The mark 2 assembly line. The advancer is at the right, the brick positioner in the centre and the press at the left-hand side.
The mark 2 assembly line. The advancer is at the right, the brick positioner in the centre and the press at the left-hand side.
After making the original assembly line, and seeing all the ways in which it could be improved, I decided to rebuild it totally with the additional object of making it automatically controlled. I did succeed in making it quite a lot better, though it was still based on the same plan - in particular, it still made the same product. The three basic mechanisms, the advancer, brick positioner and press, were still present, but in improved forms:
The advancer still served to move the partially-finished model through the assembly line, but was now a movable 2x4 brick with a studless (smooth) top on which 2 wheel bricks sat; the long axis of the car was thus across the assembly line, not along it as before. It was moved by a motor with rack and pinion as before, with a Lego rotation sensor providing positional feedback to the RCX. When the positioner was in its "home" position, a new pair of wheel bricks could be slid onto it from a rack - since the RCX has only 3 motor outputs this had to be done by hand in thisl version, but in the next version it was done automatically.
The advancer from the mark 2 assembly line. The new style motor with worm drive are at the top of the photo, driving the rack and pinion to their left which moves the 2x4 black smooth rectangle on which the wheels sit. The blue 2x4 brick represents the position of the rotation sensor, which was unfortunately unavailable when I took this photo. The yellow 1x8 smooth plates extending from the centre to the right of the photo are where the wheel bricks were stored. The object visible at the bottom of the picture is the brick positioner.
The brick positioner was again designed to lay the 2x6 plate on top of the two wheel bricks, but was somewhat less complicated in this version of the assembly line. A micromotor drove a rack and pinion which lowered a fork supporting the plate onto the wheels; the wheels then moved on through the assembly line, carrying the plate with them, before the fork was raised again. No feedback was used for this mechanism - all was done on timing.
The brick positioner from the mark 2 assembly line. The red micromotor in the top right raises and lowers the fork that takes up much of the picture; the horizontal 2x6 plate is resting on it, waiting to be placed on the wheels on the advancer below. N.B. For this picture the remainder of the assembly line, chiefly the press, was removed.
The press performed the final squeezing of the plate and the wheel bricks together, and worked on the same screw press technology as in the mark 1 version. However, the part that engaged with the two stacks of wormgears was substantially improved so that it was not continually forced apart instead of pressing the bricks together. In this iteration, instead of two opposed racks held together merely by a few studs, the teeth engaging with the screws were those of small (16-tooth) gears on axles that were held rigidly in the appropriate orientations; when a tooth was needed the ordinary vertical orientation sufficed, but when a gap between teeth was required the axle was held rotated by one sixteenth of a revolution using the (now discontinued) grey bushes with a "crown" of tabs. As this description is somewhat cumbersome, I have made LDraw models of both the original press mechanism and an alternative version which can be built using only the parts included in the Robotics Invention System. The press was driven by one new-style motor with one stage of gearing down (16-tooth gear to crown wheel), and the two screws were connected with a chain around two 40-tooth gears. For feedback on the press' position I used a light sensor pointing at the side of the press, enclosed to block ambient light from affecting the reading, and arranged a stack of black and white bricks so a junction between the two colours would be crossing the light sensor's field of view as the squeezing was completed. Thus the varying reflection from the varying ratio of white to black in the light sensor's field of view gave a continuous measurement of the position of the press and finding the right value was a matter of trial and error.
The press from the mark 2 assembly line. The motor is at top left, and the light sensor in its enclosure is at the bottom right, opened to show the face that is normally pressed against the black and white stack visible from the inside on the right hand side of the press (it is attached to the part that moves up and down). The two stacks of worm gears are visible, as are the four axles holding the gears that engage with them; note that the two outside axles are rotated by one sixteenth of a turn relative to the two inside ones, and are held there rigidly (not just by friction). The end of the advancer track comes through the centre of the press.
After a car was finished, the operator removed it from the advancer, pressed the touch sensor visible on input 1 of the RCX (which caused the machine to reset itself) and reloaded the parts. If you like you can download the NQC programme I wrote to control it.