Magnetic Linear Encoder – and it works :)

Posted by Ril3y on December 9, 2009

For awhile now I have been meaning to try to find a way to close the stepper motor “loop” on CNC devices. What started as an need / idea turned into reality last night. A few weeks back cruising thingiverse.com I found a post where Zach had created a magnetic linear encoder. However his design was off and needed fixing. Instead of re-inventing the wheel with a different linear magnetic chip I decided to use the same setup as Zach. I contacted him about fixing the problem with the board (the footprint was off) and in about 3 hours after emailing him he had it fixed.

I have blogged about laser etching PCB’s in the past (you can read about it here). The inherit problem with laser etching PCB’s is that is it very hard to get aligned if its a double sided board. So I modified Zach’s design to be a single sided board. Its tight in some spots but if you use the solder mask I created and have a somewhat steady hand its doable. Here are some pictures of the process and a video of it working!

The AS5306 can be found here:
http://www.austriamicrosystems.com/eng/Products/Magnetic-Encoders/Linear-Encoders/AS5306

Thingiverse entry and files here:
http://www.thingiverse.com/thing:1409

Tags: arduino, AS5306, austria micro systems, linear encoder, linear motion, makerbot, pcb

Comments (5)

hansi

December 11th, 2009 23:37

that’s cool, but why would you want a linear magnetic encoder? are there any benefits over an optical one? anyway, why not use a rotary encoder attached to the shaft of the motor?

Ril3y

December 11th, 2009 23:41

If you read the datasheet on the AS5306 you will see how accurate it is. An optical encoder is dependent on being free of debris too. A shaft mounted encoder will tell you rotations but what if you want to see how far something traveled without doing the math needed (if you are using belts / gears etc). There are many reasons why this would be a good idea. Also since this is going to go on a CNC setup debris is very much a big problem which magnetic fields are not affected by

hansi

December 12th, 2009 00:28

Ah, the debris, of course!
Now it makes perfect sense
The resolution of that chip is nice, but you can reach that with optical encoders, too.
Actually I’ve been working with a 2-axis linear stage system powered by linear motors with heidenhain encoders that offer 2µm resolution attached to them. Anyway, such a resolution without precision manufactured glass ceramic encoder strips or similar is amazing. It looks like they’re taking the varying field strengths between pole pairs into account to tell the position. All you need then is that magnetic strip, which brings me to the next question: Where’d you get that strip? I guess it has to be quite precise, too. Are they cheap and readily available?

Oh, while I’m at it.. I never really looked into the motor control circuits & software of reprap, makerbot & the like. While reading your blog post I was asking myself if they all use an open loop motor control approach?

Ril3y

December 13th, 2009 11:40

There is no closed loop system in reprap right now. Hopefully this work will help towards that. Also, the magnetic strips I got from austria micro seems to be 5″ long. I have tried also with $2 magnetic strips you get from craft stores or online and it pulses the encoder as well. From my simple understanding of the chip (not yet verified so bear with me) you need a change in poles at LEAST every 4mm for it to work to its fullest resolution. I am trying to finish up a work project and then I will do some logic analyser / scope tests on the encoder.

I would like a link to you work. That is if you have published anything. Also, one thing to note is the simplicity of the circuit. It will be fully single sided so we can flush mount the encoder directly to the sliding material. This will allow for optimization of range (meaning the ability to get the magnet flush and close to the encoder IC).

Ril3y

hansi

December 15th, 2009 21:52

I’ve come to realize that you usually won’t need a closed-loop system for steppers, if they do deliver enough torque. However, I’ve never been a big fan of steppers..

Unfortunately I can’t publish things about my work, as it’s been a research project about 3d printers and contracts restrict me from releasing anything
However, we we’re usually working with resolution of 600-1200DPI, so I am quite disappointed from the inaccuracy of reprap & co as well as other hobby projects involving linear stages. What I would like to do is to improve on the motor control and positioning front, but I did not yet decide on which project I’m going to work, as they’re all quite specialized systems leaving little room for doing things differently without having to rebuild large parts of the whole project. In other words, a reprap or makerbot is not modular enough on the mechanical side.
That’s why I am looking forward to getting some contraptor parts, as soon as a metric version is available (hello from europe). That contraptor thingy seems to be just right for building 3d positioning systems regardless of what tools you’re strapping onto it.

By the way, if you’re looking for cool microcontroller projects, have a look at the website of my local hackerspace: http://www.das-labor.org . Hopefully we’ll add some 3d printing projects soon

Do you have anything to say?

Synthetos