Hey fellow engineer! From the „big“ CAD programs (catia, sw, creo, nx, inventor) only NX currently has a feature to accurately sweep volumes (like a milling bit) along more generalized curves. It‘s called „Swept Volume“. And then it only works for convex shapes. All the others only have solutions for special cases like cutting heliocentrical spirals. We developed a different method to generate these surfaced, as we need them to project the path of a more complex body onto another body.
@@MnemoLogic ignore me, I was talking about sweeping a sketch in F360. Though I'd think sweeping a volume wouldn't be much more difficult for them to implement. I'm not a professional designer, but I feel like you could cut the path pretty easily like that though. Punch a hole at either end, then sweap a rectangle as long as the diameter of the bit, and as tall as the material thickness? Though this vid was about how it's surprisingly tricky, so maybe I shouldn't talk before trying it myself.
Awesome video! For a more reliable result - just use a projected curve instead of the wrap feature. Wrapped curves do not reliably maintain curve characteristics/constraints, this is why Solidworks isn't recognizing the wrapped path as tangent. If you use a projected curve onto the cylindrical surface, you can do a simple solid body sweep that is aligned to a direction vector (the center axis of the cylinder). This will produce a result that is exactly what will be milled if the endmill is pointing at the center axis of the cylinder throughout the entire path.
The sketch to wrap can contain multiple, CLOSED contours only. You cannot create a wrap feature from a sketch that contains any open contours. So, how did you get it?...
My first "for fun" solidworks project after my first semester of CAD classes was to replicate a Sten MK2. I figured it was fairly simple and would be fun to model. I ran into this exact problem with the slot in the receiver. I was humbled as I thought my CAD-fu was too weak to replicate this very simple mechanism. Good to know that over a decade later this same problem is still stumping so many engineers!
As a fellow engineer with some experience in making some wild stuff in CATIA I applaud you for pointing out the logic for the tool path. I used the offset-method quite a bit, because CATIA loves surfaces and hates boolean operations, plus you have it parametric associative, and everybody loves that. But a sweep of the tool cross section along a path of the surface, using the path as the spine and normal line to the surface as axis would be usually quicker. Greetings from Germany 🥳
I would have done this in Alias thar has very powerful surfacing & diagnostic tools and forget about the limitations of most solids modelers, but the fact that you solved the feature fully within Solidworks is a testament to the depth of your understanding of it. Congratulations, you should be top hiring material!
I mean it's still a pretty simple geometry, no need for Alias for that kind of stuff. Take a look at what modern solid modelers are able to do with "limited" surfacing tools nowadays : ruclips.net/video/IjcFotLmCx8/видео.htmlsi=QAdyi9NBdDt4wiXG
I laughed when I saw that LTT episode. I have done so many cam following grooves in SW using surfacing as it is the best way to do it. (Been there a long time ago and developed something pretty darn close to your third attempt) Surfaces let you have a lot more freedom in SW and is the best method (in my opinion) to design your parts for tooling and die/mold work. It's also applicable if you want to do helical slots in a shaft that rotate and change incline angle and taper. (like a mixing head on an extruder screw)
@@mittens9326 indeed! The only thing I wish is if SW would maintain an intersection constraint through a rebuild. Struggle with that sometimes. Fusion let's you do it! But Fusion has issues of it's own 😂
@@bradleynealdaley give Onshape a try, amazingly robust (much more than SW and Inventor in my almost 15 years working with those), great surfacing tools. Here's a great example : ruclips.net/video/IjcFotLmCx8/видео.htmlsi=QAdyi9NBdDt4wiXG
surfaces are great, no denying that, but this simple shape also could have been easily done with a loft cut with some guide lines... Also, never rely on this shaded view for interferences because your settings could be set to speed and not quality, and it will throw you off...
As an engineer with experience that I still consider humble, my first instict was wrap yes, but as soon as you showed that it extrudes with slanted walls, I immediately thought of surfaces. So I find it epic to know that I cam to the solution that quickly. Seems like 2 years of product development and DFMA work in addition to my other years of SW have done me well.
Well executed, walking through each method with detailed and understandable rational. Does a bidirectional surface thicken work in place of using offsets?
