This is why I do not drive solids. In aerospace one rev change may not match the next rev change in the way they were exported, the way they were modeled in the native system, etc. Merge the new solid into your file and overlay it. If you leave the part in aircraft coordinates initially this becomes alot easier. I change the color of the old one and compare. Usually it is easier to see differences in wireframe for me. In my experience I prefer to create drive surfaces/wireframe so that when I merge a new model revision in, I can quickly create/modify the geometry, modify the toolpaths, and move on. HTH

.001 lengths, Wow! Thanks for the update Tony.

You could save it down a version and it should kill the toolpaths too. Or click on that dynamite looking icon in the Op manager and it will delete all the toolpaths, ops, etc. then save it. I prefer to do what Ron suggests though.

Create endpoints on the spline then analyze the distance between the points. I have had this problem before and could never figure out why it was happening. If the curve is on the surface try switching project from plane to normal or vice versa.

If you physically build an upper and lower chain with your endpoints aligned EXACTLY where you want the fanning to occur then sync by entity... swarf 5-axis it may work. If there is a .00001 deviation in the points where it fans too and from then it will not. There will be a slight lead or lag vector in that case. The only way to control this type of thing that I know of is with the construction of the original geometry. You should be able to eliminate the B and Z movements, but X and Y will move with C unless the center of whatever arc you are swinging just so happens to be on the center of rotation for the table with that gage length tool and pivot distance for B axis. Also, set your point generator/ tighten your tolerance. You WANT alot of points.

quote: Heat shrink, mill chuck, side lock, er 32 as a last resort. +1 Motor City Minion We use heat shrinks exclusively for any of our 10,000 and higher RPM machines, but the mill chucks I have used for a 10,000 machine in the past and they worked very well. I wouldn't want to run a side lock over about 8,000 personally. They may not have to be balanced, but with collet type holders at least it isn't extremely off balance. I have seen them in a side lock without a flat and I have seen them pull out in aluminum. It drove me crazy to see a flat ground on a carbide EM for aluminum, but I guess if you can't afford the holders it was better than pulling out and doing some scrappage.

How long has this technology been in use? The tolerance is a bit loose for many applications yet, but I imagine they will get better. Where's the freakin laser beams?

Neato.

Looks interesting. I wonder if this is a Mastercam specific thing or available for other CAM systems as well. Looks like they might be trying to get some of Vericut's pie too.

LoL Nice!

Splines are calculated mathematically through a set of points. Polylines(I believe) are a group of lines with an association tying them together with the corner points. They aren't really compatible. It would be more like converting a chain into a polyline if I understand correctly. I don't know that this is possible.

Well maybe you could change your stepover to something extremely high(100 or something) after you are done since you are going to have to redefine it anyway when you change the cutter the next time. Then you may get only a couple passes, but the file size would be much smaller and still preserve the other settings in the toolpath. That's all I got.

Out of curiosity what kind of toolpath are you using typically in which you find yourself with huge files?

In other words you want to force your toolpaths to be dirty.(red X'd out) The only thing I could even think of is going into the tool manager and adjust the settings for that tool ... changing something irrelevant like the number of flutes, then updating it so it X's out all the toolpaths then save it. I'm still not sure how much of the NCI data that would kill.

Hey Tony. If you try it out on the machine, how about posting some feedback? I am kinda curious as to how much improvement you get time wise by eliminating arcs in addition to the degree of which it comes out faceted.

That is what I have heard Tony, but I have not really tested it so I can't say for sure. The idea does make sense though. I think the best way to output points instead of arcs on 3 axis toolpaths is to change your control definition so it does not support arcs. I am not sure however how you control the chordal tolerance for the post when it changes it from an arc into point to point. I know the NCI file is output with arcs regardless. Maybe one of the mastercam post guru's could point you in the right direction as far as that goes. HTH

No tony, just in the multiaxis toolpaths. If you are driving a line, there is no advantage to have more than 2 points. It would only be of benefit in 3-axis when driving splines or swinging arcs in point to point instead of circular interpolation in which case your cut/filter tolerance will dictate how many points are generated/remain. If I understand the SGI correctly, the control will establish a spline within the parameters set in the control through a set of points in the gcode(your program). If you had 20 points along a line that would only require 2 then it would build a spline with allowable deviation on each point hence causing the machine motion to potentially NOT go in a straight line. Kinda counter productive in my mind. The control does not necessarily NEED more points in situations like that, it will just manage it much, much better than other controls without the feature. BTW supposedly the SGI will calculate point to point circle motion better than interpolated circle motion. Surfacing toolpaths should be the same. If you are kellering a flat surface with a small radius at the end, there is no advantage to having the extra points on the flat area of the surface(It should drive lines). The radius area would benefit with extra points, the flat area would have the potential to hinder.

I would guess it depends on how many points you have. If your cutting tolerance on a complex surface is extremely small, the machine will have to adjust the acc/dec more often. Consider going a bit under your tolerance requirement for the surface without overkilling it. I would use filter to eliminate the points where they don't have any purpose. In 3 axis cutting, there is very little need for the extra points. Once again, give it what is necessary for the tolerance you desire, using filter to dictate this. IMO the place the extra points would be a benefit is in multiaxis toolpaths in the areas where there are vector changes, that is swarfing and cutting normal to the surface. I still tend to filter instead using the point generator on vector so I only get alot of points when the toolaxis is actually changing. HTH

It usually crashes if you scream obscenities at it. Maybe that happens the other way around....

Are they similar machines? If they are... I thought you could have one machine definition with multiple posts available still, so you could select the post instead of changing out the machine def repeatedly. It's been awhile though, dunno if I am thinking of MC9.

quote: Parts that were made ON PURPOSE I was starting to get excited remembering some I made back when I was being trained to run a machine until I saw the ON PURPOSE.

Personally I would try to avoid roughing with a ballnose if at all possible simply because the surface speed of the tip of the ball varies so much relative to the tangent of the ball at the flute. That might not help you much though...

Vericut is considered the best, but it is pricey. We use the machine sim module for many, but not all, applications. It will verify your gcode to your machine and control including the finer details. If you are doing 3 axis work the full machine sim is not cost effective IMO and a base Vericut with an autodiff module would be sufficient. I tend to use Mastercam's verify for visual confirmation of toolpaths only simply because it is not based on the gcode and the detail of machine kinematics and control is fairly limited.

Try unshading it. If the wireframe representation is nice and smooth, then it is just the display settings.

If I understand correctly, try to use one WCS per program then define C/T planes selecting origin the same as the WCS origin. I usually draw an axis system on the WCS origin for each of my planes prior to defining it so I have hard geometry to refer to. If the plane is relative to the WCS origin, you should be able to right click in the clearance box, use z-position and select a feature on the part to eliminate some of the guess work. HTH