[jrmach]

Heard it before,not sure I understand it entirely.
Why would I and when would I want it over the old tool path choices ?
Need some videos guys

[mmoe]

This feature is what I was most looking forward to, but ended up not really quite what I hoped for in V27 for the 3 axis Mill Pro. I'm guessing that you've discovered the option in Equidistant Offset settings.

In a perfect world, it would be so that you can run the toolpath along the drive curves, so that you can control the "flow" (for lack of a better word) of the toolpath along a direction that may be less than straight. The feature, as it stands, in Equidistant Offset still works more like a standard equidistant offset toolpath, where the steps are evenly spaced across the 3d surface, but generally follow the drive curves. I was hoping that equidistant offset would have some sort of adaptive adjustment similar to the new Adaptive Planar Slice and Adaptive Z-level where the steps vary in width, but they stuck with truly equal distant spacing even when using drive curves (I think this is a mistake personally). A more adaptive type of spacing feature is already available, but you have to pay for at least the 4 axis Pro package to use it, even if you ony have a 3 axis machine, which I think is just a bit more than I'm willing to pay for that one strategy. If you have the demo version, you can go to the "Multi-axis" toolpath strategies and use "Morph Between 2 Curves" to try the one I like and compare to the 3 Axis Pro strategies. This is what I consider the more favorable way to use drive curves and really what I was hoping that the new feature of Equidistant Offset would do, where it basically would evenly space toolpaths between 2 drive curves and follow them around along the 3d surface.

I've attached a BBCD file from the V27 Demo version that has a basic example of where you'd use a drive curve. You can see the selected drive curves in both the Equidistant Offset toolpath and the Multiaxis toolpath. I also threw in a Advanced Planar toolpath just for comparison, as I think this is closer to what I had hoped would be in 3 Axis Mill Pro, but I really do with they had made the Equidistant have an option to work like the Multiaxis toolpath, which basically looks like it is running along the surface normals with a ball end mill, even though you're only using 3 axis motion. It's a bit of a shame since everything else about V27 is really top shelf. From the CAD, to the new CAM additions to the "zoom to mouse" feature (which I've wanted in Bobcad for a decade now!). If they had a 3 axis morph between curves feature, 3 Axis Mill Pro would be hands down the best CAD/CAM system for under $7500, IMHO. What they provided with the Adaptive Z-Level and the Adaptive Planar slice is still very impressive though, and really provides functionality that I don't know if I've seen in any other 3 axis systems.

Big kudos to Bobcad on this version, which seems to be reasonably stable on my system, especially for a first release of such a radical shift in technology. I've bogged the CAD portion down to a halt a couple times now, but I was pushing it to see what it's limits are on purpose. The same actions bog down most other CAD systems I've used, so it really was not anything I found troubling at all. I would recommend that users be very patient when doing demanding model work. Even if it seems to be locked up, it is often still thinking on some of the more complex geometry calcs.


EDIT:
The zoom to mouse feature really is awesome. To be completely honest, I'd almost upgrade just for that alone. I've not been happy about the mouse zoom feature ever since the dynamic zoom disappeared back at around V22. Used to be that you could dynamic zoom by holding down CTRL, and I did not like when that went away. Glad to have it back, and even better that it doesn't require an additional key to work!

[jrmach]

Wow,thanks for your time.It's going to take awhile to learn all those new paths.

Member Forrest had a thread on here a week or so ago.Does BoB have a better solution to his problem now?I am interested in what you have to say too,as I have been there more than once or twice wishing there was a better way.
Here is the thread http://www.cnczone.com/forums/bobcad...r-surface.html

[mmoe]

Wow,thanks for your time.It's going to take awhile to learn all those new paths.

Member Forrest had a thread on here a week or so ago.Does BoB have a better solution to his problem now?I am interested in what you have to say too,as I have been there more than once or twice wishing there was a better way.
Here is the thread http://www.cnczone.com/forums/bobcad...r-surface.html

The new "Advanced Slice Planar" feature would be ideal, just from a quick skim of the thread, though Equidistant Offset with the drive curves would also work (more controls for the straight back and forth in Advanced Slice Planar I think). The drive curves function of the Equidistant Offset would work well on geometry that has parallel curves (concentric maybe the better term?). For example, if you were cutting the trim that goes over an arched door, the inner and outer curves would be equal in distance from each other, so using them as drive curves in Equidistant Offset would probably work very well. When the curves are non-parallel, such as the example I included below, Equidistant Offset produces a chevron toolpath in the wider portions of the surface between the drive curves. IMHO, it would be much better if it just equally divided the surface into curves with the toolpath closer in the narrow areas and further apart in the wider areas, if those conditions exist in the part. Much like the Advanced Planar they added, you would just give a stepover value that would be the maximum step over, and then it would squeeze the toolpath where it needs to in tighter areas.