My exact thoughts ! Thicken-remove after the initial swept surface would save a few steps. But really, the key is understanding that this geometry needs to be modeled from a center guide.
I had exactly this problem recently modelling a dry break connector with a similar pin fitting. So frustrating how Solidworks kept on distorting the cut no matter how you try to constrain it. As you say, programatically it doesn't matter, as it's easy to do in CAM, but I too am OCD when it comes to modelling, and it drove me insane with the groove being visually inaccurate. I had to move on to other more pressing jobs (the dry break was just for representation within an assembly, and really no need for that level of accuracy, but I often take the opportunity to model superfluous parts to both expand my skills and produce practical assemblies/realistic renders), but I like your approach and will revisit the dry break now!
Mechanical engineering student here, with 4-5 years of prior experience with CAD before starting. A lot of my experience is in CATIA, and the easy way is to just use generative shape design. Create the cutout with a bit of projection, create the cylinder walls around it, convert to solid. I should try this tomorrow to improve my skills.
I'm very pleased with myself. I tried to think how I'd do it in Blender (because that's all I really know) and the method I came up with was pretty much the same as your final method. I'm feeling real smart right now.
I ran into this problem awhile back but never found a good solution. I am a machinist so it didn't matter too much, but I scratched my head on it a lot
A swepted cut with guide curves on the inside and out should do it as well. You might be able to get away with a single guide curve, but it often distorts a bit so would require testing. You can also sweep a tool through a path that might work. I had a tricky one years ago where we were making helical cuts with an endmill in the side of a cylinder (bar). Getting the geometry of that cut right was tricky and it mattered because it was a flow passage and we needed to do CFD on it. Surprising though, that the whole team couldn't figure this out.
You have less control with a solid sweep because it needs to manage more directions/dimensions at once. The advantage of a central surface,which is the correct way to do these geometries if you need good control on the result, is that it allows you to tightly control the part, and then control the volume (thickness) in another operation, which is sometimes not possible with solid sweeps.
Thank you, sir, for that wonderful explanation. I don't even watch the WAN Show on RUclips. I listen to the podcast. But the algorithm still knew that I needed to see this video.
The way I visualized this in my head was just a bunch of lines projecting out from the center forming the shape of a J and then thickening to take up the diameter of the mill, I had no idea how to translate that into CAD so this was really enjoyable to watch you take this and work it out, thank you.
Wow, first off all English is not my native language so I apologise for any mistakes in advance. So I actually made pieces like this in 2009 for a company. and although it looks real difficult to draw in 3D. It is actually very simple. You do not wrap the J form around a cilinder but you first unwrap the cilinder to be a flat surface then you place the J shape where it is supposed to go and take out the materiaal (width off the J shape) completely through the dept off the flat model. Now when you are done you wrap the flat surface back to a cilinder, and that is it. On a CNC lathe with the setup as you mentioned (ours was a DMG Sprint42) we actually programmed the line or shape we wanted to cutout on a C-plane as if it was a flat surface. I hope despite my language problem my explanation is understandable. 😉
Your English seems great! Check out my second video where I address this suggestion that was also made by others. ruclips.net/video/TAlYex2l9yI/видео.htmlsi=zrWo3We4RPLU1csg You end up with a slot where the walls aren't parallel. It might be possible to fix that in CAM but the goal of the exercise was to have CAD that is accurate to the finished product
@@JamesDeBono it was extremely helpful thank you. I searched previous for bolt action and cam and groove, but none showed how to do it around a cylinder. Hopefully you’ll get 1000s more subs off the back of it.
It seems like Inventor handles this process a bit better, to do that I would create my 2d sketch and wrap it as a 3d sketch around the cylinder, create my "tool" and sweep it along the path. I've done some pretty complex splines and it seems happy enough to do it. The second for the angular control, the other way to do it, would be to use a dimension formula, take the Dia of the OD, turn it into a circumference, then create your angle from the circ/360, The math would look like "Start to End Line Dim=((Dim(OD)*PI)/360)*Angle", as long as it handles dimensions like I think it does all you would need to do to quickly adjust the arc would be to adjust the angle part of the formula.
