Arc moves instead of linear?

[Hugh.Venables]

I'm trying to mill a concave surface of revolution whose generator is an "S" shape; starts off cylindrical at 207 m.m.dia., finishes at 40 m.m. dia. cyl. and is 94 m.m. deep. I would have thought "scallop" would have been the way to go but am surprised that MC has broken up the (logically) circular paths into thousands (really) of short straight lines. Problem is, our poor old Mazak M2 doesn't seem able to be able to cope with the rate of data and stops and starts continuously while the processor catches up. Surely this job could be done with a small number of arcs.

Any ideas?

[cadcam]

Is filter on?

And if so do you have out put arcs ?

[Allan]

Scallop toolpath would be hit and miss with producing arcs, best bet is a flowline or paralell toolpath then you can be sure that the arc lies in G18 or G19 plane. Be sure that filter is on.

[Karl Oram]

Try This Hugh,

But Dont blink you might miss it generating the toolpath :

Toolpath >> next menu >> wireframe >> revolution.

Select your "S" shape (no splines - use break arcs to splines if you do)

Select your Axis endpoint and also a tool,

at this point Hold your eylids open and say OK

Make sure your post has G18 an G19 activated to out put your arcs.

Hope this adds to your future repertoire

Regards Caledonian_Karl

[Ambassador]

Karl, nice use of the old faithful superfastaccuratewireframetoolpaths. Karl, it must be difficult being in Stirling and Calgary at the same time.

Ambassador

[ 09-29-2001: Message edited by: Ambassador ]

[Hugh.Venables]

Thanks Guys for your help. I'm on holiday for a week so will try again when I get back.

Jay: I did have filter and "create arcs" on and tried some coarse tolerances, both cut tolerance and filter tolerance, and taking care to maintain the recommended ratios of tolerances, (which I found somewhere in Help, but have forgotten and can't find again). It is >2:1 filter:cut I think. The output to .NC was still.....thousands of straight lines!

Allan: Why is scallop hit and miss with producing arcs when the sections at constant z are circular? Not sure I follow you on why flowline or parallel will be any better? Would you mind expanding a little on G18/G19 planes please. Wish I was at work to try!

Karl: I am also intrigued by your considerably split existence. I am dealing with a SAT file from Solidworks so there may be some aspects of it that are not ideal for what I need to do. The surface in question exists as two semi circular surfaces. It's as though the generator ('S' curve) was revolved through 180 twice. I am not sure if the 'S' curve is still there. I guess I could create, curves, one edge but I'm green enough to suspect that if the surface was created using a large tolerance I may create an innacurate curve. I don't quite follow your spline comment; does break arcs to splines make arcs from splines? How do I find out if G18 and G19 are activated in the post. Will the result of this be a toolpath of arcs that the machine processor can keep up with?

Ambassador: Thanks for your support of Karl's suggestion.

[Karl Oram]

Hi Hugh,

Well first of all if the customer is using Solid works your best choice would be to ask him for the solid works files in its native format. Then to read it in us File>Convert>Parasolid>.......Then change the extension type to Solidworks Prt format and proceed to read in the file.

OK if you have the curves you only need one set. and I believe the double curve is because in Solid works its outputting a start edge and end edge of a solid at 360 then cutting it back to 180 so you end up with a double curve on one side only . Just delete one of them. You will however need to draw a Center line through the datum axis.

The Toolpath Revolution will not accept splines so as correctly guessed the break command will fit arcs through the splines and either delete or hide the original spline.

Once you have done the tool path you will find some cool features offered with this method of machining also - G18 or G19 output in your code because all the revolved moves will be arcs. If you are not seeing these codes Mpmaster.pst offered on Emastercam

will do an excellent job for you.

Some of the Old Toolpaths are Still The Best.

2D Swept is a prime example.

And the split personality - Caledonian opperates in Stirling Scotland and also out of Calgary Alberta.

Regards Caledonian_Karl

[Hugh.Venables]

Karl that sounds really good. I almost wish I wasn't on holiday for another week so I could get on and try it.....almost. I think I see how it is going to work. It will ask me for the cut increments which will be laid out down the "S" curve and then make a 360 arc about the axis for each increment. Still not sure about the G18/G19 though. Is this because the 360 arcs will be spirals including the cut increment and the G18/G19 is an inclined plane? Given the scalloped effect that I am going to have anyway and the fact that this part is a water flow transition, is there any similar proceedure I could use that would result in cuts down the "S" curve incrementing as the "S" curve is rotated around the axis, that is radial cuts. Or is that what I'm already going to get? Will the Mpmaster.pst drive a Mazak M2? Thanks again, Hugh.

[markd]

G'day Hugh,

the M2 post I gave you will output G18 and G19, if it can. I use multisurf scallop toolpaths a lot and rarely see any filtering affect other than G17. If you like give me a call when your back at work.

-Mark D

[BernieT]

Hugh, the G18/G19 arcs work only in the XZ and YZ planes. If your axis of rotation isn't exactly in one of these planes you will get a million little lines for a toolpath.

[Hugh.Venables]

Hi Guys, I'm back. Thanks for your help but I'm not there yet.

BerTau: don't you mean that the axis of rotation should be normal to the G18/19 plane if the arc is in the plane?

Anyway, when I described the job in the first place I should have said that the axis of rotation is the Z axis, parallel to the machine spindle axis. Surely it is possible to machine it with 360 arcs in the G17 plane.

The drawing is oriented so the top view looks down the axis of rotation. The origin is at the top of the job in the center of the surface of rotation. The cross-section curve lies in the side c'plane. If I read HELP for wireframe/revolution I find that the system calculates in the current c'plane and transforms into the current tool plane. The system will only allow me to select the cross-section curve in side c'plane. I presume I need top tool plane. When I select the curve, the system prompts me for a second "edge of the surface". I can't think of a second edge for this job so I click "done". The system then asks for a point on the axis of revolution. I give it an endpoint on the line through the Z axis. If I set Y axis and concave in revolved parameters I get 180 degrees of the shape and it is rotated 90 degrees (how do I get a degree symbol?). The system is trying to machine it as though it is tipped on it's side. What the hell am I doing wrong? Why does it only revolve through 180? How do I control this? Why is it tipped on it's side? How do I control whether it climbs or upcuts? At least if I post it I am getting arc moves.

MarkD: Thanks for your offer. I will drop you a line.

Hugh.

[BernieT]

Hi Mark,

quote:

---

BerTau: don't you mean that the axis of rotation should be normal to the G18/19 plane if the arc is in the plane?

---

Essentially, yes. What I was trying to say was that the axis of rotation has to be in the XY plane and either vertical or horizontal(ie. 0 deg or 90 deg)

Since your axis of revolution is the Z axis, you can't use the 'revolution' toolpath. Use the 'swept 2D' instead. Construct a circle in the XY plane centered on the Z axis and touching the 'S' curve. Break this circle into 2 arcs with one breakpoint on the start of the 'S' curve. Now chain the 'S' curve as the 'across' and the circle as the 'along' and the 'intersection' is where the two touch. That should do it. Good luck.

BerTau

P.S. Sometimes you have to experiment with the 'left' - 'right' settings to get the toolpath on the correct side of the geometry.

[ 10-10-2001: Message edited by: BerTau ]

[Hugh.Venables]

Yes! 2D swept is the way to go. Thanks for your help. Now I'm not sure when you would use revolution.......

Hugh.

[BernieT]

'Revoloution' toolpath is for cylindrical shapes laying on their side. Like a coke bottle mold. :-