to do anything useful.

Given what you are looking to measure my approach would be visual. Use a video camera and analyze the ball motion programmatically. Of course that too would require fairly sophisticated software and a non-programmer setting out to create this in their spare time is a large mountain to climb ... IMO.

Dave

 

[iusedtoberich]

Never used Arduino but have a little experience with Raspberry Pi / Python (and a degree in EE which included Basic/Fortran/Assembler/Machine language classes). You having no programming experience is a big drawback. Expect a lot of hours (dozens) learning the very basics, then a pile (hundreds ?) more to do anything useful.

Given what you are looking to measure my approach would be visual. Use a video camera and analyze the ball motion programmatically. Of course that too would require fairly sophisticated software and a non-programmer setting out to create this in their spare time is a large mountain to climb ... IMO.

Dave

Ha ha. Yes, I thought of the high speed video approach. But my end result desired is a device I can easily take from table to table. Push a button. Swing it on its pendulum. Then the device will spit out:
-the entrance speed into the rail
-the exit speed out of the rail
-the entrance angle into the rail
-the exit angle out of the rail.
-the entrance spin into the rail
-the exit spin out of the rail

I think a photographic approach would be harder from a software and apparatus device standpoint. I could be wrong.

 

[iusedtoberich]

I do like the vision system from the standpoint that the instrumentation would not touch the ball, as the embedded sensor would. This would eliminate any concerns like Hang-the-9 was stating.

 

[iusedtoberich]

I also thought of hanging the ball from a pendulum, where the pendulum is a stiff but light hollow tube. The pendulum has 3 rotary encoders. One at the end of the tube to measure ball spin. One to measure change in angle (into and out of the rail), and one to measure the pendulum height.

This is something more up my alley from a mechanical standpoint. I worked with EE's to make a device like this to measure the angle of swings that rock infants to sleep. (That's my day job). But, I thought this system would have a much higher friction and influence the path of the ball.

In contrast, if the sensor was inside of the ball, the pendulum can be a simple piece of fishing line. That would be the least friction, and allow the most degrees of freedom. That is why I was thinking an accelerometer/gyroscope combination. But I'm not sure if that would measure what I want (assuming I was an expert programmer...)

 

[DaveK]

I also thought of hanging the ball from a pendulum, where the pendulum is a stiff but light hollow tube. The pendulum has 3 rotary encoders. One at the end of the tube to measure ball spin. One to measure change in angle (into and out of the rail), and one to measure the pendulum height.

This is something more up my alley from a mechanical standpoint. I worked with EE's to make a device like this to measure the angle of swings that rock infants to sleep. (That's my day job). But, I thought this system would have a much higher friction and influence the path of the ball.

In contrast, if the sensor was inside of the ball, the pendulum can be a simple piece of fishing line. That would be the least friction, and allow the most degrees of freedom. That is why I was thinking an accelerometer/gyroscope combination. But I'm not sure if that would measure what I want (assuming I was an expert programmer...)

Rube ? That you man ?

Dave

 

[DaveK]

Ha ha. Yes, I thought of the high speed video approach. But my end result desired is a device I can easily take from table to table. Push a button. Swing it on its pendulum. Then the device will spit out:
-the entrance speed into the rail
-the exit speed out of the rail
-the entrance angle into the rail
-the exit angle out of the rail.
-the entrance spin into the rail
-the exit spin out of the rail

I think a photographic approach would be harder from a software and apparatus device standpoint. I could be wrong.

Here is one way to quantify the "hardness" :

Write up a detailed functional requirements.
Send it to several embedded systems / prototype developers and request quotes.
Sit down before opening responses.

Take their estimates for duration and multiply by 3 or 4 to compensate for needing to learn about sensors and coding.

Good luck.

Dave

 

[iusedtoberich]

Here is one way to quantify the "hardness" :

Write up a detailed functional requirements.
Send it to several embedded systems / prototype developers and request quotes.
Sit down before opening responses.

Take their estimates for duration and multiply by 3 or 4 to compensate for needing to learn about sensors and coding.

Good luck.

