A planer blade setup would need to be at least 4 axis, you would need to change the angle of the cutter as you moved out to the rails. I guess that's the thinking behind this machine, 5 axis means you get a ready to glass finish in the same time as a pre-shape from a 3 axis machine.
I don’t think 4 axis would be enough. You would need the head to angle and rotate perpendicular to the rail, so it would have to be 5 axis to use a planer blade.
With a flat bottom end mill they could get a similar effect to the planer blade, but it would take more passes to get a smooth rail, unless they used a 4" diameter end mill, which would probably be a safety issue.
Hey Kenz, maybe you should invent a finbox system that works with that router. One where the slot is perfectly compatible with existing fins, but the part of the base that gets glued in is rounded to the same radius as the cutter. All they'd have to do is plunge it in & they're box ready as well as glass ready... glass over the boxes is stronger & saves time anyway...
That’s a pretty good idea Benny…maybe you should have patented it before you blurted it out.
A planer blade setup would need to be at least 4 axis, you would need to change the angle of the cutter as you moved out to the rails. I guess that’s the thinking behind this machine, 5 axis means you get a ready to glass finish in the same time as a pre-shape from a 3 axis machine.
I don’t think 4 axis would be enough. You would need the head to angle and rotate perpendicular to the rail, so it would have to be 5 axis to use a planer blade.
With a flat bottom end mill they could get a similar effect to the planer blade, but it would take more passes to get a smooth rail, unless they used a 4" diameter end mill, which would probably be a safety issue.
There are a few reasons why I think Miki’s cutter is pure genius. For most (not all, most) boards his 3 axis machine with a his big cutter will do what this 5 axis machine with a big cutter will do. And his 3 axis machine with a big cutter will do the same as a 4 axis machine with a planer blade. The math behind the toolpath calculations and compensations are so simple it’s a joke compared to bullnose cutters(not even thinking about 4 axis and 5 axis). Only (big?) con I can see, it won’t do swallow or bat tails or things like that.
my main priority for the design and approach to my machine is maintaining the integrity of the HAND SHAPED design. relying on incremental cross sections to CREATE surfaces you have lost all integrity of the modeled board
A good way to reduce transition (time and reject boards) from hand shaped to machine pre-shaped. But how does this facilitate incremental changse to the design?
I rather like the idea that surfaces are created to join cross sections. This means that the designer has to understand a bit how the surfaces are likely to blend in order to locate the slices. I mean that slices may have to be moved aft or fore in order to ensure that concaves start in the right place, for instance. It’s a challenge to get control of both rail transitions and bottom contour changes without freaking the computation algorithms out (by having too many control points located in different places in the different planes) while only using a couple of slices. Maybe your point, OAK.
The end result of a good computer design seems to be smoother transitions (and a big chance of bunny ears and other wobbles during the learning process!) . These are definitely barriers to entry. Anyone can design and pre-cut a sort of decent board. Getting exactly what you want from a pre-cut is a different animal.
OAK you’re not linked to Kenny by any chance, are you?
Harvaard, I’m interested in a picture of 3 vs 4 vs 5 axis paths if you have one handy.
A good way to reduce transition (time and reject boards) from hand shaped to machine pre-shaped. But how does this facilitate incremental changse to the design?
I rather like the idea that surfaces are created to join cross sections. This means that the designer has to understand a bit how the surfaces are likely to blend in order to locate the slices. I mean that slices may have to be moved aft or fore in order to ensure that concaves start in the right place, for instance. It’s a challenge to get control of both rail transitions and bottom contour changes without freaking the computation algorithms out (by having too many control points located in different places in the different planes) while only using a couple of slices. Maybe your point, OAK.
The end result of a good computer design seems to be smoother transitions (and a big chance of bunny ears and other wobbles during the learning process!) . These are definitely barriers to entry. Anyone can design and pre-cut a sort of decent board. Getting exactly what you want from a pre-cut is a different animal.
OAK you’re not linked to Kenny by any chance, are you?
Harvaard, I’m interested in a picture of 3 vs 4 vs 5 axis paths if you have one handy.
my personal approach to scan and digitizing techniques are proprietary to my system. Being a shaper first and not a designer I have approached my method from a different angle. it is completely and totally different that what most all other 2d methods shape3D, asp3000, brazilian machine etc… in that I do not rely on any cross section (slices) to create surfaces. my method ensures accurate data true to the original master model. from this data I can then modify any axis to any dimensions. with the 2d method you are limited. and the boards are never accurate and true to the master. never. you are using only 2d and relying on algorithms to fill in the blank data. the method I have chosen utilize true 3d data and no surfaces are created.
