UPDATE - This page describes an early development of the servo control system for the DIY motion platform. As of Summer 2009 a new ready-built control card - the 40SPU-1, is available to buy which replaces all the breadboard electronics. See the 40SPU-1 page for more up-to-date details.

The Signal Processor and its PICAXE chips work with low currents – nowhere near enough to directly drive the 200W DC motors used in the cockpit. This task is carried out in the system using bought speed controllers. I've used several of these on the various electric vehicle projects you can see on the rest of the site and I thought initially I'd go with controllers I was already familiar with for the movement cockpit prototype. Specifically I looked at three of 4QD's 24V Vortex controllers - a 35Amp unit for each rotational DOF drive and a 75Amp unit for the heave drive. These did work but were troublesome; they are of a type of speed controller that uses internal relays to reverse the drive direction and the switching of the relays caused a considerable amount of electrical noise which caused instability in the Signal Processor's micro-controller chips. This was largely solved by adding capacitors and diodes to the connections between the chips and speed controllers but still left a problem with the speed and smoothness with which the drive motor direction could be reversed. I also felt that the frequent relay firing was starting to cause some mechanical problems with the relays sticking which can have interesting results - certain failures cause the controller to think it's driving the motor in the direction opposite to its actual motion!

I eventually changed the Vortex controllers out for smaller Devantech MD03 H-Bridge controllers which implement their direction reversals using solid-state switching and are much "quieter" audibly and electrically. The latter makes life much easier for the chips. Important also is that change of direction can be accomplished without any delay arising from the controller mechanical operation and this results in much smoother movement at the cockpit. The experiment with the Vortex units was very useful however as it also gave me a much better feel for how much work the speed controllers actually had to do - not as much as I'd first thought. The MD03 controllers are only 20 Amp units and only the one driving the paired Heave DOF drive motors shows any sign of warming-up in operation.

The MD03 units will accept a range of control signals including the capability to follow a 0 to 5V PWM speed demand signal and this is how they are driven by the Signal Processor Unit. They seem effective at holding the cockpit at any demanded position even when the cockpit load is unbalanced - eg they are quite happy driving the heave motion when the seat is unoccupied and there is a significant unbalanced torque acting on the heave motors.

Putting the control system together (The SPU design has now been updated - click here)

Below is a wiring diagram for the overall system. This can be read in conjunction with the wiring diagrams for the Signal Processor Units to understand how the electrical system is put together. Click on the image for a bigger printable version.

The Signal Processor Unit is connected to the PC via two serial connections - one for the chip programming and the other to receive the motion data. I have both of these connected at once to different COM ports to allow me to make changes to the PICAXE chip programming when I need to. However if you weren't interested in changing the chip programming the same COM port could be used for both tasks - the "Motion Data" connection could be made after the chips had been flashed. You will have to cut up a couple of serial cables to get the free wires to connect to the board - or fit serial connectors directly.

The high current cabling for the electric motor drive end is made with 2.5mm^2 35Amp cable which is considerably thicker than the 1/0.6 mm single core equipment wire recommended by PICAXE for the bread-board wiring.

The multi-turn position feedback potentiometers are mounted on their respective motor output shafts so that they signal the actual position of the cockpit movement. It is very important that they are wired so that the direction of increasing voltage output from the pot matches the +ve direction of movement of the motor - otherwise the position feedback loop will be unstable and the cockpit will constantly drive into its end stops.

The end-of-travel limit switches should be positioned so that they are triggered and held open as a DOF approaches its end of travel - best in fact to set them some distance short of the actual physical end-of-travel to allow some stopping distance for the drive. Often when things go wrong (just when they are needed) the system will end up driving towards the end stops with considerable speed and some distance will be needed to decelerate the motion after the power is cut. The 5V logic power supply to the MD03 motor controllers is wired through the limit switches and the master on/off switch in series - so any limit switch moving into an OPEN position will cut the drive to all the motors. It is a good idea to position the master on/off switch where it can be used as an emergency cut-off switch by either the cockpit occupant or an on-looker. Additional switches can be added in series as needed.

A 9V DC regulated power supply will be needed for the PICAXE Experimenter board, I have used a 1200mA supply which doesn't seem to have any difficulties with current capacity. The 5V supplies to the potentiometers and MD03 controllers are all taken from the Experimenter board. I am currently using two very old 12V 38AmpH Lead-Acid deep cycle batteries from previous electric vehicle projects to power the drive motors but a mains powered 24V DC power supply with a capacity of about 25 Amps might do - check the MD03 documentation or manufacturer for compatibility.

The main elements of the system are mounted on plywood and secured to ensure they don't move about, as are the cables which are secured with cable ties. The breadboard cabling is susceptible to loosening if other cables etc are dragged over it (eg as the motion cockpit moves) so make sure the Signal Processor Unit is protected with a cover. Try to keep the LEDs on the Experimenter board visible though - they flash each time a new position demand is processed and give a very useful visual indication that the system is running and how fast it is processing data.

So, that's the drive system. If you were interested in building a motion cockpit a good place to start, and one which won't break the bank instantly, might be to firstly build and get running the Signal Processor Unit and then later tie it to speed controllers and drive motors. The Motion Drive software and PICAXE chip programming are avaiable from this site, and the PICAXE Experimenter kit and chips are fairly inexpensive. Note also that you don't have to use all the outputs - you could use just the Pitch and Roll DOF drives for a 2 rather than 3 DOF system......

Question! How can the system be modified to provide a 6-DOF drive? I think probably quite straightforwardly - but that's a project for another day!

 

© This site is copyrighted, If you'd like more information or have any comments please contact me at