looking for miniature circuit to detect orientation in 3D

[en2oh]

Hi guys,
I've poked around a little to see if there is anything right on point for this question.

I'd like to be able to the orientation of a die (6 sided cube) in space. I've thought about using opposing tilt switches in the x, y and z axis'. So, if the die is turned over, one of the tilt switches will open and the opposing switch will close.

Is there a more reliable way of doing this (and doing it within the constraints of a 3/4" cube with other circuits required plus battery?

Thanks guys,
doug

 

[crutschow]

A 3-axis accelerometer chip (to sense the direction of gravity) such as this should work.

 

[Mikebits]

A 3-axis accelerometer chip (to sense the direction of gravity) such as this should work.

Amazing how small they make those now. When I was in the Navy those things were HUGE!

 

[en2oh]

Thanks Crutschow,
That seems to work for detecting tilt, but absolute special orientation, I'm not so sure about without on pretty sofisticated microprocessing power.
Doug

 

[dr pepper]

You can get analogue versions that you could use with a comparator if your not into micro's.

 

[alec_t]

Post deleted

 

[jpanhalt]

Thanks Crutschow,
That seems to work for detecting tilt, but absolute special orientation, I'm not so sure about without on pretty sofisticated microprocessing power.
Doug

Not quite sure what you mean by that. Can you illustrate with an example?

There is an application note on using a 3-axis accelerometer for pitch, roll, and yaw in an aircraft. Although the accelerometer will fit within the confines of the cube you mention, how are you going to read and/or tramsmit the information? Are you planning to use light, RF, something else?

John

 

[crutschow]

Thanks Crutschow,
That seems to work for detecting tilt, but absolute special orientation, I'm not so sure about without on pretty sofisticated microprocessing power.
Doug

The output of the device I referenced is analog with 1g giving a 300mV change in output voltage (actually this device is recommended for new design). You could determine the cube orientation by using comparators and some simple logic on the output signal.

Edit: The output changes ±300mV around its output bias point for ±1g direction change, thus you can detect the orientation of any axis as to being up or down or horizontal (no output). If you know that, then you can determine the cube orientation.

 

[en2oh]

Spelling error. Should have been spacial not special ie to determine if the sensor, and so the die it is mounted within, what side is "up", so to speak.
Nitrous

 

[en2oh]

RF would be the ideal.

 

[alec_t]

I've thought about using opposing tilt switches in the x, y and z axis'.

So all you really want to know is 'which side is up'? That should be a lot simpler than full 3D orientation (which is what some of the suggestions above are aimed at).
If the tilt switches are a bit bulky you could instead have copper patches covering all but a ~1mm perimeter region of each inside face and use a small ball bearing to short adjacent patches.

 

[en2oh]

Since there are 6 faces to any cube, I need a six pointed object, with a point at each face of the cube.
An interesting structure to build, especially with the need for a transmitter inside
I think an RF24L01 multiceiver might be what I need. I believe it has 6 receiver channels in that chip.

Finding a tiny module that would detect (tilt switches seemed like hte way to go) this reliabily, but not take up too much room.
I've got a source of 1cm long ball-tilt switches that might work but was hoping for smaller.

This is the link to the module. There are two outputs, which presumably accounts for the for four positions sensed.

https://www.electro-tech-online.com/custompdfs/2012/09/28036-4DirectionalTiltSensor-v10.pdf

Thanks for the input thus far

nitrous

 

[en2oh]

I've found a four direction Tilt Switch that is SMD (so nice and small)
Just add a couple of single direction tilt switches in opposition and that should do it.
I also thought about making a conductive sphere with a conductive ball inside.... like was suggested but when I found the SMD 4 direction tilt, it seems like I'm reinventing the wheel.
Thanks guys
nitrous

 

[dr pepper]

In the 80's there was a motion sensor that had a ball inside and 4 piezos round it, when the sensor was tilted the ball contacted the particular piezos and gave an output, you had to position the sensor back upright to reset it.
I have one somewhere never found a use for it, I dont know if it was designed to be a position sensor or a vibration sensor, the local electronics shop had a few surplus ones cheap at the time.

 

[Mikebits]

In the 80's there was a motion sensor that had a ball inside and 4 piezos round it, when the sensor was tilted the ball contacted the particular piezos and gave an output, you had to position the sensor back upright to reset it.
I have one somewhere never found a use for it, I dont know if it was designed to be a position sensor or a vibration sensor, the local electronics shop had a few surplus ones cheap at the time.

I think they had those in old car alarms.

 

[dougy83]

I think an RF24L01 multiceiver might be what I need. I believe it has 6 receiver channels in that chip.

Note that you'll need a microcontroller to use that transceiver. The interface is SPI.

 

[Mr RB]

I've found a four direction Tilt Switch that is SMD (so nice and small)
Just add a couple of single direction tilt switches in opposition and that should do it.
I also thought about making a conductive sphere with a conductive ball inside.... like was suggested but when I found the SMD 4 direction tilt, it seems like I'm reinventing the wheel.
Thanks guys
nitrous

They are very common now but mainly inside products (like cameras and 'phones) and are SMD sized parts. They are not easy to find at a hobby level but if you know the part number you can probably get one from Farnell or Digikey.

I played with a single axis one here;
https://www.romanblack.com/tilt.htm

**broken link removed**

 

[dr pepper]

A proper enthusiast has a protractor with holes drilled in it roman.

 

[Mr RB]

Of course it has holes drilled in it! How else would I attach it to something when it requires higher precision than blu-tack allows or high RPM?

All jokes aside I grabbed a few of those 4" round 360' protractors from the newsagent for I think about $1.50 each. Very handy, even if you need to cut one up to do a job.

 

[en2oh]

Since there are 6 faces to any cube, I need a six pointed object, with a point at each face of the cube.
An interesting structure to build, especially with the need for a transmitter inside
I think an RF24L01 multiceiver might be what I need. I believe it has 6 receiver channels in that chip.

Finding a tiny module that would detect (tilt switches seemed like hte way to go) this reliabily, but not take up too much room.
I've got a source of 1cm long ball-tilt switches that might work but was hoping for smaller.

Thanks for the input thus far

nitrous