I am speaking from many years of experimenting from trial and error. I have many unsatisfied customer that have used the other shaping systems and methods with less than stellar results.
Intriguing - certainly got me questioning the “industry standard”
I guess the thing I see as a benefit (interpolation of curves) are disadvantages when trying to reproduce a board or achieve an exact form. I can see the benefits of not being held back by software/hardware implementations. Your technology concept could REALLY stir the pot were it to be released on the market - perfect replicas of (other people’s) boards!
Kenny’s someone whose been making machines for a number of years.
my personal approach to scan and digitizing techniques are proprietary to my system. Being a shaper first and not a designer I have approached my method from a different angle. it is completely and totally different that what most all other 2d methods shape3D, asp3000, brazilian machine etc… in that I do not rely on any cross section (slices) to create surfaces. my method ensures accurate data true to the original master model. from this data I can then modify any axis to any dimensions. with the 2d method you are limited. and the boards are never accurate and true to the master. never. you are using only 2d and relying on algorithms to fill in the blank data. the method I have chosen utilize true 3d data and no surfaces are created.
What kind of algorithms are you using for the surfaces? There are one surfboard cad program out there, BoardCAD, that use Nurbs which are truely a 3D surface representation. However, don’t disregard the bezier slices as you can use this for mathematically accurate surface representations, with some limitations. For the standard surfboard it’s no problem. I think the key to the problems with beziers are how the interpolation between the slices are implement and how it’s represented on screen. APS3000/Akushaper has some limitations in this department, Shape3D looks better IMHO in these case, for ‘regular’ shapes the are virtually identical. Surfcad I don’t understand. Or are you saying you don’t want a smooth change in between the slices? Why?
I’ve heard about mixed experiences with scanning and modifying also. You take an inaccurate plug or board through an inaccurate scanner and get an even more inaccurate copy (not my words…). YMMW of course, as there are good and bad systems. For a software based system, you really need to know the software, it’s strengths and limitations to make the most of it. Some shapers will get it, some won’t and their experience will reflect this to some extent. If the board is designed in a program(the ‘master’), given accurate toolpath generations it should be true to this master. If it’s off, it’s easily measurable.
What kind of algorithms are you using for the surfaces? There are one surfboard cad program out there, BoardCAD, that use Nurbs which are truely a 3D surface representation. However, don’t disregard the bezier slices as you can use this for mathematically accurate surface representations, with some limitations. For the standard surfboard it’s no problem. I think the key to the problems with beziers are how the interpolation between the slices are implement and how it’s represented on screen. APS3000/Akushaper has some limitations in this department, Shape3D looks better IMHO in these case, for ‘regular’ shapes the are virtually identical. Surfcad I don’t understand. Or are you saying you don’t want a smooth change in between the slices? Why?
I’ve heard about mixed experiences with scanning and modifying also. You take an inaccurate plug or board through an inaccurate scanner and get an even more inaccurate copy (not my words…). YMMW of course, as there are good and bad systems. For a software based system, you really need to know the software, it’s strengths and limitations to make the most of it. Some shapers will get it, some won’t and their experience will reflect this to some extent. If the board is designed in a program(the ‘master’), given accurate toolpath generations it should be true to this master. If it’s off, it’s easily measurable.
haarvard, I understand your mindset with algorithms and the beauty of nurbs and bezier curves. I had to think out of the box and did not want to rely on these to accomplish my task. as they work great for creating but not for exacting a specific model. as a shaper first and coming from the old school of shaping, I trust my eyes and hands to make my creations, when you are in the shaping bay and laying out a new design it’s a symbiotic process at least for me. not in terms of algorithms to desgin my boards but more about subtle nuances that algorithms will not capture with only a few plot points, much more complex processes are taking place that one can not really explain. once my handmade creations are designed I wanted to have the ability to replicate and modify my hand made creations with out any distortions and scale any axis to my designs infinately with out losing the essence of the original board. the beauty of this metod ensures the next generation will be refined as to the last generation and your shapes evolve as you would as if you are hand shaping but you have much more control over design replication and general quality of shapes. this ensures your customers are getting your best shapes on your best day.
