Machsupport Forum

Mach Discussion => General Mach Discussion => Topic started by: nik on June 24, 2007, 07:06:19 PM

Title: about Z depth
Post by: nik on June 24, 2007, 07:06:19 PM
hello everyone
i'm trying to do something unusual....
for example: wood cutting router , cutting 2.5d square on flat surface - easy
but... if the surface is not exactly flat... and z depth is 1mm - big trouble
what i can do?
for now i have 2 ideas but i dont know how to...

1. using up and down inputs ( for plasma height regulation ) but this idea need very precision switches very close to the tool where wood chops can make troubles, or when changing direction
2. using scanner - first run g-code for square with scanner,scan only the square periphery (not raster) and make corrections on z axis only, then run the new g-code with tool

i think the second one is better, also the tool and scanner could be used without changing

               tool               scanner
________________________________________________
                  |                    |
                                 ******************xx

the question is: how to realize this?
Title: Re: about Z depth
Post by: Graham Waterworth on June 24, 2007, 07:48:24 PM
Why not adjust your machine so it cuts flat and square.

Or

As a last resort machine the table with the machine.

Or

If its the work piece that is not flat then pack it and clock it flat with a DTI.

Graham.
Title: Re: about Z depth
Post by: nik on June 24, 2007, 07:56:41 PM
the piece should NOT be flat
and the piece is not one, they are many, and different
Title: Re: about Z depth
Post by: Graham Waterworth on June 24, 2007, 08:09:51 PM
oh, that changes things a lot.

It could be done with the probe, but its not that easy, its getting the values into the g-code that is a pain.

The best way is to write a VB script to generate the new program.  Not for the faint hearted.

Post a picture of what you are talking about.

Graham.

Title: Re: about Z depth
Post by: nik on June 24, 2007, 08:35:49 PM
i didnt post a photo because one of the axis is rotate and will be more confusing, but there is no difference
when cutting on not flat surfaces this problem is same with or without rotating axis

this is what should be, i have to cut only the square
Title: Re: about Z depth
Post by: Jeff_Birt on June 26, 2007, 12:26:51 PM
I have to agree with one of Graham's suggestions.  Make a jig to hold each piece so that the bottom of each rectangular cut out is parallel with your table.  It's both a simple and repeatable system.
Title: Re: about Z depth
Post by: nik on June 26, 2007, 04:31:45 PM
1. the pieces are not identical and I CANT cut nothing more than the square, this mean we CANT talk for anything repeatable
2. who care about the bottom when I cut the top? and the top is not flat? I have to move z axis all the time to fallow the curve of the surface
I give some ideas, now i need help how to do this
may be like digitizing plug in but the moving is not on raster, moving to be form g-code
Title: Re: about Z depth
Post by: Jeff_Birt on June 26, 2007, 04:48:40 PM
Are you cutting the rectanglular groove around the leg's top? And, you are saying that the your have several similar but unique parts.  How many diffrent styles 10, 100, 1000?  If it is a limited number that you will be doing several pieces of, I would still build jigs for each one.  I suggested referencing the bottom, flat section of each leg top as it should be a repeatable known reference to the top surface you are trying to follow.  With such a known reference, the contour you are cutting on the top will be the same for each style of leg.  If EVERY part is unique it is a more difficult problem as you would have to mount each leg, digitize the surface, convert the point cloud data to surfaces, somehow dynamically generate G-code to match the surfaces.  Or, write some sort of Mach wizard that will let you touch of a certain number of points and from there it will generate the g-code for the part.
Title: Re: about Z depth
Post by: nik on June 26, 2007, 05:23:21 PM
If EVERY part is unique it is a more difficult problem as you would have to mount each leg, digitize the surface, convert the point cloud data to surfaces, somehow dynamically generate G-code to match the surfaces.  Or, write some sort of Mach wizard that will let you touch of a certain number of points and from there it will generate the g-code for the part.

finally :)
what we do now?
i talk about 100-1000 pieces, until now i sort them in 5-10-20 piles , and for each pile manually measure Z and write a new g-code, is not long - 30-40 rows, but is not the fastest way, and the quality is not the best, because even assorted they are still different
Title: Re: about Z depth
Post by: stirling on June 27, 2007, 01:21:57 PM
This is surely exactly the same problem as faced by engravers and printed circuit board makers. How to follow an irregular surface whilst maintaining precise cut depth. Sounds like a job for a floating head to me.
Title: Re: about Z depth
Post by: jimpinder on July 02, 2007, 03:19:49 PM
The only thing I can think of is to plot the co-ordinates of the area to be cut with a plotter, transfer these to your cnc and reduce the height by whatever you want the depth of cut. I did see a plotting machine at an engineering exhibition a few months ago so they are out there, but what they cost, I have no idea. I would imagine the programs will be compatible with G code, however.

You are probably doing this manually at the moment - hence you say, the result are a little imprecise. The ones I saw were accurate to a thousand of an inch, and could be programmed to make as many plots as you wanted within a given area.

Jim Pinder