Subject:
|
Re: Newbie needs Help (diff sensor)
|
Newsgroups:
|
lugnet.robotics
|
Date:
|
Wed, 7 Jun 2006 16:09:32 GMT
|
Original-From:
|
steve <sjbaker1@airmail.STOPSPAMnet>
|
Viewed:
|
4404 times
|
| |
|
|
Brian Davis wrote:
> One simple way (that honestly doesn't work very well... shows how complex
> friction is): the coefficient of friction is just the tangent of the slope angle
> when the object "breaks free" and begins sliding down the hill. So you can build
> a four-wheeled platform (with the axles locked, so the tires don't turn) and put
> it on a board, raising one end of the board until the platform starts sliding.
> I've done this, but for various reasons I don't completely trust the results.
> It's still interesting and informative, however.
The problem is that this is measuring the static friction ("stiction").
That's the wrong thing to measure if your strategy is to jam all the
motors full on and progress forwards (we hope!) with all wheels
spinning.
To measure the dynamic friction ("sliption") you can use the same
locked-up wheel rig and just pull it along flat ground with a
spring balance. Drag it along at a reasonably constant speed
where it isn't jerking along.
So you have two numbers - one for sliption and the other for
stiction. With rubber tyres, those numbers should be very different.
So which one do you want to optimise for? If you have two wheels - one
with great stiction but terrible sliption - and another with
great sliption but poor stiction - which should you choose?
Well - it depends on your strategy.
STRATEGY I:
You plan to drive mindlessly forwards at full speed with all wheels
slipping...several people here have advocated that,
Forget about the results you got for stiction on your inclined
plane - they are quite irrelevent. Look instead at the sliption
figures - your wheels are slipping all the time - so that's the
right number.
Furthermore, realise that the high school physics equations for
sliption say that the frictional force is proportional to the
weight - but NOT related to the contact area. People find that
counter-intuitive - but it's true. So don't bother putting 10
driven wheels on your vehicle - it won't help one bit once you
are out there with all wheels spinning. Tank tracks are useless
for the same reason - they increase the contact area - but if
they are slipping, that's irrelevent.
Stiffer rubber is better because you aren't wasting energy deforming
the rubber in the tyres as they spin and you want to pick the wheel
design with the highest DYNAMIC friction.
STRATEGY II:
You plan to take the slow and steady approach where you don't let
the wheels slip and carefully meter the power to make sure they
aren't slipping. I guess I'd measure the speed difference between
a driven wheel and an idler wheel - then progressively increase
the motor power until the wheel slips - then back off until it
stop slipping - then you can gradually increase power again until
it slips again...just like the ABS in a car.
You'll be taking advantage of that higher stiction coefficient - and
you'll need to be aware that stictional forces ARE proportional to the
contact area...so you want as much rubber in contact with the ground
as possible. Nice fat, soft tyres - the ones that have the highest
stictional coefficient. MAYBE even consider tank tracks IF you
support the bottom of the track with a bunch of little idler wheels.
WHICH APPROACH WORKS BEST?
Well - let's look at some real-world vehicles. My car is a heavily
modified MINI Cooper'S - if I turn off the traction control software
and floor the gas pedal from a dead stop, the wheels will spin like
crazy - there will be tons of smoke and my 0-60 time will be 8 seconds.
If I carefully apply the gas so the wheels don't spin - but are just
on the point where they would if I pushed just a little bit harder -
then my 0-60 time is close to 6 seconds. That's the difference between
stiction and dynamic friction.
Look at a dragster - huge tyres - made of the softest rubber compound,
pre-heated to make them yet softer and run at very low tyre pressure
to keep as much rubber on the ground as possible.
So - *IF* you can manage the mechanical and software sophistication
of measuring and controlling the motors so you apply the maximum
possible power without the wheels slipping - and *IF* the laws of
physics for rubber tyres are similar at the scale of a car versus
a small robot - then going the slow-and-steady route with lots and
lots of big, fat soft tyres should theoretically produce much better
results than just jamming the motors on at full speed.
|
|
Message has 1 Reply:
 | | Re: Newbie needs Help (diff sensor)
|
| (...) As you note in your mini-cooper example, *if* you can keep the wheels from "spinning out", but instead are always in rolling contact with the ground, static friction is what's important... and generally, the static coefficient of friction is (...) (18 years ago, 7-Jun-06, to lugnet.robotics)
|
Message is in Reply To:
| | Re: Newbie needs Help (diff sensor)
|
| (...) A rather inventive young man that I know did try exactly that - using the ribbed tubing from the Mindstorms set inside the Mindstorms motorcycle tire. It's a good fit, and a good idea, but I'm not sure how effective it was. (...) One simple (...) (18 years ago, 7-Jun-06, to lugnet.robotics)
|
39 Messages in This Thread:
 
 
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
This Message and its Replies on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
