EMC2 hexapod: the beginning

The project aims to:
 - prove the possibility to build 5 axis machine tool with decent performance out of cheap imprecise components;
- develop simple and effective desigh of joints and actuators for parallel machine tools;
- elaborate automated methods to calibrate parallel machine tool without expensive equipment.

Available information on existing multiaxis parallel machines was analysed and evaluated. The following considerations and requirements to parallel mechanism and its components were taken into account:
1. Single-type actuators for all dof to minimize the number of components. That leads to fully parallel machines only.
2. The simpliest possible design of all components to minimize manufacturing cost.
3. The lowest precision of base frame so that it could be weld and bolted with no additional equipment.
4. The servo motors preferred to steppers - they do not lose steps, have uniform torque up to maximum rotation speed.
5. The base frame should be tall enough to provide free access to working envelope and wide enough to be steady.



Multiaxis machining requires 5 dof, which leads to 6-dof Stewart-Gough platform or 5-dof pentapod. Pentapods have dextrous end effector and one actuator less than hexapods, but their spindle joints are pretty complicated. Thus, good old hexapod was chosen.

Why not hexaglide? First of all, it is practically impossible to outperform brilliant IMAC hexaglide. Secondly, additional cost of linear guides would exceed small project budget. Thirdly, high requirements to mutual parallellism of rails significantly complicate the base frame manufacturing. Fourthly, linear guides add some extra compliance.

Ballscrews choice was very important to start the project. Market analysis showed that building cheap machine is impossible with local components. This is why six chinese ballscrews were purchased from ebay.

Ballscrews

The next crucial component is motors. Though not desirable, stepper motors were main candidates. Steppers with torque at least 2 Nm require separate drivers, and it already looked like expensive but not really effective purchase. Fortunately, DMM technology 400W servo motors were found and immediately purchased.

Servos

Non captive shaft actuators were chosen as they have the simpliest possible desigh and do not require linear guides to prevent rotation. To increase the workspace, actuators stroke should be significant. This is why ballscrews were made through the joints, so that min/max actuator lenght is 400/1300 mm. To ensure wide range of joints, the motor is mounted as close to the ballscrew as possible. Timing belt and equal pulleys 1:1 used for transmission. Jount bearings taken from auto transmission U joints, the nut rotates in two angular contact roller bearings.

Bearings

The actuator prototype was manufactured to check its functionality. The motor was installed, driver connected to PSU and computer through breakout board. First attempts to move the motor were unsuccessful. No matter which settings were set in stepconf, no smooth motioin was obtained (often there was no motion at all). The problem was the breakout board which inverted the signal. With "Invert" box checked on "Step" pins everything run perfect.

The base frame design was much inspired by IMAC project. The most difficult problem was to get 45 degree inclined planes out of straight channel bars without bendind and sophisticated cutting. The channel bars were cut by abrasive wheel, 60 degrees angles were additionally milled, the mounting holes were drilled. The channel bars of hexagons were tacked with a few weldings, the frame was assembled and hexagons were weld completely and thoroughly. Then it was disassembled, deburred and painted.

The base frame

To be continued...