Dave

Oh, I know! I'd estimate the cost of one such prototype would be in the 10,000 USD range to have a company do it. But, from a hobbiest standpoint, I believe the electronics could be had for less than 300 using the Arduino stuff. Whether it is accelerometer or rotary encoders. The mechanical stuff I can design and prototype myself. (I'm a mechanical engineer).

 

[jrctherake]

Ha ha. Yes, I thought of the high speed video approach. But my end result desired is a device I can easily take from table to table. Push a button. Swing it on its pendulum. Then the device will spit out:
-the entrance speed into the rail
-the exit speed out of the rail
-the entrance angle into the rail
-the exit angle out of the rail.
-the entrance spin into the rail
-the exit spin out of the rail

I think a photographic approach would be harder from a software and apparatus device standpoint. I could be wrong.

The end goal you seek would be very nice but you not not having a real strong background in programming is a MAJOR hurdle. One in which, in my opinion, will kill the project unless you bring a someone with with strong programming skills on board.

Even then, its gonna be tough but doable given enough time and $$$.

 

[Mosens]

To get you started hardware-wise here's some suggestions...

-MPU-6000 (3-axis Acceleromters/Gyroscope)
-ATSAMD21E17A-MU (Cortex processor to control the sensors, somewhat compatible with Arduino 'Zero', built-in USB)
-ABS07-32.768KHZ-T (32.768kHz Crystal, for the processor)
-MIC5205-3.3YM5-TR (voltage regulator)
-RN4871-V/RM118 (Bluetooth 4.2 radio, optional?)
-MCP73831T-2ATI/OT (LIPO battery charger, optional?, or just expect to keep changing batteries)
-RN4871-V/RM118 (4.2 Bluetooth module, optional?, depending on how you want to get the data out)
-Various capacitors, resistors (0402 SMD is a reasonable size, half dozen or so of each)
-3.7 volt LIPO battery (or whatever will deliver enough voltage/mAH)

The above could likely be crammed into something the size of a stack of about 4 quarters, designing the PCB will be very challenging.

Arduino's programmed mostly in C++, the base language of C is great to know as Arduino, Java (Android) and Swift (iOS) is based from it.

Programming is 90% of the 'secret sauce' to get things to work, don't discount the time/frustration for programming particularly if you're new at it.

I was involved with much of the development of the QMD3 cue orientation/motion sensor (www.cue-md.com, pardon the plug), we have roughly 100 years worth of programming, embedded systems and PCB development experience under our belts so I know of what we speak.

I think this project is do-able, I'd love to see the results.

 

[Dan White]

measure the angle of swings that rock infants to sleep. (That's my day job).

Your day job is rocking infants to sleep? How do you get that gig? Do you have a bunch of them on a leash like a dog walker in their cribs? lol just kidding.

All About Circuits is a good forum where I've gotten lots of help on various electronics projects. You might get lucky and catch the attention of someone who takes interest in what you are doing.

 

[iusedtoberich]

To get you started hardware-wise here's some suggestions...

-MPU-6000 (3-axis Acceleromters/Gyroscope)
-ATSAMD21E17A-MU (Cortex processor to control the sensors, somewhat compatible with Arduino 'Zero', built-in USB)
-ABS07-32.768KHZ-T (32.768kHz Crystal, for the processor)
-MIC5205-3.3YM5-TR (voltage regulator)
-RN4871-V/RM118 (Bluetooth 4.2 radio, optional?)
-MCP73831T-2ATI/OT (LIPO battery charger, optional?, or just expect to keep changing batteries)
-RN4871-V/RM118 (4.2 Bluetooth module, optional?, depending on how you want to get the data out)
-Various capacitors, resistors (0402 SMD is a reasonable size, half dozen or so of each)
-3.7 volt LIPO battery (or whatever will deliver enough voltage/mAH)

The above could likely be crammed into something the size of a stack of about 4 quarters, designing the PCB will be very challenging.

Arduino's programmed mostly in C++, the base language of C is great to know as Arduino, Java (Android) and Swift (iOS) is based from it.

