Absolute Linear feedback


Thread Starter

C Seffens

I have a project in which the customer desires to use a linear motor to drive a pick and place unit over 16 feet of travel. The normal position feedback device is a linear incremental tape scale ala Renishaw / RSF. They want absolute feedback to eliminate homing AND true positional accuracy of .002" over the 16 feet of travel. Speed is moderate, peak of 2 meter/sec, accel of 1 g. I have used Temposonics in past and but am concerned about it's accuracy over this distance. Can anyone suggest another technology / brand to investigate? I need to be able to provide an incremental pulse stream to the drive for velocity feedback and am open to dual looping with tape scale and absolute device.

Peter Nachtwey

Why not use a SSI, synchronous serial interface? Stegmann makes multi encoders and Tempsonic and Balluff make the MDT rods with which you are already familiar. All can return 24 bits of absolute position. Balluff claims that they linearize their rods but I don't know if they have rods that long.
SIKO has a magnetic tape system called MAGLINE that comes in both incremental and absolute flavors. It doesn't have the resolution of the magnetostrictive devices but it can handle the distance. Ed Schmidt Associated Technologies, Inc.
One thing to consider with the magnetorestrictive devices is that the update times proportional to signal transmission distance . . . i.e. long rod = long update times. These are typically 9usec / inch. As such if you go beyond 100 inches you start getting some pretty slow update rates / data delays that are going to severely limit your ability to close a position loop . . . Ken Brown

Alex Ruderman

Tuesday February 06, @03:33AM, by Alex Ruderman

I suggest to end up with incremental feedback and Smart Position Initialization.

Conventional technique for AC (sinusoidal) brushless initialization is to use Hall sensors to start as DC (trapezoidal) brushless, find zero encoder pulse (absolute position) and then switch to sin usoidal operation.

In the absence of position Hall sensors, one can bring the rotor to predetermined position by feeding two motor phases. This random "magnetic alignment" rotor movement (180 el.deg. in worst case) is unacceptable for many applications hence the need for smart position initialization that eliminates long trial movement for an arbitrary rotor position.

I wrote a report about such a smart position initialization (too long to be published overhere). The report comprises 5 pages of text, 4 drawings and 6 equations with the following contents:
1. Introduction
2. 3-Phase Sinusoidal Brushless Basics (Torque Generation, Current Loop etc)
3. Smart Initialization Principle
4. Rotor Movement Equation
5. Control Algorithm Convergence
6. Conclusions

Contact me for details.
[email protected]


Alex Ruderman

I prepared revised 10-pages Smart Initialization report that contains 7 figures and 10 formulas.

Smart initialization, or soft start-up, procedure that eliminates a long trial movement for an arbitrary initial rotor position. After appropriate tuning, initialization trial movement is typically as short as just a few encoder counts.

Position initialization error ideally equals zero. We discuss how to reduce net load, Coulomb friction and cogging torque negative impact on position initialization error.

[email protected]