However, if you're doing what I'm calling parallel curves where the distance between them is constant, the current drive curve function of Equidistant Offset does perform well. Here's an example of an archway type moulding that might benefit from that toolpath, where it used to be very difficult to produce in Bobcad 3 Axis Pro. In the straight part, I used the new Advanced Planar Slice, which keeps the step over even regardless of the change in slope of the surface. You could have used the drive curves function of Equidistant Offset to get the same result here. On the arched version of the same profile, I used the drive curves function of E.O., and you can see that they are remarkably similar toolpaths, just one in circular motions and the other in linear motions. Overall, the new stuff they added is definitely a huge upgrade from where we were at in V26, just wish the drive curves function was a little better when the curves are not what I'm calling parallel like this example. Again, you could not have cut this moulding this well before without 4 Axis Mill Pro, so hard to complain too much.

[BurrMan]

Why would I and when would I want it over the old tool path choices ?:

It drives the toolpath more in the shape of the surface, which will always give you a better surface finish (One of the things that creating a toolpath of "equal distances always", is good at)

mmoe gives some good pointers and insight. I'll touch on a couple things too. (Speculation)

The equaldistant toolpath was a pre-existing toolpath for BobCad.. They added a new feature to make it better...

mmoe is discussing the "Morphing, or blended toolpath option. The multi axis toolpaths are very good for 3 axis too!

Really though, it's a 5 axis toolpath that's just having it's tilt angles locked in the z....

The new toolpaths come from Moduleworks and surely BobCad is paying a pretty penny to them when they license one out....

Having to pay the multi axis price to get them is more than likely just paying the Moduleworks licensing....

With that being said, if surface finish of that kind of magnitude is required, it's an investment you'll want to make, even for a 3 axis machine... (ie; cutting with the side of a ball mill instead of the tip)

[mmoe]

It drives the toolpath more in the shape of the surface, which will always give you a better surface finish (One of the things that creating a toolpath of "equal distances always", is good at)

mmoe gives some good pointers and insight. I'll touch on a couple things too. (Speculation)

The equaldistant toolpath was a pre-existing toolpath for BobCad.. They added a new feature to make it better...

mmoe is discussing the "Morphing, or blended toolpath option. The multi axis toolpaths are very good for 3 axis too!

Really though, it's a 5 axis toolpath that's just having it's tilt angles locked in the z....

The new toolpaths come from Moduleworks and surely BobCad is paying a pretty penny to them when they license one out....

Having to pay the multi axis price to get them is more than likely just paying the Moduleworks licensing....

With that being said, if surface finish of that kind of magnitude is required, it's an investment you'll want to make, even for a 3 axis machine... (ie; cutting with the side of a ball mill instead of the tip)

All good points. From looking at the Module Works website, and how they seem to classify their toolpaths, I do wonder if they could still put together a 3 axis only version of morphing toolpath and still be only subject to the 3 axis licensing. It just seems like they are so close on the Equidistant Offset with drive curves that it would be a minor derivative to add something along the lines of an "advanced" step over to go along with the drive curves. This would prevent it from being used as a 4 or 5 axis toolpath, but would be a huge improvement for 3 axis users. From a business perspective, I need that type of toolpath only once in a while, so hard to justify the 4 Axis Pro level price for half a dozen yearly jobs (the quote I got for 4 Axis Mill Pro last year was more than double 3 Axis Mill Pro with several add-ons). I think I'll submit a feature request for that morphing-type step over option in Equidistant Offset (it would still be equidistant, just equally divided, rather than equally stepped ). It couldn't hurt to ask and it may just be that they hadn't thought of doing it that way.

If it can be done outside of the 4/5 axis dialog window, I suspect it may be able to be done under the existing 3 axis licensing, but that's purely speculation based on what it looks like MW uses to delineate between levels of licenses. If we had that available, I think Bobcad would have the hands down best 3 axis product on the market when it comes to available toolpaths strategies (they are close as it is). If you have to go to 4 Axis Mill Pro, Bobcad does have some competition there with Mecsoft products, and the pricing/features aren't dramatically different between them at that level. Bobcad seems to enjoy it's best pricing and feature advantage at the 3 Axis Mill Pro level, when comparing features to other products.

