Please critique my design

cparker

I'm looking for any and all feedback on this design idea for a robot. This isn't meant to be a 'battle' robot of any kind. I want to experiment with steering and suspension a bit on this project. I've build robots before with tank-style steering, and also using omni wheels on the front. Something about depending on specialty parts like omni wheels just doesn't appeal to me.

I'm wondering how important different steering angles on the front wheels are (so called 'Ackerman' steering)?

The suspension is an attempt to cut down on vibration a little.

The forks are much taller than needed in order to accommodate larger wheels as the budget allows. I'm starting with rollerblade wheels that happen to have on hand.

The material is mostly 5/8" steel square tube which I found on ebay for a good price.

I have to admit it's pretty fun learning and using google sketchup.

Please let me know what you think. Too complicated? Too fragile? Too heavy?

Thanks!
Video of model
Picture

erco

Looks very reasonable if automobile drive is what you're after, a refreshing change from the glut of differential and omniwheel designs out there. Rollerblade wheels are durable and grippy. Scooter wheels are similar, just larger diameter. You could get slick and take advantage of your virtual differential: speed up the outside rear drive wheel proportional to the steering servo angle.

Did you consider powering your front steering wheels, AKA front wheel drive? A bit more complex mechanically, but your maneuverability, stability and repeatability would improve.

BTW, Ackerman will also improve maneuverability and repeatability. But you knew that, or you wouldn't have asked. Looks like you factored it into your sketch already anyhow!

You've probably seen the good info at http://en.wikipedia.org/wiki/Ackermann_steering_geometry

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
·"If you build it, they will come."

cparker

Thanks Erco. I was considering the virtual differential idea you mentioned. So, like you described, I'll vary the speed of the rear wheels proportionally to the deflection of the servo. With that in place, I suppose Ackerman steering isn't necessary?

I like your front-wheel drive idea. How would I couple the wheels to the motors?

CP

erco

No, I wouldn't forget about Ackerman steering. You're on the noble path to a real usable chassis, much better than the differential steering robots that are a dime a dozen. IMHO that's about what they're worth for useability and repeatability. The best you can say is that they're simple to build and to theoretically use. But nothing in the world can make a diffdrive go straight and/or repeatable for long. Not all the encoders, wheel watchers, dedicated processors and algorithms.

If you can't tell, I'm no fan of diffdrive robots. Reliable whole-house navigation is where it's at as far as I'm concerned, and you need a solid mechanical chassis to start with. I wrote to encourage you to pursue your chassis just to say that you're going in a better direction than most. You want good traction, large grippy rubber wheels, plenty of torque, no wheel slip, no backlash in your steering train, no slop. Good machining skills, tight tolerances, proper materials and lubrication, attention to every detail is necessary. Most people just get a cheap plastic gearmotor and hope for the best. That's a bad start.

So keep the Ackerman steering. Build your chassis first without any drive motors and make your steering SOLID. Forget suspension. Make your steering accurate and repeatable. Push your chassis around like a skateboard just to make sure you can make it go STRAIGHT when you want, and also track in accurate circles to the left and right. Making that steering linkage robust and repeatable is a challenge all by itself. A standard servo is NOT accurate enough. You'll need a nice custom linkage with a steering motor encoder, mechanical stops, and perhaps microswitch limit switches to verify your straight ahead position. The chassis (without motors) should roll very freely and accurately with Ackerman steering. If the front wheels are scrubbing, you're DOA. Mechanics is where it all starts.

Enough tirade for now, I've probably scared you back to a diffdrive platform. I've got two brand new drive platforms coming out this year that track dead straight and turn reliably and repeatably. But keep on your project, just go slow and accurately.

BTW, if you want front wheel drive, the best way is just to make two mirror-image gearmotor wheel units and pivot them seperately, exactly like your existing drawing shows, with a steering linkage. Use Ackerman steering and throw in differential motor speeds and you should have a great chassis IF you have a good mechanical design and minimal slop.
Cheers!

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
·"If you build it, they will come."

cparker

Thanks! It sounds like we're on the same page. I am interested more in the mechanical chassis on this project than anything else.

So that wheel 'scrubbing' is what can happen if the different angles aren't correct huh? I can see that the wheels would fight each other a bit if the angles were wrong. I.e., they would be trying to go off in different directions.

I appreciate your recommendation on the servo. I was wondering if a pretty standard hobby servo would cut it. I'll try to plan accordingly.

I have my eye on these HG20 motors from robot market place:
www.robotmarketplace.com/products/0-BHG20.html

The HG20s are larger than the motors I have modeled. I'm getting to that part.

I'm not sure how good my machining skills are... hopefully I can get some help with the trickier stuff.

Thanks again!

CP

erco

Those HG20s look like great motors. I'd definitely go for the higher-geared 62:1 ratio than the 20:1 ratio, and stick to 12 volts. Lower RPM and higher torque means more accurate control, and you won't bog your motor down.

And yes, consider a nice accurate gearmotor for your steering control. Rack & pinion-type steering, perhaps. Again, slow with plenty of torque. Slop and backlash are certain death to accuracy & repeatability! You might want to review typical hobby grade R/C type steering designs to see what does & doesn't work for them. Wear & tear eventually creates slop & backlash in any mechanical system, so build it well to make it long-lasting. Consider designing in some type of replaceable bushings that you can swap in and out for maintenance.
Cheers!

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
·"If you build it, they will come."