Linear position sensor

[Toby Catlin]

Hello Everyone,

I was hoping to pick your collective brains about an idea i had. 

I am working with two friends on a shapeoko cnc kit (hopefully the very start of a norwich hackspace) and we were chatting about adding a position feedback mechanism. It obviously has to be accurate and reliable, at least as much as the steppers. Cheap is also important. 

We were talking about sticking a grid pattern along X & Y axis and using a mouse to measure the progress of the gantry. After some googling it seems this is good in theory but hard in practice.

So that lead us to a webcam + scale markings + opencv to measure where we were and I thought seemed unreliable. I thought about sticking some magnetically encoded tape (video or cassette tape ) and then attaching a read head to the gantry. I could record a series of pulses on the tape and count them back at the gantry moved. 

Before i spend time testing this out does anyone have any opinion on the practicality of this idea?

thanks

toby

[Nick Johnson]

What are you trying to achieve? Zeroing at the limits and counting steps is generally considered perfectly reliable, as long as you don't skip steps - in which case you're in trouble anyway.

[Toby Catlin]

Better accuracy & reliability I suppose. We haven't got the thing moving yet so it may well be totally unnecessary and I am prematurely optimising.

[Nigel Worsley]

The proper way to do it is with one of these:
www.ebay.co.uk/itm/230735255966

Although not mentioned in that listing, these have a data port where the
measurement can be extracted, they continuously spit out the data in
format very similar to SPI.

A bit expensive, especially as you probably won't need them.

Nigle

[Toby Catlin]

That looks perfect and the price is not ridiculous at all.

I love having access to a hive mind. Not only do i have a better solution, I also probably don't even need it which is the best solution of all.

thanks

toby

[Adrian Godwin]

They have quite a slow update rate, though. Some can be pushed into a higher speed, but that's possibly only the report rate not the measurement rate. It may be possible to move them fast enough that they lose position - they're not an absolute measurement (even when called that) but I'm unsure of the limits.

If they're fast enough, there are 150mm calipers often going cheap at Aldi etc. These use the same position encoders as the machine-mounted scales and may have the data port inside.

There's a company called US Digital which makes accurate magnetic scales as you describe yet still at prices below the precision glass scales (also available from the company Nigle linked). A Yahoo group called cad_cam_edm_dro is a good place to get advice about them.

I've also wondered about using optical (pattern sensing) mice to measure relative position. I'm not sure of linearity and it may be difficult to keep them sufficiently clean on a CNC machine, but they're very cheap and don't require a mechanical link to the surface.

[Simon Howes]

Just chuck an encoder on the other side of the stepper and be done with  it.

Some cnc programs like mach3 actually support this and will restart/realign the machine if your stepper  stall.