How to move Mount Fuji (and the rest)

[Yaytay]

Folks,

I want to take an illuminated globe and rig it with a pi so that its position (relative to the tilt) matches the real Earth.
This is entirely pointless, but I think it would be cool.

The globe will also have a clock and will glow when the kids are allowed to get up.
Which is not pointless, but is less cool.

My question is, how should I power the motion of the globe in such a way that I know where it is, given the fundamental requirement that the kids will turn it manually no matter how much I tell them not to?
My thinking is that if the globe is manually turned its motion should pick up again when it is in the right position.

My thoughts so far are:

• Power it with a stepper motor, but how will that react to being forced?
• Power it with a standard geared DC motor and have some kind of sensor that tells me where it currently is.
• I can't fix a magnetometer to the globe because the cables would twist, but could I fix a magnetometer to the core and a magnet to the globe?
• If the power is a standard geared DC motor the drive should be a rubber wheel against a ring (slipper clutch) - but I can't do that if the motor is the position sensor.

Have I missed some simple component that would make any of this simpler?
I write software, I know very little about electronics.

Thanks

Jim

[mthomason]

I'm thinking you may want to look into rotary encoders.

There's a few different types, including conductive ones that make physical contact as something turns, to optical which basically just use a wheel with cutouts and an optical sensor to count as the slits move past. You may or may not find yourself able to mount the shaft-type ones to your globe, so you may want to look into making a little custom 3d-printed piece to fit to it, possibly with 24 notches to roughly match up to time zones (in which case you'll end up with something akin to a sundial ) You may also find a code wheel you can use off-the-shelf.

The simpler ones just count the markings as they move past. More complicated ones can encode absolute positions, so you don't have to count rotations.

Some links to take a look at:
http://machinedesign.com/sensors/basics ... nologies-0
http://johntrimble.com/blog/2013/07/11/ ... y-encoder/
https://www.solo-labs.com/rotary-encode ... mentation/

[Yaytay]

Oh thank you.

I'd already considered and rejected incremental rotary encoders, but wasn't aware of absolute rotary encoders - they sound like exactly what I need.

Jim

[Burngate]

I've been thinking about this, since you first posted, but haven't come up with anything useful
Would it be possible to post pictures of your specific globe (or link to where there are details) - specifically the mounting points, so we can see if there's an easy way to drive it?

[Moe]

Nice idea

If you can hack into the globe itself, you could mount the magnetic compass inside (along with lights, lasers, speakers and anything else that might be cool) and use a slipring. If have one of those globes with a light inside it should already be openable. The advantage of a compass is it always knows it's true position, even if the base is moved.

I have done something similar - the scanning turret on my robot is sits on a 10mm dia hollow shaft which the wires go through, mounted in a 10mm bearing, driven by cogs, one fixed to the tube (10mm bore - other dimensions are available) and and one fixed to the spindle of an offset stepper motor (3mm bore), so it can't 'slip'. Works nicely, but (in my case) is a bit too slow. I have no experience of magnetic sensors though.

I would definitely use a stepper motor. IIRC mine does 4096 steps / rev, which should be almost imperceptible. I seriously doubt you could control a normal DC motor to that level of precision without a gearbox the size of a house. A servo maybe, but only with encoders.

Forcing it isn't a problem in my experience, and in your case it's going to be idle 99.9999% of the time so you could simply switch it off between steps to save power.

[Yaytay]

Ah, a slip ring, another thing I didn't know existed.
I have an Adafruit magnetomer, but I couldn't work out how to use it as a compass for the globe because of the obvious cable tangle. So a slip ring is brilliant.

I don't think I'm going to be able to crack the globe in two - I think it's taped around the equator - but there is an approximately 5cm diameter hole in the base for the existing bulb (240V bulb, to be replaced with a 12V led connected via a Mosfet and a bit of PWM to dim).

I don't think there is going to be any way to get a motor hidden inside the globe, but if I can work out a way to get a stepper motor with a rubber wheel rubbing against the bottom of the globe it might do the trick.
And I guess I can hide a stepper motor in the wooden base - not much worry of it overheating if it's on for a fraction of a second per minute!
Thanks for pointing out that I could just turn it off

I can't do pictures at the moment 'cos my phone is broken (it's taken me until now to join the ranks of mobile phone owners with a cracked screen), but I'll get some up as soon as I can.

Thanks all.

Jim