Shop OBEX P1 Docs P2 Docs Learn Events
1st project with SX chip Power Wheel Jeep R/C mod — Parallax Forums

1st project with SX chip Power Wheel Jeep R/C mod

Jeremy HodderJeremy Hodder Posts: 2
edited 2007-07-20 13:12 in General Discussion
I'm looking for suggestions here what i got so far.
Starting idea
Sx48 proto board
R/C (Vex transmitter and receiver)··
Manual analog input (brake and throttle, Potentiometer or Hall Effect)
H Bridge 30-40amp (drive motors) Any designs that work good with the SX chip ?
Steering Servo (needs to be strong, gearbox and pot?, 2nd H bridge and analog input)

Advance idea
Follow robot
Temperature sensor
Amp sensor
RPM sensor (speed governor)

thanks
Jeremy

Comments

  • Philip GamblinPhilip Gamblin Posts: 202
    edited 2007-07-20 04:55
    You gotta look at the HB25 by parallax. Great motor control.
  • Dennis FerronDennis Ferron Posts: 480
    edited 2007-07-20 07:46
    Those powerwheels gearboxes are great. I built a robot that uses 4 of them. I decided to forget about the whole steering issue, and I just run four fixed wheels with "tank" steering (differential).

    I had trouble finding an affordable motor controller that could handle the current I needed, so I decided to build my own. 18 months later I am still working on the blasted motor controller nonstop, because every time I think I've fixed all the issues, something else comes up. The basic idea of PWM control is simple, but when you're controlling motors as big as the ones in a powerwheels car, issues that don't show up with smaller motors become a big deal suddenly. I would have been better off to spend $200 for a good off the shelf one. On the other hand, I'm wondering if some of the commercial ones maybe have issues too, because I uncovered a lot of things that do not stop the controller from working, but which would contribute to it wearing out faster or sending voltage spikes back out over the power supply lines.

    For the steering servo, here is what I would do, if I were going to use analog steering on my 'bot. I'd simply use an extra powerwheels motor and gearbox, same as the drive wheels, to turn the steering wheel. Instead of using another motor controller, you can use this gearbox to "amplify" the power from a very small servo. Using some plywood or something, just make a plate that goes on top of the steering wheel, and mount 2 single pole double throw microswitches mounted on it. Put the small servo in the middle of the "giant servo" so that when the servo turns one way, it's arm bumps one microswitch, and if it turns the other way, it bumps the other microswitch. The idea is that bumping one turns the wheel left, and the other turns the wheel right. So if you move the servo, it will bump the switch and make the wheel move, but as soon as the wheel moves far enough (the switch moves with the steering wheel) then the switch won't be "bumped" anymore and it'll stop turning.

    I've also seen powerwheels steering done using a screw apparatus (like from an automatic car window). And I've seen yet another that used a bicycle sprocket and chain.

    You definitely want the speed sensor. I didn't have one and it causes me trouble because the motors don't have a linear response. They'll tend to do nothing... do nothing... and then suddenly jump to nearly full speed, as you turn up the PWM power. So you want that speed control so you can get a reasonable response and not jumpy. It's difficult to mount anything to the motors for measuring speed. What I plan to do now, is, you know those shaft couplings between the wheels and the gearboxes? They look like cups? I'm going to print a striped "bar code" out on a long strip of paper, and wrap the striped paper around the shaft coupling. Then I'm going to just use an infrared sensor to read the pattern of light and dark. You don't need full positional encoding since you know which way you are moving your motor, you just need a count for speed.

    Temperature: definitely. I burned up the MOSFETs in my first version of the controller. Literally had a fire in the electronics box. If I had had temperature sensors in there, I could have had the controller automatically shut down long before things got so roasty toasty. I'm using thermistors directly in my new design. This is a dumb way to measure temperature, I now know: thermistors are highly nonlinear and difficult to read and calibrate for. Instead, there are digital temperature sensors that handle all of the thermistor calibration, calculations, etc, and are accessible through a serial interface. Don't do what I did here - use these digital temperature sensors. They come in 3-lead packages just like transistors, and there's no reason not to use them.

    Current sensing: Why mess with a shunt resistor? That's the old way to do it, but we have better ways now. Allegro makes these really neat hall-effect current sensors that provide an output between 0 and 5 volts that is proportional to the current flowing through the isolated sense leads. It doesn't dissipate any power like a shunt resistor, and you can take the current sense output and run it right into an A/D converter. You can also locate the current sense lines anywhere in your circuit, not just certain places like a shunt. I'm using two of the Allegro current sensors on my motor controller right now.

    Post Edited (Dennis Ferron) : 7/20/2007 7:52:36 AM GMT
  • Jeremy HodderJeremy Hodder Posts: 2
    edited 2007-07-20 13:12
    I’m thinking of using something along the same idea as here http://www.fieroaddiction.com/servo.html for steering.
    Anyone have some broken servos I could take apart?
    ·
    And for powering the drive motors, scooter controller, looks like they use a Hall Effect sensor that could be faked with a digital to analog chip. Or http://www.robotpower.com/osmc_info/
    ·
    I’ll have to look in to digital temperature sensors and hall-effect current sensors
Sign In or Register to comment.