Prototype Direct Drive Telescope Motor
 

Prototype Direct Drive Telescope Project Goals

This project is construction of a prototype direct drive telescope motor (brushless axial flux motor) and a PC based control system. The project goals are:
  1. Construct the motor and control system from ordinary, inexpensive components
  2. Develop the programming for a PC (laptop) based control system 
  3. Evaluate the prototype motor and determine if it can be scaled up to control a small to medium sized telescope mount
  4. Find a viable alternative to purchasing an expensive encoder (to reduce project costs)
  5. Keep the entire project costs for the final scaled up system to less than $300-$400 USD (dependent upon the encoder costs)

Background

Direct drive telescope mounts are not a new idea, and are commonly found on professional observatory telescopes and very expensive high-end telescope mounts. Unfortunately the price of a direct drive telescope mount is far outside the budget of most amateur astronomers (on the order of $15,000 USD or more). This project is not to reinvent something that already exists, but rather to find an inexpensive alternative. The project goal is to construct a direct drive telescope motor and control system that can be scaled up to a small to midsized telescope mount for less than $300-$400 USD. The system will be constructed from basic electronic components and common materials.

The greatest challenge to keeping the total scaled up system cost at less than $300-$400 USD is finding an inexpensive, but accurate encoder. I constructed the motor control system for less than $80 USD and the scaled up motor will probably cost several hundred dollars for larger magnets, motor wire, hardware, and misc. components. The actual costs for the motor control system were on the order of $20 USD because I already had an I/O board from another project; the I/O board contributed about $60 USD to the total cost, everything else was only $20 USD for electric surplus components. The prototype motor cost about $20 USD to construct: $5 USD for a spool of 24 gauge magnet wire and $10 USD for 20 small Neodymium magnets. I have recently replaced the PC based control system with a programmable microprocessor, further reducing costs (the $60 USD I/O board was replaced by an $8 USD microprocessor).

Prototype Axial Flux Motor 

The prototype direct drive motor is an axial flux motor. This is the type of motor often found in homemade wind generators where the permanent magnets and coils are arranged around the perimeter of two closely separated plates. The axial flux motor is a brushless motor with permanent magnets fixed to a rotating top plate. The motor coils are fixed to a non-rotating bottom plate. Since the motor coils don't move as in an ordinary brush type motor, there isn't the need for brushes to supply power to rotating coils. An electronic control system supplies power to the different motor coils to cause rotation of the top plate with the permanent magnets. The below left diagram shows the axial flux motor components: the non-rotating bottom plate with the motor coils (green) and the rotating top plate with permanent magnets arranged with alternating polarity (red = +, blue = -). The below center and right photos show the prototype motor bottom plate with the coils and assembled with the top plate and permanent magnets, respectively. The rotating top plate contains 20 Neodymium magnets, epoxied to a 2mm thick polycarbonate plate. The non-rotating bottom plate contains 15 coils of 24 gauge magnet wire, connected to give 3 phases (5 coils/phase). Note that the prototype motor top plate is constructed from clear polycarbonate to give better visualization of the magnet positions. I will replace the polycarbonate and wood with metal in the final motor to better utilize the full magnetic field.

Hall effect sensors are placed between three adjacent motor phase coils to provide the commutation logic (the equivalent of placement at 120 mechanical degrees separation). The Hall effect sensor feedback gives trapezoidal commutation, where there are 6 different positions (Hall states) per 360 electrical degrees that determine the rotor position. The 6 Hall effect sensor states determine which phases are used and the current direction through each phase.
Current is driven through two phases while the third phase remains open. 

   

Motor Control System Prototype 1: PC Control via Relays

The first prototype control system (excluding the laptop) is shown below. Because I already had an I/O box from another project, the control system and motor cost less than $50 USD to construct (excluding scrap materials). Everything looks messy, but this is just because of the solderless prototype boards and temporary connections. All of the components can easily be loaded into a single electrical box with a few multilead wires and a USB cable. The control system is actually pretty simple; the PC interprets the Hall effect sensor logic and activates the appropriate motor phases based on the permanent magnet position. I have written a simple test program for motor control (below right screen-shot). The program gives real time displays of active Hall effect sensors and motor phases, and logs other relevant test data. The program allows user specification of the commutation logic, motor direction, and a series of timer and delay loops control the pulse-width modulation and time betwen pulses. 
  


                

I have tested the motor with generally positive results. The motor and control program function without any significant problems (within the limitations imposed by the system components). I have determined that several system modifications are required to reduce the motor speed to directly control a telescope. Currently there is a minimum deliverable pulse width modulation, most likely due to limitations in the hardware and software:
  • I constructed the control system using extra electrical components (mechanical relays) I had from other projects. These components may not be able to respond fast enough to deliver the required pulse width modulation. 
  • Solution: I will replace the switching components with high speed transistors.
  • The motor control program is running in a Windows environment. The Windows operating system is also constantly performing other functions and this could affect the maximum speed that the program can execute commands.
  • Solution: I will switch the motor control program from a PC based system to a programmable microprocessor with a quartz crystal oscillator. This should allow faster execution of commands to give the required pulse width modulation.
An advantage of switching to a microprocessor based system is that it further reduces the control system price: the $60 USD I/O board can be replaced with an $8 USD microprocessor.

Motor Control System Prototype 2: Microprocessor Control via Transistors

April 2014 Update:
It took me several months to learn how to program the microprocessor: I/O operations, delays and interrupts, LCDs, and AD conversion. Transferring the PC control system to the microprocessor was surprisingly easy and the coding was less complicated than the PC based system. The below photo shows the microprocessor and transistor based control system. I wrote a first test program for one direction rotation with fixed pulse-width modulation, and everything functioned satisfactorily. The microprocessor and transistors gave microsecond order pulse-width modulation; this is a significant improvement over the PC-relay based system (millisecond order pulse-width modulation). The next jobs are to modify the control program to display motor parameters on the LCD, accept speed and direction input via the keypad, allow rotation in both directions, and to improve timing by adding a quartz crystal oscillator. It still remains to find an appropriate encoder to give position feedback control to the microprocessor.

 

All original graphics, photographs, content, and writings are copywrited © 2007-2015 by the author and all rights are reserved. Do not copy or reproduce in any form without prior written consent.