Not to poo-poo what they have done though. I can see the drive curves option being very valuable for the woodworker who processes a lot of custom moulding. In fact, I may be able to get new clients based on that alone, so it is a huge add. I also think the Advanced Planar Slice is going to become a go-to 3d option, where in the past it just wasn't practical to use Planar Slice for many jobs due to the way it reacts to steep walls. I've never been a fan of the way Equidistant Offset looks on the majority of parts I use it on, where the new Advanced Planar will probably produce a better looking surface finish. I do a lot of aluminum and plastics, so the dwell from the hard stop/start in the chevron shaped toolpaths that Equidistant Offset frequently produces creates a bit of an ugly spot in the part. The main reason I've used Equidistant Offset on those parts is simply that there has not really been any other option for good finishes on both flat and steep surfaces. The new Advanced Planar Slice is just fantastic as it incorporates the best of Equidistant Offset type utility with a better overall motion for what I expect will be very smooth finishes, especially in finicky plastics.

On an unrelated topic, I remember that you were looking for a good subdivision modeler a while back? Form-Z 8 was just released and has added that functionality to it. I can't speak to how good it is, but they seem to do things right in Bonzai and Form-Z, so I'm guessing it's probably pretty solid. It looks like you can take your model and convert to NURBS, which might be a handy tool in other ways as well.

[BurrMan]

It couldn't hurt to ask and it may just be that they hadn't thought of doing it that way. .

For sure! Actually, I think (could be wrong) that the equal distance and slice planars are still separate from Moduleworks (We had those before)... Adding intermediate steps to equal distance would have it not be equal distance.. But, they HAVE added some functionality to the slice planar which is getting it closer to what's being described.. I think it may be a better fit to address it in there?.....Anyway

On an unrelated topic, I remember that you were looking for a good subdivision modeler a while back? Form-Z 8 was just released and has added that functionality to it. I can't speak to how good it is, but they seem to do things right in Bonzai and Form-Z, so I'm guessing it's probably pretty solid.

Yeah, I was trying to find a good replacement for Hexagon. Just a pure box modeler. Unsuccessful so far

I've always had some drool for Autodesys.. Almost pulled the trigger a few times.. A plethora of tools (some good ones I would like to have)... I don't really like the "mix" though of all the different types of geometry.. It's just personal for sure... But, I always look in from time to time.... For me, they fall into the same category with Punch stuff... FormZ has been around along time... I have good things to say about both...

[mmoe]

For sure! Actually, I think (could be wrong) that the equal distance and slice planars are still separate from Moduleworks (We had those before)... Adding intermediate steps to equal distance would have it not be equal distance.. But, they HAVE added some functionality to the slice planar which is getting it closer to what's being described.. I think it may be a better fit to address it in there?.....Anyway



Yeah, I was trying to find a good replacement for Hexagon. Just a pure box modeler. Unsuccessful so far

I've always had some drool for Autodesys.. Almost pulled the trigger a few times.. A plethora of tools (some good ones I would like to have)... I don't really like the "mix" though of all the different types of geometry.. It's just personal for sure... But, I always look in from time to time.... For me, they fall into the same category with Punch stuff... FormZ has been around along time... I have good things to say about both...

So you are thinking perhaps adding drive curves to the Advanced Slice Planar? That would be interesting for sure. It already has the capability of essentially variable width step-overs, so not really a stretch to think drive curves could be incorporated as well.

I've had Bonzai 3d for a while now. I still prefer to use Viacad for anything that has extensive 2d wireframing, but Bonzai is definitely superior in most 3d applications. I often use them both, 2d in Viacad and then 3d in Bonzai. Bonzai has been exceptionally stable compared to just about any other CAD product I can think of. The only thing I have not liked about Bonzai is that it has more limited file import/export options than Viacad and some other CAD products (no IGES?!)

[BurrMan]

So you are thinking perhaps adding drive curves to the Advanced Slice Planar? That would be interesting for sure. It already has the capability of essentially variable width step-overs, so not really a stretch to think drive curves could be incorporated as well.

I've had Bonzai 3d for a while now. I still prefer to use Viacad for anything that has extensive 2d wireframing, but Bonzai is definitely superior in most 3d applications. I often use them both, 2d in Viacad and then 3d in Bonzai. Bonzai has been exceptionally stable compared to just about any other CAD product I can think of. The only thing I have not liked about Bonzai is that it has more limited file import/export options than Viacad and some other CAD products (no IGES?!)

Yes. Add the driving to slice. But then that removes the "planar" lol... could just be "simplified 3d surfacing"....

Bonzai = formz light with the new features of 7 and 8..... def a step above punch, but punch has been getting passed around to foster homes... it may be finding a family with poly to fitted nurbs that they are incorporating now...

[Forrest C]

Are all the same toolpaths and added functions of V27 available in the latest V4 for SW's?

