Steering Servo Torque

     Actually, it will probably require more torque to steer when
stopped than at any other time.

     When we did our DARPA Grand Challenge vehicle, which was a Polaris
ATV without power steering, we first used a spring scale to find out
how much force was required to steer the thing manually when parked
on a rubber mat, the worst case.  From that, the torque was calculated
and a motor (a large Faulhaber servomotor with an inline planetary
speed reducer) drove the steering shaft directly.  The steering
wheel wasn't connected to the wheels; it drove only a pot and an
encoder, so even when driven manually, the vehicle was entirely
steer by wire.

    We put a polyurethane spider coupling on the shaft to keep road
shocks from hitting the gear train too hard.  That's so that if you
hit a curb, you don't break gear teeth.

     I'd have to go back to the records to get the exact specs,
but I know that we current-limited the motor to 11A at 24VDC, and
that was always enough to get the job done.  Lock to lock time for
the steering (which had quite a wide range) was six seconds.

     Our vehicle weighed just about a ton, so you should be
able to do this with less power.

                    John Nagle

You might need to make a simple test rig and then spin the tires
up to the rpm they will be turning at 130 mph. Then test the
force needed to turn the tires.The tires will be gyroscopes and
more force may be needed than you think to change their axs of
rotation. Be careful, as things tend to explode when things are
spun faster than their design specs.

Thanks for all the good advice.  I was thinking the same thing that
the force to turn the tires may be greater at speed than standing
still.

I'll do some research on gyroscopic effects and forces they generate.

John used an 11A @24VDC motor to turn a 1 ton ATV. I was looking at
some DC geared motors.

One was 18RPM with 35 in-lbs of running torque @ .5A, and 65 in-lbs
stall torque @ 1.5A at 12V.
The other was a little more powerfull. I believe it was a windshield
wiper motor and has a running current of 1.5A at 40RPM.

These motors weigh in around 4 lbs, and I feel they would probably be
an overkill considering even the largest RC car servos will only
output around 350 in-oz of torque.  The smaller DC geared servo motor
I was looking at will do 1040 in-oz of torque.

I plan on building my own H-Bridge driver using some MOSFETS, and
wiring into a standard RC car servo board along with an external pot
for the feedback.  I may just layout a whole new PCB with servo board
electronics and MOSFETS all together. I'm still learning about speed
controls and servo drivers, but those are minor things I'll work out.

Since my profession is wireless technology, I'll build my own radio
controller with two way communications and as many channels as I need
for steering, brakes, throttle, engine kill, parachute, and most
importantly a fail safe feature that will actuate brakes, parachute,
and ignition kill in case of lost communications.  I'll add some
diagnostics like MPH and engine RPM that can be displayed on the
controller.

Thanks again for the information.  And any more info or comments are
greatly appreciated.

You will find that stiffness is important too - the interaction of a
gyroscopic wheel with a bump may well set off interesting oscillations. This
all becomes even more interesting if a closed servo loop is involved ...

I suspect that the turning force requirement will turn out to be a
relatively minor aspect of the problem.

I'd look for a non-reversable stiff actuator with no play - something like a
screwed rod type.

Careful testing is called for

Dave

Do people use linear actuators for steering mechanisms?  I've seen
them with stepper motors.  Would a servo system with feedback be
better on a screw and nut type linear system as you mentioned.

I've mainly been focusing on a servo system since I understand them
well, and that's what I typically see for RC steering mechanisms.  As
far as steppers, haven't really played around with them much.  I have
a basic undestanding of steppers, but haven't studied the driver
circuits.  I know the steppers have a lot of holding power which is
good, but I imagine a high torque DC motor (with gear reduction)
should be able to have just as much holding power.

As far as the mechanicals for the steering system, it will be
something very rigid with no play.  I see where using a linear system
may have a more rigid setup vs. a rotating arm onto some steering
rods. Their will be a lot of testing and tuning to achieve the right
combination.

But if anyone knows of a steering mechanism that has been tested and
may suit my application better, please let me know.

Thanks.

Just as a point of reference you might want
to take a look at http://www.nuge.com/~jjn/jeep/
it is a MUCH SLOWER vehicle, but you might find
some of the information there in the steering section useful.

Best of luck with your project

Thanks for the link.  My first attempt (and hopefully last) at the
steering system will be similar to a go cart.  Not sure what the exact
terminology is, but it will have 2 rods with adjustable round ends
that will house some sort of hard rubber bushing.  One end will
connect to an arm which is controlled by the servo, and the other end
will connect to a knuckle/spindle assembly where the tire will mount.
Since I will only be turning between 10 and 20 degrees, and mostly
going straight, I feel this will give me a solid front end.  It can be
set up very stiff depending on how tight I crank everything down.  And
with a high power servo setup I'm hoping it will steer fine without
any free play.

This may sound dumb, but the TV show Myth Busters (Discovery Channel)
often use large RC servos.

There was an episode where Jamie Heinemann used a servo on a rocket
propelled Chevy Impala (you just *have* to see it to believe it). He
operated the serrvo with a regular RC controller (from a chopper) I
don't know the specs, but you could email them and they might let you
know what to use.

The servo was surprisingly small for that sort of job  - a
rocket-propelled car!

:-]

Dale

Terry wrote:

Hmmmm.... the choice of car might be telling. The Impala (at least the
older models, which I've driven) had non-variable assist power steering
with a feather touch. You could steer it with your fingertip.

I kinda doubt that a robotic car that weighs 50 lbs will itself have
hydraulic power steering as an assist mechanism, so the motorized
steering has to be strong enough on its own. That pretty much rules out
even the largest of RC servos, which top out at around 400 oz-in and
have a maximum swing of about 160 degrees.

In any case, Jamie writes for RC Driver magaxine and a couple others.
Maybe he's talked about this techniques there.

-- Gordon

That's an interesting point.

The difference here in Australia is that even cars from the 1980s often
don't have power steering, and rarely I have never heard of power
steering in the locally made standard Fords and Holdens of the 60's and
70's.

Cheers

:-]

Dale

Gordon McComb wrote:

I have seen a few remote control full size cars.  Some use air power
or hydraulics.  I do watch mythbusters and have seen their remote
control cars.  For my 50lbs RC car, I plan on using a dc motor in the
range of 800 to 1600 oz-in of torque.  If I need more I can gear
reduce it further, but I feel that this will be more than enough for
my setup.  And if I design the front it very stiff, it should be
pretty stable at top speed.  Of course the tire width will be a
factor, so I plan on using the thinest contact patch I possibly can.

I am most likely going to use the MC33030 servo driver IC with my own
circuit for a closed loop feedback and H-Bridge to handle the extra
power.   I have a decent setup that allows me to make my own PCB
boards (2 layers) with 10mil min spacing, so I can really play around
with my design untill I get it just right.  Cost only being the 2
sided .063 copper boards.

The servo motor I was going to order is no longer available from the
surplus store, so I'm looking at a few others.  As soon as I purchase
the motor, then I will build the circuit for the right amount of power
and begin some testing.

An email to the mythbuster team sounds lke a good idea.  I'll look
into it.

Thanks.