Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

How do I size a motor to steer a vehicle?

  1. Aug 6, 2004 #1
    I'm looking into implementing some parts of drive by wire on an SUV and need assistance determing the load requirement for steering effort. Braking force is widely published as roughly 150 lbs of pressure but I am unable to find a similar estimate of force needed to turn a steering wheel. I'm looking for a simple rule of thumb for sizing a 12v DC motor in a chain drive application. The motor might be a stepper motor or servo for ease of control.

    My current thinking is to have a sprocket welded (or similarly attached) to the steering shaft between the steering column and the power steering unit. The motor would be able to drive the steering shaft via a chain, and when the electric motor is unpowered the vehicle would be manually steerable (assuming that the ignition key is in place and the steering column is unlocked). The chain would retain alignment in spite of any flexing due to rough road conditions, out of balance tires etc. Some older marine autopilot drives work this way; they even have a nice failsafe clutch in the middle!

    I'll have to work out how many teeth to use on the motor and the shaft sprockets to give an output RPM in the desired range. How do I figure out much power I need in the DC motor to turn the steering shaft under load? I don't want it to be too small to turn or hold the load (which is why a stepper or servo might be a good motor choice).

    Thanks for any assistance ... :)
  2. jcsd
  3. Aug 6, 2004 #2


    User Avatar
    Science Advisor

    Empirical data will be your friend. You can use a torque wrench to attempt to calculate the effort required, BUT you will need to make sure you test under realistic conditions. If you test on a smooth painted concrete surface that is wet your measurements will be very low. If instead you tested the steering effort with the motor turned off in a soft dirt field you're readings will be tremendously high. Surface conditions would vary in-between, where is this system to be used?

    You are going to need some sort of feedback, a servo could be a good choice. However running the DC motor at stall could cause it to overheat (depending on cooling method) and burn out the brushes if they arc (which could easily happen depending on design and the position they're in).

    A stepper works best at stall and generates the most torque near 0 RPM. The torque falls off tremendously as RPM increases and looks similar to a capacitor discharge curve. Since a stepper can lose steps and you will be able to operate the system without the steppers you will need feedback as well, and probably want it to have absolute positioning so the system will have the tires pointed straight ahead when you want it to and not turned! A potentiometer would be far easier than an optical encoder.

    This could be very involved depeding on how well you wish to design this. Oversized would be better than undersized and you'll likely want a gear reduction motor or use a gearbox to offer torque multiplication and sacrifice speed. The safety aspect needs to be addressed too, hopefully you have that well covered.

  4. Aug 9, 2004 #3


    User Avatar
    Science Advisor

    Just a suggestion which may help you visualize what you need to do would be to look at an RC car, which uses servo's for the steering. Then it would simply be a matter of scaling it up. You could perhaps look at a 1/4 scale off road truck, get an idea of how they made it work.

    What are you planning on use to control the motors? You could go as far as simply making the steering wheel control the motor as well, so that it would still be drivable in the old fashion, but not require any excess parts.

    Can't give ya much help on the technical specs, but I think you would be wise to check out the current electrically steered vehicles out there.
  5. Aug 13, 2004 #4
    Thanks for the ideas etc. I've found a gearmotor with an optical encoder which can be controlled easily in a closed loop interface. I'll use the sprocket gear ratios to get whatever RPM works best in terms of RPM on the driven shaft. The gearmotor I have in mind is backdrivable when unpowered so all I'll have to do is drop power to the motor to make the vehicle manually steerable.

    I'll probably use a Roboteq 2850 motor controller with the new OE support. It can be driven by a PC or microcontroller, or even via a joystick or R/C interface!

    In terms of safety, I will wear a t-shirt that says "Televehicle Test Team Member" on the front and "If you see me running try to keep up!" on the back ... just kidding, it won't even be mobile until an e-stop interface is working.
  6. Aug 13, 2004 #5


    User Avatar
    Science Advisor

    That sounds awesome. I've always wondered how one would go about hooking up this type of system. Any chances on getting some detailed info from ya about it someday?

    Ok, I admit it, playing grand turismo spoiled me to using the analog controllers, and I wanted to drive with the ps2 controller (which, btw, you can get usb convertors for.)
  7. Aug 20, 2004 #6
    Alternate fly-by-wire

    Basically what you are trying to do is what most power steering units already do; they just use the steering wheel as the control sensor. You should find it fairly straightforward adapting the power steering unit to do what you want, especially if it is one of the newer electric units. I know that some Saturns have electric units. For the hydraulic ones you would simply have to connect to the hydraulic inputs to the pump and the ps unit from your own servo valve. For some idea how that works see the following:


    It also discusses "Fly-by-wire"

    Best of luck, but be very careful on however you want to do it: it can be dangerous if you don't handle it carefully. Also check your local and state laws. They may have something to say about what you are allowed to do.