Wow, I missed this Wan show but I actually had this problem a few months ago as well. It took me a while to figure out how to model it up but it looks like we went through the exact same process.
Multiple ways to do it, that's for sure. I've done similar type of paths before back in SW2012 or so when we had a lecture on pierce and cuts that follow a path. So in a similar instance of cutting curves on a round pipe for 4th axis milling, we just drew the curve ( projected it ) onto the cylinder. Then at one end, created a reference plane that was perp to the end of the line. Created a rectangle ( To represent how an endmill is gona cut it ) and performed an swept cut that follows the line. Worked perfectly and accurealy represented what the endmill cuts..Took a second to verify this in SW2023 and got the same result. I guess sometimes you just know useless stuff ( me ) as for me SW is just a tool I dork around with. I just do CNC by trade. So I use solidworks based on how I know tooling works..if that makes sense. I like the 5+ axis CNC machinist for years --> Solidworks/Solidcam --> Engineer direction VS engineering degree --> solidworks --> HowDoMachineWork?? heh
LLT's problem is their asking engineers questions only machinists can answer. James looks like youre one of the rare individuals that understand this. If you were an engineer i had to work with, if gladly work with you man.
I had an issue with this once. A twist and and what not. Create the Body your cutting out of create the "milling bit" in Cad as a seperate Body Rotate it up along a path to cut it. I'll dig out the two features i used when i'm back in the office. Thats what our tech support was recomending as well.
I‘ve encountered that same problem a few weeks ago when I was designing a bayonet-lock into some part I needed to 3D print and it was extremely frustrating until I came to the same conclusion as you did!
I am not a solid works person but a 3dsMax one. It is mostly trivial to do in max. You can either do your J curve on the cylinder surface and extrude it flat toward the cylinder center, or make a ton of milling bit copies alogside the J to substract from the cylinder, or many other options. Keep in mind that you will have impresisions in anysoftware anyway so you'll have to build in tolerances in your cut anyway. Regarding your gap it is because of triangulation. A non flat quad polygon can be triangulated either concave or convex but never flat that is just math.
Try it and I think you will find it's not trivial. If you extrude the J towards the cylinder centre you will end up with slot walls that are not parallel. Making a ton of milling bit copies is also equally possible in solid works but would leave you with an equal amount of tiny curved cuts instead of a straight cut so wouldn't satisfy the end goal of having CAD accurate to the final milled piece. Lastly, there are no quads, tris or polygons of any kind in parametric CAD. It's purely vector based
I think the bigger issue is that although yeah, it's totally possible. It doesn't really look like a J because the curve makes it appear basically like a U.
This is one of the few times I'm glad I model in rhino. I'm an architecture 3D modeler, so I've never even touched solid works. Can't understand solid object modeling. Surfaces only for this guy
i don't get the layers am i supposed to put a sketch here a surface there and the resulting solid in the next? everything in one layer? just show me the objects
Thid is why solidworks pisses me off so much. This should be so simple, but the fact that SW has an option to keep the cutting face perpendicular to the round face of the tube, but somehow still fucks it up without explanation, is just classic solidworks.
As a machinist who have spent north of 500 hours in CAD (Fusion, Inventor, NX and some SW) the past few years for toolmaking and general quality of life things around the shop, this seems like a simple operation that should be possible in most environments. Just wrap the centerline of the profile and extrude a rectangle along the path
I only have a couple day's experience in Solidworks, but if I were making this in NX, I would wrap just the center line of the J shape onto the tube, then use the Law Extension tool, which basically extrudes the line, but the direction can be complex, and if I remember correctly "normal to face" is one of the options, So not I have a sheet going through the tube that is normal to the surface. Then I would thicken that sheet to the thickness of the slot that I wish to cut, and subtract that solid from the tube. Reading the comments, I see there is a simpler way, but that is just using the tools that I have experience with.
Also, for the first attempt you use, there is also a way that is similar to what you did the third try, that is to extrude a curved even surface. I only know how to do it in fusion 360 so I will put the workflow this way: 1. Emboss the J 2. thin extrude the J into a sheet body. 3. Cut the profile with the sheet body you just created. 4. Remove the face you just created, and loft them back together. I think it will work, and although It did not show clip in fusion, I believe it works.
