Physical "contact" connections between modular robot parts?

[corrado33]

Hey guys!
I have an idea.  (Of course, that's why I'm writing a thread  )  Anyway, what do you guys think would be the best way to communicate between two modular components of a robot?  Basically, if I ever finish my robot I'm building now, it's really just a "locomotion" section.  Kind of like an ERP that I can mount stuff on.  Except I want to be able to swap things on and off of it quickly, without plugging and unplugging wires. 

So I my question is, what's the best way to connect two MCUs using simple contact connections.  Basically, when something is mounted to the locomotion section, it'll have a piece of metal that "contacts" another piece of metal on the locomotion section.  This will act as a connection between the two MCUs.

It's kind of hard to explain it, so I hope you guys understand what I'm trying to say. 

Anyway, one way I though of was to use the things that hold the modular parts on to the locomotion section to communicate.  For example, if I used bolts to connect other parts to the locomotion section, I could connect a wire to the one side of the bolt, then connect a wire to the nut on the other side, in the other part.  Would something like this work?  Of course I'd have to use a very conductive metal, but would there be really bad noise if I did something like this?

Oh, I thought of an example.  If you've ever taken the steering wheel off of a car, you've probably wondered, how does the horn work if there are no wires to unplug?  Well, if you look at it closely, you'll notice (I believe) copper contacts that connect the steering wheel to the steering column.  The ones on the wheel act like springs, so they always press down on the copper plate on the steering column, no matter which way the wheel is turned.  Then again, I may be making this up, but at least I think this is how it works... 

If none of this makes sense, just ignore me, I ramble sometimes...

[Cristi_Neagu]

Depending on the joints you use, you could use optocouplers. Having two pieces of metal in contact can be unreliable. They could get corroded, or develop increased electric resistance...

[corrado33]

Depending on the joints you use, you could use optocouplers. Having two pieces of metal in contact can be unreliable. They could get corroded, or develop increased electric resistance...

Hmm.. interesting, I didn't know such things existed... That's very cool, and would only require (precisely) aligned holes in each modular part. 

And I figured as much for the two pieces of metal.  That's why I wasn't sure if anyone had any better ideas!

[Ro-Bot-X]

Take a look at this:
http://www.societyofrobots.com/robotforum/index.php?topic=9461.msg73832#msg73832

[Soeren]

Hi,

[...] Except I want to be able to swap things on and off of it quickly, without plugging and unplugging wires. 

Anyway, one way I though of was to use the things that hold the modular parts on to the locomotion section to communicate.  For example, if I used bolts to connect other parts to the locomotion section, I could connect a wire to the one side of the bolt, then connect a wire to the nut on the other side, in the other part.  Would something like this work?  Of course I'd have to use a very conductive metal, but would there be really bad noise if I did something like this?

You want to bolt/unbolt a piece of equipment, remove nuts and the like, but you think a plug, eg. a modular plug or a Molex will take too much time to remove?

The best communication you get by wire. It's cheap and reliable compared to optical solutions and you still need power, so the optical solution means having a separate battery in each part you wanna be able to swap in.

The horn contacts on cars (some are made optical, with IR to a ring shaped piece of perspex for comms from steering wheel "keyboards" for eg. a CD player) will allow power - but you don't want it to rotate where you mount it anyway, as this would weaken the mounting.

If you use a PCB with a grid of pads of say 2mm each and then mount some ball shaped pogo pins in the bottom of your auxiliary devices, with slightly less than the 1/4" travel sticking out, it would go together not unlike the contacts of a battery handle on a DSLR camera.

Pogo pins are originally intended for ATE (Automated Test Equipment) and comes with various points, some with one point, some with a grid of points, but for your use, the ball shaped "points" will be best.

If you make a template out of metal to begin with, you can make all the equipment have the same foot print and for power hungry parts, you might wanna reserve 2 or 3 points for both V+ and 0V (or you can rely on 0V through the metallic connection). Remember that the pogo pins need to be mounted isolated, so a nylon center block in the add-on may be a good idea.

[corrado33]

Hi,

[...] Except I want to be able to swap things on and off of it quickly, without plugging and unplugging wires.  

Anyway, one way I though of was to use the things that hold the modular parts on to the locomotion section to communicate.  For example, if I used bolts to connect other parts to the locomotion section, I could connect a wire to the one side of the bolt, then connect a wire to the nut on the other side, in the other part.  Would something like this work?  Of course I'd have to use a very conductive metal, but would there be really bad noise if I did something like this?

You want to bolt/unbolt a piece of equipment, remove nuts and the like, but you think a plug, eg. a modular plug or a Molex will take too much time to remove?

I couldn't think of any other example at the time, and I was thinking further down the road, when I wouldn't be swapping the auxiliary devices on and off.  (I.E. it'd be automated) I was just using the bolt as an example, I haven't put much thought into it yet, but I was thinking of using some linear actuators to hold the aux. devices on to the main section.  So, I wouldn't really be doing anything, and that's why I was trying to avoid using wires.  

If you use a PCB with a grid of pads of say 2mm each and then mount some ball shaped pogo pins in the bottom of your auxiliary devices, with slightly less than the 1/4" travel sticking out, it would go together not unlike the contacts of a battery handle on a DSLR camera.

Pogo pins are originally intended for ATE (Automated Test Equipment) and comes with various points, some with one point, some with a grid of points, but for your use, the ball shaped "points" will be best.

Hm, those pogo pins look interesting.  Pretty much exactly what I had originally thought to use.  I like the idea, but I mainly make these threads to find other ways of doing things.  (Or to be told that my idea is the best way of doing things.)

Regardless, Thanks!

Oh...

Remember that the pogo pins need to be mounted isolated, so a nylon center block in the add-on may be a good idea.

What did you mean by this?  I know the isolation part, but a nylon center block, is this to make sure nothing else touches the pogo pin?

[Soeren]

Hi,

[...] I was thinking further down the road, when I wouldn't be swapping the auxiliary devices on and off.  (I.E. it'd be automated) [...]

Cross that bridge when you reach it. If you plan and plan and plan some more, for the time that it's a super-robot taking over the world, you'll never get anything done.

What did you mean by this?  I know the isolation part, but a nylon center block, is this to make sure nothing else touches the pogo pin?

I assume that whatever you want to add would not be massive so it will have a hollow center, which is where you place an isolator holding the pins.
You might urge for a much more sophisticated solution, but it is good enough for industrial applications, so I have a feeling that it may serve it's purpose on your robot as well