Let's Make Robots!

What's on your dream robotics shield?

Right now, I have 3 platform modules,
  • LCD + UI
  • Prototyper
  • Propeller

I'm going to make 4 more modules in October.  3 have already been completed;
  • DMX I/O
  • PICaxe
  • Video + IP

I'm thinking the 4th should be some sort of robotics module.  I'm not sure what to put on it, though, so I'm looking for any suggestions.

Remember that the ucontroller would be on another module, and the module size is 3.8" x 2.5".  What do you think about;
  • Just H-Bridges / servo controller with some power control?
  • Including mounting holes that fit some of the dagu robots?
  • Should it have sensors, or would it be better to add sensors to another module?
  • What else would be cool / useful to have on the board?

I'm still doing my initial homework, any input would be really helpful.  Is it better to throw a few motor controllers on a board and call it a day, or is there a way to make it more useful?

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

My dream shield would be a servo shield a la this one BUT with 2 x 16-channel analog multiplexers and an 8-channel shift register (for controling the multiplexers). It would then have 36 analog in's + 16 easy-to-use tripple servo pins + 7 free digital I/O's (also capable of controlling servos). The remaining space would be a tiny breadboard.

Man I'm drooling allready :)

I can do partially pre-assembeled, like put on the surface mount parts, but leave the rest to the builder. Or fully assembled (but if you want to drive a servo, you're probably comfortable with a soldering iron).  Yeah, I was assuming that I'd have to put something on there to convert the i2c for the h-bridges.

I2C does does add complexity & cost.  The ucontrollers I'm supporting (28X, 40X, and Propeller) all have a ton of I/O, too.

If people who use those ucontrollers are running out of I/O, then I2C makes sense.  Otherwise it would be more cost effective to let the host do it.  Honestly, my robotics experience is limited - do most folks run out of I/O with those bigger ucontrollers?  The 40X and Prop have 32 I/O.

Agreed, if it's just an h-bridge or two, an added micro is overkill. Like robologist said, more complication and delay.

If it has two h-bridges, 4 quad encoders, and two temp sensors, that's 18 to 22 IO right there (2 or 3 per motor, 2 per encoder x 4 motors). Still not really pushng the 32 IO of a prop or 40x, but close enough that running out of IO could become a possibility.

I admit, my idea is probably overkill in general. I imagine that someone building a bot which needs that control over 4 motors is likely to knock out their own solution instead of buying a prebuilt one.

As a different preference, I really dislike talking to another processor to do motor drive commands. I'd rather directly drive the h-bridge input signals from whatever micro I'm using. I2C for me just adds another layer of complication and delay, which I've been annoyed with in the Devantech stuff about.  Just a different perspective.
Ha! Well - the whole idea of a modular system is that some of those things would be in different modules.  The protoshield already is available and GPS is on the shortlist of modules for the next run.  So this module doesn't need those things because you can just stack a GPS and protoboard on.  

For power, I'm thinking I might just have pin headers to an RC pack or other battery.  I'll eventually do a charging module, but I'm not sure if that should be part of the robotics shield.  That flexel is pretty cool, too.  

@Rudolph, it sounds like you think I'm better off keeping the modules 'modular', so there isn't too much non-motor control stuff on this board.  It would also be easier to design, and other designers who want to use it for non robotics based motor control could use this board. The only magic would be including a straightforward way to address the motors.  

Is the big difference between this and ladyada's motor shield the addressing features?  

How much peak current is enough?  (aka, how wide should I make the traces) The more the better, but do you think 3A would be about right?

The way I understand LadyAda's motor shield is 2 servos or 4 motors or 2 steppers (I could be wrong). The other major difference would be the encoder inputs being right there on the motor shield. And yes, the addressing would be different, her's just uses straight Arduino pins (most of them even), "mine" would use a maximum of 2 (TX/RX or SDA/SCL).

For the H bridge I'd probably go with the L298 which is 3A peak. A drawback is it may not be happy with lower voltage motors (< 6V, so I've been told)

Looking back on my first post I realize it wasn't clear that my idea would require its own onboard microcontroller.That'd make the I2C easy, the encoder functions would be on board, all the user would have to do is send an I2C packet of "set motor A 50%" or "both motors forward X ticks". I think it should take ticks, not mm/inch, because wheel diameters would vary by builder. Unless part of the command set included "set wheel diameter to X" and "use metric/imperial units".

Are you looking to offer a pre-built shield or a kit? A surface mount ATMega would be great but not if it's a kit. A through hole ATMega might be too big to stuff on a board with 2 L298s with heatsinks, a wad o' diodes, and a bunch of pin headers.

Chances are I've thunk out a $80 board though... ;)

this is going to end in a 2 foot thick ard, but 40 servo ports. A gps Sensor, 3 ports for Gyros, a proto shield, usb datalogging capabilities, tons more memory, etc.

I'm not sure which would be more appealing, but some form of Wireless, like the 434mhz RF kits, XBEE, or Bluetooth.

Also, maybe a recharging circuit for the battery, including the port for charging while in use or whatever.. haha, sorry, haven't used one of the chips before, just heard about it! http://letsmakerobots.com/node/9272   there it is. I'm not exactly sure how to work out specs.. or if it would even be worth while, since there are so many different variations.

But of course the H-bridges, but you may be better of designing your own, to handle higher amperages. (Because you know, bigger is ALWAYS better:D)

Also, I seen a post about a chip that has USB host capabilities.. which isn't a MUST, but would be very nice for.. alot of uses! For example, wireless cameras that are USB based, storing data from your robot on a thumb drive. (this is just an Arduino post, I believe the chip is accessible with Serial, so pretty damn universal:  http://www.practicalarduino.com/news/id/371  )

Here we go:

http://letsmakerobots.com/node/4373

That would be a nice addition.

But of course, there's always the cost issue.. and I didn't put that into consideration, just threw out ideas. :D 

 

I'm sure I'll think of more shortly! :) I'm anxious to see what you come up with.

A full-tilt motor controller board would be slick. Two high power h-bridges to run four motors individually, and quadrature encoder inputs for each of the four motors. Make it handle both I2C and Serial (TTL, maybe with optional spots for RS232 components) so it's stupidly easy to talk to it from any controller. The ability to handle servos on top of it all would be really nice, as many as possible. If there's a couple analog inputs available after that, a pair of tempurature sensors (one for each h-bridge) could be neat too.

I'd leave all other inputs to the "host" controller. Since its I/O won't be killed by the h-bridges and encoders there'll be plenty of space for other sensor inputs there.

i want one of these, would make the design of robots so simple and effective
plus its got loads of cool features
nice one rudolph