Go Down

Topic: Accelerometer for sensing how fast a poi is spinning? (Read 6179 times)previous topic - next topic

jfeldstein

I'm in the drafting stage of making a set of poi that are aware how fast they're being spun with the intention of doing something cool with an array of LED's, currently looking for an accelerometer sensor to put in the end of the poi. My hope is that this would give me a reading of the acceleration the poi are undergoing and I could use that to calculate how fast they're moving and run different sequences in the LEDs to make it look cool.

What would be a good sensor/accelerometer for getting an estimate of how fast the poi are spinning? I've found this accelerometer, but I'm concerned that this and most components like it, would be more appropriate for orientation sensing and not such a good fit for acceleration sensing, like I would need.

Thanks guys!

#1
Jul 01, 2011, 08:09 pm
Seems you want to measure centripital force
http://regentsprep.org/regents/physics/phys06/bcentrif/default.htm

How about a force sensor at the end of the poi, such that faster speed would compress the force sensor more, while no spinning (hanging straight down) would only see 1G of force.
Designing & building electrical circuits for over 25 years.  Screw Shield for Mega/Due/Uno,  Bobuino with ATMega1284P, & other '328P & '1284P creations & offerings at  my website.

jfeldstein

#2
Jul 01, 2011, 08:44 pm
Exactly the component I'm looking for, for exactly what I'm trying to measure.

Can you recommend one? I'm not sure this one on parallax is what I'm looking for..

#3
Jul 01, 2011, 09:11 pm
sparkfun carries them, in different sizes

http://www.sparkfun.com/products/9673

You'd have to do some experimenting and see what range you needed.
Or some calculating, I think the link I sent earlier can give you expected range based on how heavy your poi, and some estimates by you on how fast it moves, etc.
Designing & building electrical circuits for over 25 years.  Screw Shield for Mega/Due/Uno,  Bobuino with ATMega1284P, & other '328P & '1284P creations & offerings at  my website.

MarkT

#4
Jul 02, 2011, 03:48 am
Many MEMS accelerometers have a +/-2g scale and a +/-8g scale - the latter will be needed at least (and it probably will overload), since even a pendulum swinging +/-90 degrees pulls 3g at the bottom of the swing.
[ I will NOT respond to personal messages, I WILL delete them, use the forum please ]

johnwasser

#5
Jul 02, 2011, 04:02 am
Maybe instead of detecting the acceleration shifts you could have an infrared beacon on the ground and an IR sensor in the poi to detect when it was passing near the beacon.  A pulse per revolution will tell you how fast they are spinning.  Something like a BlinkM (ATtiny plus RGB LED) can probably do the trick.  If you use an IR beacon you might even use it to send a pattern allowing you to switch light programs.
Send Bitcoin tips to: 1G2qoGwMRXx8az71DVP1E81jShxtbSh5Hp

Magician

#6
Jul 02, 2011, 05:07 am
The really "ingenious" solution I've seen in the book:
"tinyAVR Microcontroller Projects for the Evil Geniusâ„˘".
Author refers to Elektor magazine, December 2008("LED Top with Special Effects"), (magazine probably would be easier to find in the library)
Motion sensor include two coils and OPA's, that sense Earth magnetic field. It allows to get speed, direction of rotation, plus synchronize leds for POV effects. Last feature is impossible to obtain with accelerometers or force sensors.
I'm not sure why hall effect sensors were not used in design?

jfeldstein

#7
Jul 02, 2011, 06:27 pm
Wow! So many ways to approach the same problem.

IR beacon is tough, because I won't always be in the same place, and may often move while playing.

I'm going to read up on how the magnetic option works. I think it'll wind up being too complicated for what I'm doing, but sounds really cool to know.

What do you guys think about measuring tension in the rope? Instead of measuring gravity at the end of the poi, measuring the poi's pull on the handle in your hand..?