Programming is 90% of the 'secret sauce' to get things to work, don't discount the time/frustration for programming particularly if you're new at it.

I was involved with much of the development of the QMD3 cue orientation/motion sensor (www.cue-md.com, pardon the plug), we have roughly 100 years worth of programming, embedded systems and PCB development experience under our belts so I know of what we speak.

I think this project is do-able, I'd love to see the results.

Yes, I own the QMD version 1

Thank you for the helpful rough component list. My fundamental question that maybe you can answer, is does the MPU-6000 (accelerator plus gyroscope) or similar sensor from another manufacturer, give me what I need to calculate the 6 things I want to: Speed before/after, angle before/after, spin before/after?

So if I was an expert in programming, or if I hired someone who is, could this sensor give me what I'm after? I'm not asking the math or programming specifics. I'm just asking if someone did know that, is it possible?

 

[iusedtoberich]

Your day job is rocking infants to sleep? How do you get that gig? Do you have a bunch of them on a leash like a dog walker in their cribs? lol just kidding.

All About Circuits is a good forum where I've gotten lots of help on various electronics projects. You might get lucky and catch the attention of someone who takes interest in what you are doing.

Ha ha I've designed these swings (along with a team of course) for the past 15 years. Actually I did, as I just got laid off

 

[nataddrho]

Rich,

I might have beaten you to this. I've been working on this exact problem for the past few years, and it is a super tough problem to solve once you get down to the details like moment of inertia restoration, the power consumption of MEMS gyroscopes, a way to charge an embedded battery reliably, low cost, potting (epoxy is the way to go because phenloic resin is poured at 120degC, says the president of Cyclops when I asked him), shock survivability (Dr. Dave calculated for me impact to be around 10,000 Gs of acceleration), and a myriad of other things. I filed a provisional patent for some of my ideas already.

But this problem is so hard anyway, I just want to see someone make a product that is useful and worth the money. Please continue with this effort anyway! I want to have a prototype ready for the SBE if I can get some more things working properly.

Nate

 

[iusedtoberich]

Rich,

I might have beaten you to this. I've been working on this exact problem for the past few years, and it is a super tough problem to solve once you get down to the details like moment of inertia restoration, the power consumption of MEMS gyroscopes, a way to charge an embedded battery reliably, low cost, potting (epoxy is the way to go because phenloic resin is poured at 120degC, says the president of Cyclops when I asked him), shock survivability (Dr. Dave calculated for me impact to be around 10,000 Gs of acceleration), and a myriad of other things. I filed a provisional patent for some of my ideas already.

But this problem is so hard anyway, I just want to see someone make a product that is useful and worth the money. Please continue with this effort anyway! I want to have a prototype ready for the SBE if I can get some more things working properly.

Nate

Yes, I'm certain it would be hard! It sounds like you are doing a ball to be used for actual playing, based on your descriptions, especially of the impact shock values. I hope you do come up with something to market. I'm sure you already bought and took apart the smart golf balls on the market!

The intent of my device is more to measure the differences in tables, and try to quantify it. I don't think there would ever be a consumer or commercial market for it. It would be more to understand what is happening with different brands of tables, and even the same table with different installation techniques.

But I can see from a consumer standpoint, a CB and/or OB that they can buy, might be a profitable product.

 

[iusedtoberich]

PS, to our two resident cue stick accelerometer designers, have either of you thought of making an app that interfaces with Apple Watch? The shooter can put the watch on their shooting hand, and you might have all the sensors you need. So you wouldnt need to ship hardware. Selling an app for that for $5 or $10 might net you much more cash than selling the complete package of hardware plus software for $100 to $200.

Just a thought! Don't shoot the messenger.

 

[Bob Jewett]

... I think I'd like the ball mounted on a pendulum frame, so that the bed cloth does not influence the cushion measurements....

I think that removing the ball from the bed cloth would make the cushion impact measurements nearly worthless. Further, since the impact on the cushion is not at the equator of the ball, the ball will yank on the pendulum and start oscillating vertically.

I think you need to review the measurements that people have already made before you plan to make your own. In particular, ask yourself what have they failed to measure that your approach would allow.