Forrest

[mmoe]

Are all the same toolpaths and added functions of V27 available in the latest V4 for SW's?

Forrest

I can't say 100% for sure, but it appears that V4 is pretty much the same. I know it has the strategy additions we're talking about here. Bobcad just emailed out a sales message that had a link to a video of V4, and all the new V27 features appeared to be there. Whatever differences there may be, they are not significant that I can tell.

[SBC Cycle]

Wouldn't the "morph between two curves" essentially be a ruled or revolved surface with an applied offset and a specified maximum stepover? V21 had the ability to create Ruled and Revolved toolpath that I would love to have back. If I wanted to machine a twisted surface like this one I would simply want to follow the u and v lines, dividing one or the other in a specified interval.

No 5 axis should be required for such a simple construction. 5 axis would make it better of course, either allowing you to maintain a tilt angle to avoid bottom cutting with the ballnose or keeping the tool perpendicular to the surface.

[mmoe]

I had V21, but I had the earlier version which only had the basic 3d toolpath strategies (Planar, Radial, Spiral) for 3d work. I think they completely redid Bobcad with something like V21.5 where they also added more of a toolpath wizard (IIRC), but I didn't update Bobcad from V21.2 until V24. I'm really curious to see how that would have worked. For the person who does parts such as guitar necks, or something of that sort that is 3 axis only, a morph between 2 curves 3 axis option would be the cat's meow. I suspect the overall finish quality would be considerably better than the standard parallel cut that most would use. Otherwise, you end up with a bit of a serrated edge to the part. I just can't imagine it would be any more difficult to make calculations for than a radial toolpath on a 3d object, and the current Bobcad developers seem to have some serious kung fu coding abilities based on the updates between V26 and V27 done in only a year (or less?).

[BurrMan]

A morph toolpath is a blending of 2 curves. There is not a constant step produced from that...

Attachment 249042

The surface that creates a "Blended isocurve structure would be a swept surface, where the 4 endpoints connected would sweep along the "2 curves'....

Attachment 249044

BobCads new "ruled surface" s more like a "lofted surface" Which creates an analytical surface with an isocurve structure that doesn't equate to a blend.

Attachment 249046

If you create a surface with the "swung surface" command, we're only given the ability to select "1 sweep path", which then just does a default selection/calc for the other side (Most likely based off the UCS...

This generates a 3d surface, but to morph calculate that, you have to now use the normal of the surface also... It becomes a very complex calc that is done in the 5 axis modules with all the functionality they provide for reading surfaces, we can then lock the z angle... You couldn't perform that calc from a simple z projection down to the surface...

[mmoe]

Interesting stuff Burrman!

I guess that when I'm looking at how the new Advance Planar Slice works visually, I'm picturing that it must be taking sections through the geometry into account. In the past, I've observed that Bobcad only generates curves along the XY, XZ, and YZ planes when you have arc fit selected in 3d strategies, so it's safe to say that any surface that would benefit from a morph-like toolpath would also be one that does not meet those constraints. That would mean that the toolpaths would have to be line segments anyways, which I really have no problem with since that's how Equidistant Offset works already, and the quality of toolpath output in that format is excellent (other than my dislike of the chevron direction change on some parts). With those thoughts in mind, could it be that taking sections of the part in a similar manner to how they seem to already do so would allow a similar "adaptive" spacing to occur between the drive curves, where line segments would connect along the proportions of the sections taken? Of course, it's tricky in the sense that the sections would be varying in their placement angle, which would require the "maximum segment length" parameter, but at the same time, I think it is not so different from how I suspect is must be done for Advanced Planar Slice already. What are your thoughts on how they are getting Advanced Planar done? I really liked your idea of proposing that drive curves be added there, and do wonder if it's already doing similar functions.

[jrmach]

since that's how Equidistant Offset works already


arc fit spits bad code out,,,I have to use line segments,,,,is that what you are referring to ?

[mmoe]

If I remember right, arc fit did something weird in Equidistant Offset when they introduced it where the code was not as smooth as it was under the same conditions with arc fit disabled. I also recall that if you increased the tolerance setting, arc fit would come back to about the same level of quality, but at the expense of a lot more calculation time.

