Home‎ > ‎Router‎ > ‎

$5 digitizing probe

As the workshop is not built yet I am doing the work in the house. As I don't want to fire it up in the house and spend the text week vacuuming the house I'm doing less messy things.

One of the things i am keen to do is work with very organic shapes and being able to scan in objects. The $5 digitizing probe was a quick and dirty hack to just see what could be done with the things I had at hand. This was plan B, plan A was to buy a digitizing probe on EBay. Then I saw how much a real probe cost!

So is what I made instead. A darning needle glued to a microswitch. Actually I headed up the needle and melted it into the microswitch button. To stabilize the needle it passes through a piece of plastic cut from a milk carton.

So the $5 question is how well does it perform? I tested it by scanning the rear of an Australian 5c coin.

This is the test object. It is just big enough to fit a 10mm by 10mm scanning area.

I used the example gcode program gridprobe to scan a 10mm by 10mm square at intervals of 0.1mm.
This was the result brought into meshlab.

How to configure gridprobe and bring the data into meshlab is covered in the gridprobe project

After I did a back of the envelope calculation as to how long it would take to scan 10000 points I set the probe speed to 320mm/min as that would lake about an hour.

The result above was pretty impressive considering what was used to make it.

I did some repeatability tests, getting the probe to hit the same spot 100 times at different speeds, 320mm/s as in the scan above and then slowed right down to 30mm/min.
The graph below shows the results. Clearly slower is better and by quite a lot.

This is the same coin but scanned at 30mm/min. It took a long time, about 10 hours.

For a quick and dirty hack the results are amazing. The probe does have some limitations. It is really slow and it only comes down from above and slide off objects when the angle between the object and the needle is shallow.

I wont be doing any more with the $5 probe. With bit more complexity we can make a very cheap probe that registers touch both from below and from the side. With this type of probe a much more sophisticated scanning process can be used and slipping off the object is not an issue. This is the subject of better touch probe.