Gyroscope project : Journal

Ralph Hempel's interface and the RCX platform allow only two states:

  1. motor power is set to 6: this causes the RCX to send a positive pulse of 7 msec and a 0-pulse for 1 msec. The interface inverts the signal.
  2. motor power 5: 6 msec HIGH-state; 2 msec LOW-state

The programmer has to observe the turning speed of the servo, or -as overkill- to add angle-reading for exact adjusting.

/* servo.c*/
#include <unistd.h>
#include <dmotor.h>
#include <conio.h>
int main(int argc,char **argv)
{
  int k;
/*start the motor at power 0*/
  motor_a_speed(0);
  motor_a_dir(fwd);
  msleep(1000);
  cls();
  lcd_int(0);
/*speed up gradually and display value*/
  for(k=0;k<=MAX_SPEED;k++)
     {
       /*speed the motor up a notch*/
       motor_a_speed(k);
	   lcd_int(k);
       /*wait a second to observe*/
       msleep(1000);
     }
  motor_a_dir(off);
  cls();
  return(0);  
}

Conclusions: there is no possibility to send a pulse between 1 and 2 msec using the RCX with legOS or the standard firmware. We'll have to design an own interface.

10/4/2001

Andreas Peter has designed a Two RC Servo Motors Interface to the RCX. He has some very usefull ideas! Perhaps he could give us the schematics and the PIC-program? We'll ask him.

13/4/2001

We contacted the manufacturer of the IKARUS Gyro, but the details we got about the sensor are nevertheless scarce. After a short analyses, it becomes obvious, that the brain of the gyro is a PIC 16C711. This PIC has a build-in A/C converter. The conditioned information from the TOKIN ceramic sensor must enter the PIC somewhere. We found that it is actually Pin 11 of the PIC. So we tried to buffer the signal first to our lab-instruments, then to the RCX via Mike Gasperi's RCX-input-buffer. (The 30mA current from an RCX-input is not sufficient to drive the gyro. So an external battery is needed). Don't use an RCX-motor output!! Because of ground-problems.

The gyro was mounted on the RCX (It is possible to stick it between two Lego-bricks!) Gyro-sensibility manually to 100%. We chose our direction-master (I) for the tests. We programmed some heading changes and stored the data in the memory through datalogging. We captured the values towards the PC. The heading is expressed in degrees, dH/dt is the derivative of the heading through time. The gyro values are raw values from 0..1023. To fit them in the graphics, we used the formula y=800-gyro_raw.

--> The gyro gives the derivate of the heading. Some mathematical transformation has to be done. First the zero-point has to be detected, then probably a proportional adjusting. 

17/4/2001

If the gyro is used as designed by the manufacturer, there is NO need to build an own servo-interface, since Ralph Hempel's interface may be used on his PBForth platform. PBForth has two keywords SERVO_INIT and SERVO_KEY to generate any pulse width from 0.50 to 2.50 ms. It's even possible to drive two servos from one output (by setting the mode-option and replaceing the Graetz-diode-bridge by an unilateral circuit. (book: 'Extreme Mindstorms' p.313)

20/5/2001

After a longer pause we work again on this project.

We operated some additional tests using the following device. The rotation sensor is connected to RCX input 2 :

RESULTS: SEE THE ROBOLAB-FILES

30/10/01

We are working now with Robolab 2.5 ! This marvelous program has some very interesting features such as the possibility to integrate textual program-parts. Many things are improved from the previous version, but there remain some bugs, specially in the graphical part of the compute-area. Nevertheless, we are delighted with all these new possibilities. This software makes robot-programming reachable for secundary students !

 

RetourMain Page