What I was referring to is that the arc fit function only produces arcs that are planar to either the XY, XZ or YZ plane. Even if you have a perfect sphere, cut at a 45 degree angle in half, when the equidistant toolpath runs along that 45 degree edge, it will not generate any arcs despite the fact that it is running up or down a perfect arc surface. Since that motion does not correspond to the XY, XZ or YZ planes, no arc is generated. If you rotate the same model 45 degrees so that the face dividing the sphere is along either the X axis or the Y axis, it will generate an arc motion in that direction. However, it only does so for a few passes before the adjacent passes start to stray from a straight line and it switches to lines because it no longer follows the geometry on the XY, XZ or YZ planes. That is the nature of Equidistant Offset. It doesn't follow straight lines, it equally spaces the pass from the previous pass, or it would behave like the old Planar Slice.The first pass along an edge will be perfectly in line with the axis, but the following passes will slowly get less and less straight. As soon as they exceed the arc fit tolerance setting, lines start to be substituted instead because they are no longer planar with XY, XZ or YZ. Hope that makes sense. I don't see anything wrong with Equidistant Offset. It's just not a toolpath that will ever generate many arcs because of the fundamental reasons we use it.

In Equidistant Offset, I just leave it with arc fit unchecked because the benefit is minimal (maybe only 1% will ever be fit with arcs) and the calculation time for equivalent quality is as much as quadruple. The finish quality of Equidistant Offset does not leave anything to be desired, other than the chevron turns I complain about, so I have no problem with it using line segments. My machine runs those line segments just as smoothly as it would an arc anyways, so smoothly that I actually thought it was arc fitting the toolpath originally.

I'm not 100% sure and perhaps Burrman can correct me if I'm wrong, but my understanding of Module Works is that it is still working off of meshes, which it generates during each toolpath calculation off of the solids in the viewport. While we appear to be working directly from solids (and graphically we are), I think there is actually a translation from solid to mesh to toolpath. As such, recognizing how to fit arcs to those mesh surfaces is likely a very difficult task, very similar to what Viacad is doing in their CAD system to allow you to convert mesh objects back to solid objects. If I'm right, Module Works creates a set of triangulated surfaces, which are easy to get points to travel through (tighter tolerance equals a tighter mesh), but then to arc fit it has to convert those now faceted pathways back to an arc moving through them. I do not believe that it reads the arcs directly from the solid, but rather it has to fit a mesh to the solid, then fit an arc to the mesh. Not easy if that's the case. Again, that's just what I've read into the various bits of info I've got here or there, so I could be wrong on that account.

[BurrMan]

I'm not 100% sure and perhaps Burrman can correct me if I'm wrong, but my understanding of Module Works is that it is still working off of meshes, which it generates during each toolpath calculation off of the solids in the viewport. While we appear to be working directly from solids (and graphically we are), I think there is actually a translation from solid to mesh to toolpath..

I can take a best guess but don't have the required behind the scene's info to state "this is for sure"......

I would say you have added a step in the moduleworks process that they don't use... Although, it generally is correct! lol.. (I'll try not to be confusing) The process you describe of converting to a mesh, then toolpathing that, I would say is used in less desirable strategies.... I think they are generating the toolpath from the actual read surface, but, gcode and machine controllers don't have a "bspline" or don't read knot vectors and weights and such, so the calculation actually is a triangulated one, with points plotted that way... (Cheaper softwares will use the mesh that's generated by the video card for display, then plot that... very quick, but not really your model.... So basically, the "Toolpath is a generated mesh".... per the projected calculations from the Moduleworks math...

I think your discussing the same thing though, in posts 15 and 17... Really when you move an arc into 3d space you don't have an arc, you have a conic, which no longer has anything to do with gcode or a machine (the arc fitting/equal distant thing you were referring to) when you are calcing an algorithm that is projection plotting surfaces with triangulation...

Most simply, I don't think using just 3 xyz values you can properly define an arc without the assumption of a plane (a fourth value that would be ever changing on a twisting surface)....

I need to read #15 a bit more....

[SBC Cycle]

Every time a new version comes out I get a chance to toy with the multiaxis stuff but I've never had a chance to dig deep into it. I used "morph between two curves" on relatively simple surfaces that I couldn't get great toolpath on using the 3 axis Pro strategies and the toolpath appeared very similar to what I got in V21 with the Ruled surface and an even older software called "Powersurf". It is very possible (even likely) that there is much more to the morph than that.

Someday I'll have that multiaxis. Someday.

[mmoe]

Burrman,
Did you replace your graphics card with a workstation card yet? I remember (I think) that you had a card fail a while back, and wondered what you replaced it with. I'm looking at picking up one of the new K2200 Quadros, and curious what's working for you? I've been using my old Geforce 560ti, which works pretty reliably and has always been stable, but I've always thought a workstation card might be a good upgrade. The rest of my system is all pretty much workstation quality, so it would finish it off nicely, and the K2200 looks like it's even a bit better than the K4000 was. Sorry it's off topic, but I think we've already played the topic out for the most part anyways.....