Digitizing arm options

[turmite]

Originally Posted by dirdim Definitely stick to the Immersion arm for interior car spaces. For the price, you can not beat its performance. The trickiest thing is stablizing the base of this, or any other arm digitizer. The Immersion, being small and light, is just easier. Meanwhile, the Faro line of digitizer products starts much higher than the $10K figure you note, unless you get a deal on a used system. Figure their line between $35-85K, depending on options. These are typically for much more demanding manufacturing applications. Another nice option for the Immersion concept is that you can add a non-contact laser scanner. For a relatively low price package of around $30K, you can have the contact digitizer, the laser scanner, and some decent software for modeling the data into high quality polygon models for downstream use. We show all of these products and applications on our website at www.directdimensions.com. And I attached a graphic showing the scanner package used for a small complex toy. Michael Direct Dimensions, Inc. Baltimore, MD

Hi Michael,

I think you missed the gist of this thread. It isn't about buying an arm, but about building one. For most on here, the lowest price arm available is simply out of their pay grade. So, a diy plan is developed and eventually someone comes up with a working solution and makes it available either as a low cost ready built, they sell the plans, or they give the plans away.

For myself, if I had the money, I would already own one of your arms....well not your arms, but one of the digitizing arms! Laser produces models that are just too heavy for my needs.

Mike

[Lston]

I have ruled out using a Avago Technologies laser mouse sensor because I found out it requires an expensive power calibration tool but Cypress Semiconductor has a new laser mouse sensor. The OvationONS, it's self calibrating, surface mount, smaller and even beter specs.

CYONS1001U:
No power calibration or optical alignment required.
Hardware resolution to 3200 cpi "Counts Per Inch"
Ability to maintain full resolution at speeds up to 50 ips "Inches per Second"
4-wire SPI port
Native 16-bit x and y directional reports from the sensor.
Market-leading 40 kHz positioning sampling rate.

http://download.cypress.com.edgesuit...=1228097355670

It would still be a challenge to create a design where the sensor is focused on the shaft and the circuit board might be tricky. "Any CAD guys out there?" If I went with regular rotary optical encoders, I need to know if the Microscribe uses hollow bore or shaft type encoders which make a huge difference in price. "Hollow bore approx $200 U.S. versus about $70 for shaft." Anyone Know?? U.S digital now has the best prices.

[Lston]

Here are some photo's of the microscribe. Two good quality close-ups and one not so good of the internals of the base with a additioinal circuit board added. If anyone has more please post.
"For max size, click on the Flash generated picture then click on the picture that opens up in a new window then wait for the maximize icon to appear over the picture and click on it."

[Lston]

Avnet sells single quantities of the Cypress laser sensor for $9.38 U.S. but they have no stock and a 6 weeks lead time. I emailed to find out which sensor is included in the sensor/lens bundle they sell. There are five different resoulutions listed in their data sheet.
http://avnetexpress.avnet.com/store/...oducts&x=7&y=7

[Lston]

I just heard from Avnet and the bundle they sell in single quanities does come with the 3200cpi sensor.
http://www.cypress.com/?app=search&s...ords=CYONS1001

[kevlar129bp]

Not sure if this project has been round-filed, but I thought I would ask. Working on one myself. Just lurking for some thoughts...

Thanks for comments and thoughts,

Kevlar

[TheMeatHead]

any progress with this?

[Graham S]

Not from me at least, I lost interest, possibly because I don't need one.

If I were to build one, I'd use the avago encoders, some decent bearings and a Parallax propeller microcontroller linked to a PC via USB emulating a serial port. Then some VB or python to get x,y,z from joint angles. In many ways it is a simple project, the difficulty is making it work really well and calibrating the arm etc.

Just noticed some stuff about mouse sensors, a waste of time I think.

Graham

[D.L]

I'm building an arm less precise, its a diy pantograph arm standing 90 deg with a ball bearing tip on one side and a laser mouse on the other end. It will have an x axis slide for taking contour slice intervals, its mainly for surface work, 'streaming' splines for 3d modelling. It's only accurate to about 0.2mm, still, it gives me splines instead of mesh or point clouds which saves hours of cad work.

(i really hate mesh!)

[vicen]

Interesting thread even if old (like me). I have not read the entire thread, but think I got the gist of it, and have a question.

Instead of having to put up with the accumulated error of several circular position encoders, why not track directly, and in an absolute manner, the position of, say, the head of the stylus? Then the only additional information required to determine the position in 3D space of the tip of the stylus, would be the position of the tip with respect to the head of the stylus.

Of course I have assumed that the head of the stylus is more easily trackable directly than is the tip, or I would have suggested tracking the tip directly.

One simple way to track the head would be to locate an LED there. This is just an example, as the LED would probably not provide sufficient accuracy.

I know there are various inexpensive and accurate techniques for 3D scanning (e.g. photogrammetric techniques) that would allow determination of the 3D position of the LED (or whatever point we chose as target).

Perhaps this idea has already been explored? If so I would be interested in a pointer.