4th axis newcomer

[kendo]

I've never programmed anything rotary before and now people here are talking about getting a new router soon with a rotary 4th axis
They want to use it to run a non-indexable sanding wheel on a frame thats being rotated so the grain around the perimeter is continuous and follows the radii.
Except, our frames aren't circular, they're rectangular with various radii in the corners.

This doesn't seem like a very complex thing to do, but again, zero experience here.
What you're looking at below is one frame transposed at 90 degrees to itself
It measures about 35 x 18 with corner radius of about 5"

the frame will rotate about its center point while the sanding wheel goes from position 1 straight to 2, then sweep an (XY) arc from 2 to 3 as the part turns 90 degrees, then straight to position 4, then sweep an (XY) arc back to position 1 and repeat to do the whole frame.




Anyway, thats what I'm trying to do and Im just starting to envision the necessary code to make this happen: this is where I need help.

I wrote the code below and was wondering if anyone wanted to tell me where I might be going wrong.

G0 G90 G54 X15. Y8.1716
G43 H1 Z1.
G1 X11. Z0. F20.
G1 X-11.8713
G2 X2.8437 Y17.1992 R12.2071 A90.
G1 X-2.8437
G2 X11.8713 Y8.1716 R12.2071 A180.
G1 X-11.8713
G2 X2.8437 Y17.1992 R12.2071 A270.
G1 X-2.8437
G2 X11.8713 Y8.1716 R12.2071 A360.
G1 X11.
X10. Z1.
G0 G91 G28 Z0.
G28
M30

My main concern is the G2 arc moves coupled with the 4th axis move, will they work together?
Besides that, am I even close with this code?
If I'm totally off base, can someone point me to where I might learn about this?

[kendo]

Ok no help so far, but the plans have changed anyway

The local Haas apps engineer was here today and they want to sell us a smaller router and remove its spindle to install an HRT110 rotary in its place. This will have an electric sanding wheel mounted to it which will then be able to contour around the frame as it sits still on the table.
This all sounds pretty awesome to me, I was wishing they made something like this, now they're going to build it custom for us.

But, I still have the problem of entering the 4th dimension with my programming skills and I still have no idea what a 4ax program looks like.
In particular, contouring around a corner with an XY arc move while the rotary turns 90 degrees and all axes arrive simultaneously at the next arc quadrant.

For a frame that was 20 x 20 with 5" corner radius wouldn't it look like this?

X0,Y0 at center of the frame:

G1 X-5 Y10
G3 X-10 Y5 R5. A90.
G1 Y-5
G3 X-5 Y-10 R5. A180.
G1 X5
G3 X10 Y-5 R5. A270.
G1 Y5
G3 X5 Y10 R5. A360.
G1 X0

[Donkey Hotey]

I guess I'm not understanding what it is you're trying to do. That doesn't look like a rotary problem to me. Which axis is the sander in? Is it a disk or a drum? I'm still not seeing why that isn't just a planar, 2-axis move.

[Geof]

Originally Posted by Donkey Hotey I guess I'm not understanding what it is you're trying to do....

My interpretation is that they want a sanding wheel to travel around the perimeter of the shape and always keep the axis of rotation of the sanding wheel perpendicular to the direction of travel; so the sanding marks are parallel to the perimeter all the way around.

[kendo]

Geof's right on the money

sorry for being a little vague
this is a whole new realm to me

look here:
we want something like this, except the frames aren't on a conveyor and one wheel does the whole job

http://www.youtube.com/watch?v=Dj7_SFW99kA&feature=related

http://www.youtube.com/watch?v=Dj7_SFW99kA&feature=related

[Donkey Hotey]

OK, why not use a flap-sander or drum sander and run around the perimeter in a standard setup?

[kendo]

I dont follow,
whats a "standard setup" for using a cnc router to hold an electric sander to sand the perimeter of an aluminum window frame continuously and parallel all they way around?