Intriguing - certainly got me questioning the “industry standard”
I guess the thing I see as a benefit (interpolation of curves) are disadvantages when trying to reproduce a board or achieve an exact form. I can see the benefits of not being held back by software/hardware implementations. Your technology concept could REALLY stir the pot were it to be released on the market - perfect replicas of (other people’s) boards!
Kenny’s someone whose been making machines for a number of years.
as intrguing as it is, I have been tempted from a couple “over seas” ventures and have turn down some lucritive offers. knowing the possible tumoil this could do in the wrong hands I have politely refused their offers. back when these possible ventures were available I didn’t think anyone here in the states would sell out our future. funny, now looking back I can’t believe the long list of shapers that have. I am still commited to helping shapers here in the cali maitain what little future we have left. It has been a tough battle.
haarvard, I understand your mindset with algorithms and the beauty of nurbs and bezier curves. I had to think out of the box and did not want to rely on these to accomplish my task. as they work great for creating but not for exacting a specific model. as a shaper first and coming from the old school of shaping, I trust my eyes and hands to make my creations, when you are in the shaping bay and laying out a new design it’s a symbiotic process at least for me. not in terms of algorithms to desgin my boards but more about subtle nuances that algorithms will not capture with only a few plot points, much more complex processes are taking place that one can not really explain. once my handmade creations are designed I wanted to have the ability to replicate and modify my hand made creations with out any distortions and scale any axis to my designs infinately with out losing the essence of the original board. the beauty of this metod ensures the next generation will be refined as to the last generation and your shapes evolve as you would as if you are hand shaping but you have much more control over design replication and general quality of shapes. this ensures your customers are getting your best shapes on your best day.
If you scan a board given a finite number of points, you are effectually creating a mesh of the board. There are no more way to fill in the blanks in a mesh than in bezier model or a nurbs model. If you want to be that specific in bezier, scan a slice at the same granularity as you otherwise would, for nurbs, scan and build segments with the same granularity. Advanced 3D modeling programs (3D max, lightwave, etc.) support several if not all of these tools. Each have their strengths and weaknesses. Nurbs are thought of as the tool for organic shapes, beziers are easy to use, meshes are the workhorse of the industry. Just my $.02 worth.
If you scan a board given a finite number of points, you are effectually creating a mesh of the board. There are no more way to fill in the blanks in a mesh than in bezier model or a nurbs model. If you want to be that specific in bezier, scan a slice at the same granularity as you otherwise would, for nurbs, scan and build segments with the same granularity. Advanced 3D modeling programs (3D max, lightwave, etc.) support several if not all of these tools. Each have their strengths and weaknesses. Nurbs are thought of as the tool for organic shapes, beziers are easy to use, meshes are the workhorse of the industry. Just my $.02 worth.
Yes, in theory you are correct. It comes down to the number of points the closer the points the more accurate. With the programs that are being delivered with the over the counter machines the scanned board in reality is not remotely close to the discerning eye. I understand where you are trying to create segments in surfaces and that’s in where the difference is. Believe me, I have many customers that have wasted mucho time and money trying to get their shapes accurate to the model board.
Relying on algorithms is a great way to create surfaces and does a good job of smoothing out point to point type of surfacing but if your want an exact shape in terms of replicating a hand shaped board it will have a tendency to average out the surface and that is where the discerning eye will be able to differentiate the difference.
I’m sure there are many shapers here that have experienced this and understand the difference.
I like to compare the difference like that of recording of music tracks. There is synthesized version of recorded music where the notes and type of sound are created through processing. Then on the other hand, you have analogue music that is created and played by a musician, then track is recorded into a digital format. I liken to think my process is like the later. Nothing wrong with synthesized music I actually dig it. But there is something lost from the original recording. Same notes different sound. Same dimensions but not accurate to the real deal.
Nice discussion, keep it going. Bezier, nurbs, meshes - all good stuff for an uneducated tecchie like me. Now I can direct my research.
My design philosophy encompases belief in smooth transitions between surfaces so the interpolation works for me. But you’re right, it’s easier to design a pre-cut board from scratch using the industry standard than to copy an existing model.
Got any good pointers for cutting path calculation algorithms?