A new approach !?

[raj]

I am trying out adifferent approach to handling data for the shop bot. Thought I would try to see if anyone has tried or worked on something similar.
I am building a front end user interface that can scan shop bot files and display pertinent information about what's in it such as size in xy and max z etc. I have been scaning a lot of detailed ornaments with the probe. Although it takes long it is infinitely faster than drawing the same part in a cad program and it happen while I sleep! Anyway the aim is to have a graphical front end that allows the ornaments to be trimmed, sized rotated, sized and placed on a final product much like tiling a floor. The software is then supposed to create a full blown shop bot file that has put all the relevant pieces of code together. To keep my graphical interface simple I'm just planning on using coloured rectangles for the different ornaments with names so as to communicate what has been done. The other thing I am trying to include is scanning that gives two scan files one for a rough cut and one for a finer bit. I also did some changes to the scanning software that has speeded things up a bit by removing some of the redundancy of scanning points on flat surfaces at the bottom of the probe. Anyone alse working on this kind of problem that wants to collaborate or share to build a interface that will be a quantum leap forward for folks who want to do designs with scanning as their main input rather that cad drawings, ie a integrated front end for a new approach to the end ?

[Ted Hall, ShopBot Tools]

Hi Raj,

Sounds interesting! I'm sure Bill Young has probably seen your posting here. But he will definately have an interest in what you are up to as he has struggled with similar projects.

FYI, we expect to put the next beta of the new software on the website on Monday. It has a number of new features that will help integrate with what you are up to. Especially in terms of making it easier to easier to use a 'front end' and in terms of providing more file info.

There won't be a lot of instructions with the beta, but if you look at some of the examples you'll get the idea, I'm sure.

[valensign]

I would love to help you But I have no clue about Writing Programs. I like your Idea.. Currently I use a probe and then import the point cloud into rhino. I just got a plugin for rhino that creates a mesh now from the Point cloud so that should make things easier I hope.. Havent had a chance to try a large file yet. Then I export the 3d model into Artcam and create my cutting files from there.. Your approch would make things easier and quicker. Would you include the abblity to clean up the model or would we still need a 3d program... If I can be any assistance feel free to email me wish I was more help.

[raj]

CAD programs are great when you want to design or alter significantly a scanned part but if you have a good original scan why spend all the effort at trying to end up with a tool path that is the same as you started with (for the scanned area in question. One design of a frieze which I had was four feet by about a foot and it took a whole day to scan it and if I were to try to take it into rhino just so I could add the other 4 feet I want and to allign it and trim it, it would be time consuming. Trimming and alligning should take me a few seconds because it is a fairly simple task. Then if I want to complete the product with a few more ornaments in different locations they are simple trim, place , rotate, size type operation. Should take seconds and again I think CAD tools like Rhino is like trying to crack a nut with a sledgehammer ! But to really avoid the tool pathing route via CAD tool plus layout tool means a radical change at the probing end. What I am trying for is two files to be generated when probing : one for roughing out and another for the finer cut. The paths would have to be generated for the two bits you are going to use. The algorithm can easily be speeded up several times with a little bit of change so that much smaller files are produced quicker. One small change I did already dramaticall chaged the size and speed of probing by scanning sideways as if skimming when the probe is at the bottom (it can't go further down). This produces start and end points over those flat areas much like artcam would instead of thousands of points.
If there was a probe that could look ahead a few steps, boy that would be nice! But you can be clever by looking at your previous raster as a guide to the type of terrain you are dealing with and change step size accordingly. Use statistical analysis. Somehow get accurate smaller files, faster. I'd like to do all this stuff in Lisp because that is what I am more familiar with, but don't exactly know if I could control the shop bot probe through lisp through some mechanism that would allow me to talk to shop bot?! Maybe bill knows!
I downloaded a great set of development tools free from xanalysp and am playing with them at the momentto see if I can manipulate the bot files and build a front end interface.

[bill.young]

Hi Raj,

It's great to see someone working on speeding up the probing software. The Windows software will have a new Virtual Tool ( called the "Copy Machine" at the moment) that combines all the probing and edgefinding routines and gives then an easier to use interface, but it still uses the existing routines with a few added features (like being able to pick the starting point of the scan with the keyboard control).

As far as using LISP, I don't know much about it except as an AutoCAD scripting language, but if it has the ability to do text and string manipulation then you should be able to use it to pass information to the new Windows software and get info back. How much of the actual probe control you can do within LISP, though, I really don't know.

FYI, one of the sessions at the Jamboree will be on Part Files and Programming. The bulk of it will be on programming basics like using variables and comments, but we'll try to talk a little about the programming features of the new Windows software and ways to interact with it.

Hope to see you there,
Bill

[raj]

Thanks Bill,
I am looking forward to looking at the new release. One thought I had for the probing process was to probe pertinent data and leave the creation of shop bot file or DXF file or somethig else as a post processor operation. Perhaps a data format that defined each raster line, also partial rasters in that line which were the flat bottom parts (if you know what I mean). Then we would end up with a data structure that would be easy to operate on and produce whatever was needed for the cut file. One file say for the ornamental detail, another for the flat bottom. In practical terms ornaments tend to often fit into recesessed areas. For example if you had a square recess and wanted to put an ornamental rose into it, you want nice edges for the square recess and want the rose centred at cut out without ruining any of the square edge. Having a data structure type file allows an easier access to the pertinent data which then gets reformatted to do the cuts. In this example of the rose the operation needed would be to trim the scan, to a size and place it in a recess (move in z), and then give a final cut on the square to get a fine edge. At other times it may be necessary to cut only ornaments without the bottom flat parts because another ornament overlaps into that area and so you dont want the area cleared. Storing the probed data in a data structure I think will allow more ease of access and will make the job easier for each feature that one wants to add. Even interfacing to CAD vendors format may become easier.
Does ayone know if using the existing software it would be possible to store the previous raster line as a reference to making decisions on step size ? How hard would it be to reference nearby-X data for Z value. The language seems fairly simple but for some reason little changes gave me a hard time and I spent my life glued to the panic button incase I lost my precious probe by being mangled. So far survived !