MarkT

#8
Jul 03, 2011, 12:53 am

The really "ingenious" solution I've seen in the book:
"tinyAVR Microcontroller Projects for the Evil Geniusâ„˘".
Author refers to Elektor magazine, December 2008("LED Top with Special Effects"), (magazine probably would be easier to find in the library)
Motion sensor include two coils and OPA's, that sense Earth magnetic field. It allows to get speed, direction of rotation, plus synchronize leds for POV effects. Last feature is impossible to obtain with accelerometers or force sensors.
I'm not sure why hall effect sensors were not used in design?

Accelerometer wlll give synchronization, the centripetal force has a gravitational component that's a rough sinusoid(*) added onto a mean force - so long as its spinning in a vertical plane (if not it will be hard to see the PoV pattern anyway).  High speed spinning will possibly drown out the gravitational signal, but smooth low speed circles should show it up nicely.  Implementing a software PLL to pick out the phase will be an interesting problem too

(*) the maths isn't trivial because the angular velocity isn't constant due to gravity.
[ I will NOT respond to personal messages, I WILL delete them, use the forum please ]

johnwasser

#9
Jul 03, 2011, 01:27 am

Author refers to Elektor magazine, December 2008 ("LED Top with Special Effects").

http://www.elektor.com/magazines/2008/december/led-top-with-special-effects.739805.lynkx?tab=2

To get the article in PDF form you have to spend "10 Elektor Credits", whatever that means.
Send Bitcoin tips to: 1G2qoGwMRXx8az71DVP1E81jShxtbSh5Hp

Magician

#10
Jul 03, 2011, 03:20 am
Quote
Accelerometer will give synchronization,

Agree, in vertical plane.  I was thinking to give more dimension, step up from 1D (gravitational force) to 2D or even 3D space. Project looks really interesting, so I check on magnetic sensors sensitivity and find out there are a lot of them able to sense 30-60 uTesla Earth magnetic field.
Some bases on hall effect, other magneto resistive, most popular application - compass.
And they fast enough to provide at least 10 Hz reading
Sparkfun offers a 3D board: http://www.sparkfun.com/products/9371
It'd cost more than OPA and coils, but not much compare with accelerometers

liuzengqiang

#11
Jul 03, 2011, 04:02 pm

Many MEMS accelerometers have a +/-2g scale and a +/-8g scale - the latter will be needed at least (and it probably will overload), since even a pendulum swinging +/-90 degrees pulls 3g at the bottom of the swing.

happy to see some physics calculation here
as a matter of fact, the acceleration is only 2g but since accelerometer only measures force, it reports 3g. net force on bob is 3mg up and 1mg down so 2mg.

I use one adxl345, with accuracy of 4 milli-g per LSB. It is only 30 bucks, including a level shifter you need to buy for 5v arduinos.

What is a poi?

jfeldstein

#12
Jul 03, 2011, 05:35 pm
A poi is a flow toy, a ball on a string that you play with by twirling around you. Most common types are just a weight on the end of a string/rope/chain, but they come in lots of different shapes. Mine is two on either end of a longer wrope, called a "meteor". Oh, yea. Sometimes you light them on fire, which looks really cool when you're good.

If that megnetomoter can do all of what you claim it can, it'd be reallly powerful. =^] Going to warm up to arduino with some simpler projects (remote control LED lamps, and motorized blinds on a timer) before I drop \$100 for the pair.

MarkT

#13
Jul 05, 2011, 03:08 am
Quote
as a matter of fact, the acceleration is only 2g but since accelerometer only measures force, it reports 3g

That depends whether you are working in an inertial reference frame....  3D accelerometers are neat precisely because they understand F=ma and general relativity!  They only read zero in free-fall.
[ I will NOT respond to personal messages, I WILL delete them, use the forum please ]

liuzengqiang

#14
Jul 05, 2011, 06:03 am

Quote
as a matter of fact, the acceleration is only 2g but since accelerometer only measures force, it reports 3g

That depends whether you are working in an inertial reference frame....  3D accelerometers are neat precisely because they understand F=ma and general relativity!  They only read zero in free-fall.

Which part depends on inertial frame? 2g acceleration or 3mg force measured by the accelerometer?
Why is general relativity mentioned here?