[kendo]

the problem is, if the wheel doesn't turn through the corners on a 4th axis, the grain will come out like this:



instead of what we need:

[Geof]

Originally Posted by kendo Ok no help so far, but the plans have changed anyway The local Haas apps engineer was here today and they want to sell us a smaller router and remove its spindle to install an HRT110 rotary in its place. This will have an electric sanding wheel mounted to it which will then be able to contour around the frame as it sits still on the table....

Are they also going to rewrite the mill control software; it will not work without this. Alternatively they may be thinking about adapting software from a four axis lathe; other than that you will be face with writing a moderately complex macro or buying a $50,000 five axis CAM system

Perhaps considering I am sticking my neck out making predictions I should explain.

The rotary axis about the spindle centerline is the C axis (I think that is the correct nomenclature.)

The contact between the sanding wheel and the workpiece is, theoretically, a line as long as the width of the wheel but in fact the wheel distorts slightly so it is a rectangle.

The C axis has to pass through the centerpoint of the line/rectangle so the wheel 'scuffs' evenly as it goes around a corner and so that you use centerline programming for both clockwise and counterclockwise corners.

In X and Y the program is trivial; it is a simple profile with G01, G02 or G03.

The linear, G01, moves are trivial but the G02 and G03 are not because during these moves the C axis has to rotate and the Mills do not have any command that can combine circular interpolation with axis motion. Or at least none that I am aware of but maybe Haas does have it.

The C axis rotation is also a little bit complex because the rate (feed) needs to vary in a sinusoidal manner around the arc. In circular interpolation the two linear motion vectors have to vary in a sinusoidal manner so that their sum gives a constant angular vector for the circular motion. In a similar manner the C axis rotation accelerates toward the 45 degree point and deccelerates from there.

On a mill with standard operating software this could be done with macros or multi-axis CAM.

However, I think this motion is identical to what is needed to mill a square with radiused corners on the end of a bar in the chuck of a four axis lathe with live tooling.

So you don't need a modified gantry machine you need the world's largest four axis lathe standing on end.

You can save some money though because the live tooling will not be needed.

[Geof]

Originally Posted by kendo the problem is, if the wheel doesn't turn through the corners on a 4th axis, the grain will come out like this: instead of what we need:...

What are you building? High pressure hatch covers and you are worried about gasket extrusion? Window frames for a new SST airplane or a new Shuttle. Maybe it is the fancy new nuclear subs the Brits are building because they want big picture windows like Jules Verne's Nautilus.

[Donkey Hotey]

OK, I see the problem now. The original pictures showed a round cutter in plan view. Now I understand that we're talking about a flap sander on its end.

If it were me, I'd make a jig for a sander that has spring loaded rollers that grip the sides of the extrusion. I'd just run the sander around the frames manually. I can't imagine that sanding debris flying all over the linear guides is going to be good for that router. That's aside from all of the programming problems.

Was Haas sure that you couldn't do this through spindle orientation? I'm envisioning a jig that holds the flap sander dead-center on the axis of the machine and has a stub to install into a normal tool holder. Being on-axis, the spindle wouldn't need much holding torque. The machine could just track the tangency axis as it ran around the corners.

Now that I've tossed that out there, here's a third option: make a similar jig with an M-code fired air cylinder. The air cylinder would engage rollers like I described in the hand method. The idea is you'd orient the spindle (angle) so it is aligned with the extrusion. The sander would set down over the extrusion. You'd set an M-code for a pair of tracking rollers to grab the extrusion. From there, it would be normal two-axis moves around the part. The rollers would guide the sander. At the end of the pass, disengage the M-code to release the rollers, pick the sander up and Bob's your uncle.

[Geof]

Originally Posted by Donkey Hotey .... The machine could just track the tangency axis as it ran around the corners...

This would be a very tedious programming challenge. You would have to break the 90 degree arc up into many segments and re-orient the spindle after each segment. The feed would be stationary during the re-orientation so you would get funny witness marks at every stop.