Is there a practical method of “scanning” a large surface?

[steve_g]

We are often asked to V-carve text or patterns on boards that are twisted or warped to badly to be practical to surface them flat first. Recently a nice pecan plank 2” X 12” X 60” had a bow of about 1.5” in its length. It would have worked beautifully for the customer’s needs if I could have V-carved his message on the existing surface.

Is there a practical method of “scanning” a large surface into VCP or Aspire so I can project the toolpath unto its 3D surface? Ideally, I’m thinking something I could chuck in the collet and “drive” around the board getting Z points. Certainly the ShopBot 3D Digitizing Probe would seem to me to be too slow.

Any ideas out there?
Thanks!
SG

[myxpykalix]

Well you could use the probe and set the distance per point further apart then if you were scanning a 3d object (like a bust of someone). but i'm sure that what you would end up with is straight lines between points and what you would have ultimately is a faceted object.

[Burkhardt]

I know it is not going to help you that much...but to explain what is possible: My Planet-CNC controller has a "warp" function. You can scan a surface of any size with a touch probe (for your purpose a scan point every 4 inch is probably sufficient) and then "warp" your gcode file to match this surface as "zero" plane. This was originally intended for machining high precision printed circuit board traces on a copper clad blank that is not entirely flat but it would do what you want. I have used this function only as an experiment so far. Pretty cool but not what I need every day. Below my test to project the vcarve path of that piranha on a wildly curved piece of wood.

Here is a video of the process:

https://www.youtube.com/watch?v=SAP5LOocDcQ

27679

[knight_toolworks]

What I do is make a simple jig so I can screw it to the table. the only trick is getting it flat. get someone to stand on it then just a couple blocks of plywood on the edges to screw it down. I can show you how if interested.

[genek]

We are often asked to V-carve text or patterns on boards that are twisted or warped to badly to be practical to surface them flat first. Recently a nice pecan plank 2” X 12” X 60” had a bow of about 1.5” in its length. It would have worked beautifully for the customer’s needs if I could have V-carved his message on the existing surface.

Is there a practical method of “scanning” a large surface into VCP or Aspire so I can project the toolpath unto its 3D surface? Ideally, I’m thinking something I could chuck in the collet and “drive” around the board getting Z points. Certainly the ShopBot 3D Digitizing Probe would seem to me to be too slow.

Any ideas out there?
Thanks!
SG


What I do is rip the board flip the ends and re-glue the baord back together. Then plan the sides by 1/16 to 1/8 Back side doens not have to be perfect but needs to be sojewaht level.

[myxpykalix]

Because the probe has a limited up and down travel inside of the unit, i wonder if you could fashion something with a deeper travel then be able to slowly drag a tip across the surface that would plot the hill and valley of your warp.
Then take your profile of that surface and in your program lower your cut to only cut your high spots. Flip it over and repeat.
The trick would be writing the code to translate the lateral movements instead of thousands of up and down "pecks"
Does that make sense?

[bleeth]

Steve:

A bowed board should be doable by securing it to the table and then taking some measurements of the bow to create the arc or curve that you could use as as a control for a two rail sweep in modelling.
Could likely do the same thing with a twisted board by combining a few curves along a sweep.

Dave

[steve_g]

Thanks for your responses!
The boards we’re usually asked to V-carve are often bench backs made by someone else. The boards are dry, hard and thick. Flattening them by any method is impractical! I think the bot would bow if we seriously tried to hold them flat!

Positing my query and the answers I’ve received convince me I haven’t overlooked anything obvious! We do these wholesale for $70.00 so I can’t spend much time developing the 3D surface. Right now we surface them flat or turn down the job if the board isn’t suitable.

The attached image is my idea of how I see something that could be fast enough to “drive” around the perimeter of the board taking coordinates in less than two minutes. I’ve asked an engineering firm to look into using a resolver or possibly an accelerometer to achieve this.

Just for grins… If the price was right, does anyone else see a use for something like this?
SG

[myxpykalix]

Steve this was what i was trying to explain was something that would draw a line in a profile to basically map the surface. By doing that you could target the high spots rather then just a wholesale pocketing of the entire surface wasting a lot of time and material. Somehow your probe needs to write to a cad file the profile of the curves then be able to lower the cutting depth to the bottom of that "hill" (using hills and valleys analogy)

[dmidkiff]

Just a thought but could you use a plotter pen to draw lines all across the surface and then reverse engineer the sbp? It's been a while since I did that process but you would have to remove the "non cutting" vectors. Like I said just a thought.

[willnewton]

The fastest way I can imagine to do this is to take a photo along the long axis of the board, import that into your 3D software, and then bend the type layout to match the curve in the photo. It is very quick and dirty, but may be effective enough.

I have a digitizing probe and you could use a large grid (maybe 2"x2")to get a basic representation of a surface quickly, but could probably not hit your $70 price point for a one-off design, due to the time required for the 3D scan.

A third possibility would be a 2d scan. In a similar fashion to the photo method, which would give you a rough guide of the curve on a single axis, you can do a single pass probe scan down the board. Instead of a grid, you would probe a straight line of sample points down the long axis of the board, generating a curved line the shape of the board.

[ADAPT_Shop-Brady]

We just got a 3d scanner in our shop, it's a $500 upfront investment, but would make quick work of your problem.
http://skanect.occipital.com/download/

[GeneMpls]

I would love to hear how this is working for you and see an example. Thanks Gene

[tri4sale]

We just got a 3d scanner in our shop, it's a $500 upfront investment, but would make quick work of your problem.
http://skanect.occipital.com/download/

interesting, especially the support for the Kinect for X-Box 360, you can pick those up cheap, so easy to try out the free version and see if it'll work for your needs.

[bobmoore]

At the prices you are talking about I personally don't see a high tech solution being too practical. You might try breaking your vcarve into smaller sections and level and carve in a few smaller setups. Bob

[steve_g]

I’m with Gene… I’d like to see what others have done! $500 would be a cheap solution!
SG

[ADAPT_Shop-Brady]

So just for a simulation of what you guys are talking about, Here is a scan of a floor mat with my computer bag under it to give it some "warp".

https://skfb.ly/PTGB

I can export this as an stl file which could be imported to your design software.

[GeneMpls]

Thanks Adapt- even picts would be helpful. Thanks Gene

[Tom-3344]

Adapt_Shop-Brady very interested in the scanner STL. The structure sensor doesn't quite say it can give you measurable 3D models. Can both the sensor and software yield a measurable model to reference at your fingertips, and thus useful?

[ADAPT_Shop-Brady]

Adapt_Shop-Brady very interested in the scanner STL. The structure sensor doesn't quite say it can give you measurable 3D models. Can both the sensor and software yield a measurable model to reference at your fingertips, and thus useful?

We use ours to scan bead filled vacuum formed cushions that we then carve into foam on our bot.

We use these for children with disabilities (make seat and back cushions for their seating) and I can attest to the accuracy being REALLY good. Might not be perfect, but our kids can be pretty unique and the form/scan/carve/fit process works for us.