Tactile 2D + scanner idea....

Electronics and electrotechnical stuff
Post Reply
Downunder35m
Posts: 401
Joined: Sun Aug 16, 2020 5:32 am
Location: Australia
Contact:

Tactile 2D + scanner idea....

Post by Downunder35m »

Every now and then I need a thing 3D scanned to make a proper model, especially if shapes are "fluid" so to say.
I usually take my cam, rotating plaform and then painfully stich and clean up the model until it is acceptable.
Quite time consuming....

Most of the time I use thse scans to create models to be printed.
And not always you actually need micron accuracy or details.
I thought that there might be a compromise possible to avoid the endless hours of creating a 3D model from images or laser scans.....

In milling machines we often find the neat feature to "tap" a to be machined part to reference the outlines and true dimensions, or just to set a reference point for the tool path.
And those things are incredibly accurate....
Right now all of it is just swirling around inmy head with nothing designed yet, so bare with me when I try to explain the idea....

If we use the basic machins of a 3D printer and insted of a hotend a probe head we should be able to fabricate a 2d plus depth scanner.
Have the object to be scanned as a block for the max outside dimensions and have it centered on the platform.
Instead of printing the machine uses the probe to first to a "rough" scan to determine the basic depth profile from dead straight ahead.
Like poking a finger down to measure the height at that point.
Depending on the height differences in the grid pattern additional scans take care of the depth details required.
For a simple square block this would result in a very exact replica as we only have to deal with straight lines and faces.
A ball shape would be subject to either approximation or as many scans to check the depth as you would expect to have layers, just because we only get the depth info, but not a true 3D shape from this sort of probing ;)
Accuracy or resolution would be relation between how often the probe goes down per mm and how straight angles and surfaces are.
With a max of 6 scans to cover all orientations a vector based model can be created.
A 3D capable probe would also allow for a "rotary" scan of an object's outline if there is no interfering structures.

I think it should even be possible to just convert a cheap laser cutter or milling machine for this and to create a modified Marlin firmware for the purpose.
Or if doubt the classic open source milling firmwares.
Of course the real limits are how far the probe can move down and how quickly we can do this without compromising on accuracy.
Such a scanning system would just be totally useless if the process would take longer than a 3D photo or laser scan from start to clean model.
If you like the general idea so far how about this to control a probe head:
A motor means we need to have something to stop the movement before the motr can damage the probe - these things only tolerate very little abuse.
But what if we don't use an expensive probe head like used on industrial CNC and milling machines? ;)
Proportional laser sensors to measure distances are not really that expensive anymore and offer a very high accuracy.
A free-fall-piston is easy to make.
As long as you have positive ressure it stays up, open the valve and clse the one for air supply and it just falls down.
Offer much more travel then using something like an electromagnet, although which works best I guess depends on the size on "build volume" or the repurposed machine.
I guess round 10cm shouldn't be too much of a problem with a coil instead of air.
Either way, we just move the piston up by whatever means, move it to next position and let it fall down, the laser sensor takes a reading and the next position is addressed in the same way.
Shoule be rather fast compared to a classic scan for basic shapes.
For the more complex ones I guess one could try to use some of the algorithms used for slicing 3D models.
You know, take a limited amount of readings and just connect the points with straight lines.
Like doing the first rough scan and either use A.I. or user input to determine how fine the probing grid should be or to specify just certain areas requiring more scans with a higher resolution.

And tinkerer out there who would like to throw in some thoughts?
Exploring the works of the old inventors, mixng them up with a modern touch.
To tinker and create means to be alive.
Bringing the long lost back means history comes alive again.
tony_arduino
Posts: 22
Joined: Fri Aug 07, 2020 5:53 am
Location: New Zealand
Contact:

Re: Tactile 2D + scanner idea....

Post by tony_arduino »

This gives me visions of a 3D copier - 2 beds, one with the object to be copied, the second with the replica being printed ...
I like re-using/re-purposing/recycling old pieces of technology and have boxes of 'useful' looking salvaged parts. One day I might actually make something original ;)
User avatar
Orngrimm
Site Admin
Posts: 277
Joined: Wed Jul 29, 2020 8:02 am
Location: Switzerland
Contact:

Re: Tactile 2D + scanner idea....

Post by Orngrimm »

maybe something like this?
Builder of stuff, creator of things, inventor of many and master of none.
Tinkerer by heart, archer by choice and electronics engineer by trade.
Downunder35m
Posts: 401
Joined: Sun Aug 16, 2020 5:32 am
Location: Australia
Contact:

Re: Tactile 2D + scanner idea....

Post by Downunder35m »

LOL I loved those as a kid!
Had a little "toy" one my grandfather made for me.
Used a standard electric drill and was rather crappy but with enough time ....

But I was more thinking in the region of these nail face frames - creating enough reference points mechanically to get a more or less accurate replica.
A 3D scanner often really struggles with certain shapes and accuracy, like when scanning some housing parts or really small stuff.
Image scanner can result in a lot of inaccuracies here as well as both systems prefer smooth and easy surfaces.
For anything else you need know how and good software.
So I thought that there might be a middle way, something still accurate enough to compete but simple enough for everyone to use.
Exploring the works of the old inventors, mixng them up with a modern touch.
To tinker and create means to be alive.
Bringing the long lost back means history comes alive again.
User avatar
Orngrimm
Site Admin
Posts: 277
Joined: Wed Jul 29, 2020 8:02 am
Location: Switzerland
Contact:

Re: Tactile 2D + scanner idea....

Post by Orngrimm »

I use meshroom from Alicevision (https://alicevision.org/) for my 3D scanning-needs. Free and the initial "workflow" already gives quite good results.
http://3d-scan.eu/ scanned once a grip for me with 0.05mm resolution with a load of normal Photos alone.
See the result @ https://www.thingiverse.com/thing:4710973
His 3D-Scanning-Rigs can be found @ the website (Links to Thingiverse)

If you fancy the code of Meshroom, go github: https://github.com/alicevision/meshroom

But... Thats not the content of the thread. Here we have a thread about tactile and 2D-scanner. Not optical and 3D.
Carry on, sorry.
Attachments
2021-01-05 14-39-03.png
2021-01-05 14-39-03.png (132.02 KiB) Viewed 663 times
Builder of stuff, creator of things, inventor of many and master of none.
Tinkerer by heart, archer by choice and electronics engineer by trade.
Downunder35m
Posts: 401
Joined: Sun Aug 16, 2020 5:32 am
Location: Australia
Contact:

Re: Tactile 2D + scanner idea....

Post by Downunder35m »

Well, there might be no use for a "tactile" scanner by the looks of it.
Maybe it catches up one day though.
Like those pneumatic 3D displays and "beds" .... ;)
Exploring the works of the old inventors, mixng them up with a modern touch.
To tinker and create means to be alive.
Bringing the long lost back means history comes alive again.
Post Reply

Who is online

Users browsing this forum: No registered users and 1 guest