  8. Aug 22, 2004 #7
    Manual takeover needed

    This would be more a matter of actuating the existing steering shaft (ahead of the power steering unit) via chain driven from a servo motor. I would not want to play with the hydraulics, although that is possible. The problem with that approach is that it would not be manually drivable. I want to drop power to the servo motor and drive normally, which is similar to the way that HDS equipped vehicles are designed. Some of those have a nice clutch between the motor and the steering shaft to engage the steering motor.

    In case anyone asks, the rule of thumb for torque is 200 in lbs to actuate steering. I found that at the EMC site (digi-drive.com) where they discuss their primary driving controls products.
  9. Aug 22, 2004 #8
    Remote control of a full size vehicle

    Take a look at http://www.roboteq.com/rcauto.shtml for an overview. They don't say how to get control over the drive shaft but anything which can drive it via chain, spider coupler, or even a replacement gear set instead of steering wheel would work. In my case I'm favoring a gearmotor parallel to the steering shaft with a sprocket and chain coupling between them.

    Unfortunately my funding is slow to arrive so I might not be able to actually implement drive-by-wire on a vehicle although I've got some of the technology researched. Throttle is easiest, even a cruise control kit can control that, steering and brakes are harder since they are not routinely controlled by an autopilot.

    If I am able to implement what I'm trying to do I'll send some video.
  10. Aug 22, 2004 #9
    I must disagree with your assumption. Not only would the approach that I mentioned to you be contrallable with the steering wheel; it would be active while the fly-by-wire system is engaged. In effect, the systems would be additive. Anything that the remote/automated servo did could be corrected (or augmented) from the steering wheel. This is done by connecting the steering wheel servo and the fly-by-wire servo in parallel. The catch is that these two would have to be carefully balanced to keep one from effectively overwhelming the other when they are both active. This means that the F-B-W servo would have to be proportional like the steering wheel servo, and its pressure ranges would have to be roughly the same.

  11. Aug 24, 2004 #10
    So adding a steering wheel servo to balance out the first servo ... overkill? It certainly sounds more complicated than adding a single servo to control steering which is the desired result. Take a look at the Roboteq site I mentioned above to see an overview of what I'm trying to accomplish.

  12. Aug 25, 2004 #11
    Several observations, I believe, are in order. You may evaluate them as you wish.

    1. The power steering mechanism on the steering wheel already has its servo, You'd have to find a servo valve that balances it.

    2. The idea I suggested is a parallel approach whereas the sprocket drive is a serial approach. Generally speaking, parallel approaches are more difficult to grasp, but simpler to implement.

    3. The addition of a hydraulic servo valve puts the steering wheel and the F-B-W servo not only in parallel, but makes them additive. The only reason desired for balance is so that (mainly) the F-B-W system won't overpower the steering wheel input, which could be a safety problem if the F-B-W system went out of control. This precaution allows you to always correct with the steering wheel. If the parallel servo simply failed, you'd have the same problem that you have when power steering fails. You have to muscle steer, but you still have steering.

    4. With your approach - you have a serial system. You are basically using the motor and chain to drive the PS servo. When you were seeking the motor torque necessary you were seeking a crude measure of balance. Of course, balance is less important in this case because you can't correct anyway for an out of control F-B-W system until you turn it off or disengage the motor, an added step. Its like having to turn the engine off to use the brakes.

    5. Some steering shafts are offset, using universal joints. If you try to add an external sprocket drive to one of these, you're asking for serious trouble.

    6. Mounting a motor and sprocket drive in most vehicles is a serious engineering feat. The manufacturers don't normally leave you a convenient place to put these, yet they must be carfully placed to work properly. Building this mounting assembly will probably be an expensive engineering job. A parallel hydraulic servo can be placed almost anywhere. That said, I couldn't say where you'd find a proportional servo valve these days but I can suggest how to make one. Use an identical PS servo valve and a stepper motor to drive it.

    7. To make the parallel system work even better, put a detent on the steering wheel. It wouldn't hinder using the steering wheel, but would take out any tendency it might have to want to wander.

    Also, I did read the article, and have not changed my opinion. I would suggest, however that you take their caveat very seriously. It is a dangerous undertaking (either way, but especially the parallel approach).

Share this great discussion with others via Reddit, Google+, Twitter, or Facebook