This reminded me of my studies. Don't bring your non-distorted drilling bit to a distorted world. Just imagine you make everything flat again, not just the pencil, also the drilling bit.
I would have tried starting with the tube as two surfaces each with the j cut out with the same width. Created the inside surface between the two Js, capped the end of the tube. And finally knitted it into a solid.
Fun video, glad youtube recommended this to me :) Small detail I noticed, and perhaps this is not important to the functionality, is that the "reference" pens from the forum post have a more angular inner corner instead of the basic round J shape, not sure if this is done for stylistic purposes, ease of manufacturing or are vital to the mechanism...
If you define the center axis of the cylinder, open options in your swept cut dialogue, and choose "Profile Twist" is "Specify Direction Vector" and define the axis as the vector it does what you're looking for.
I've tried this too and it also doesn't work. I wish I could just send a screenshot here so I don't have to make another follow up. The feature totally glitches out even more than it did when set to "tangent to adjacent faces" I made a temporary image link tinypic.host/image/UcZ3d
In reality I believe the first “solution” creates a valid geometry that can be readily machined via 4th axis CNC as long as you use a cutter that is smaller than the width of the slot. This is generally preferred so you have a more consistent finish by only climb cutting or only conventional cutting the outline, rather than full-slot cutting. That being said, it’s worth considering whether or not you want a slot that’s tapered through the thickness, as it could affect the friction/feel of the bolt-action.
when performing the machining operation you specify the edge you want the end mill to follow, so it doesn't matter if it's tapered or not; you can create an offset in the toolpath with stock to leave (which can be a positive or negative value). while this is a "hacky" way to approach it, it doesn't really matter for the end product
Great work. And also detailed explanation on how you solved the issue. I would probably be to lazy and go "eeh as long as the outside parameter is correct then in the CAM software i will use the outside to set the toolpath
That's fair but if you want to try a 3D print proto or if you want the model to look correct it's nice to make the model perfect. It's definitely not always needed though I just can't help myself 😅
Depending on the cam software you can Export a 3D model of the finished machining simulation. Then clean up the mesh; Not the prettiest method but would work.
To get such a simple J shape it can be done with 3 extrude-cuts and 4 full face fillets. At 6:00 instead of offset+loft+knit you could use thicken-cut or thicken+(fillets)+subtract combine.
For this very specific shape maybe, but if you need to make some non-cylindrical shapes and in general anything else than just a straight line and a radius, surfacing is the more flexible tool. The goal of that video I think was to show a foundational method to get more complex results, not one that can barely achieve the result shown here. There is no need to build a closed surface and knit it though, you can work from the simple center surface and do everything from there with a thicken-remove feature and some radiuses for the end.
This. You're both right. It could be done in this case with extrude but it the j shape changes it could easily become much harder without the surface method
@@JamesDeBono I watched your recent video. Prior to commenting I tried this in Fusion360 and it worked. So I assumed Solidworks can do it too. In Fusion360 it can be spread to about 140° before full face fillet starts to fail. Today I tried in it solidworks and it worked only up to about 45°. I don't know why. Out of interest I played with Extrude-Bosses instead of cuts (and then subtract combine), and I can make face fillet work up to 140°. But it is very weird and finicky. More problems than it is worth. In regard to Solid profile cut sweep: something that could be helpful is to make 3d sketch-> convert entities (choose all those lines) -> Fit Spline. This creates one continuous curve, that might work. There is some "error" in the shape of that curve, but you can control that with Tolerance slider to minimize it below micron.
@@primoz6578 oh nice investigation work. Interesting to know the angular limitation of the face fillet feature! good work! It's unfortunate that there isn't one clearly perfect CAD program haha. They all have their issues and they are all expensive! The fit spline is a nice idea indeed but I think it bugs me knowing that it's not quite perfect. Also would it update if you changed the dimensions?
@@JamesDeBono It does update if you choose Constrained option. As long as converted entities don't break and stay connected in continuous line, it should work. " it's not quite perfect." No it is not, but you can think that it is better? Fit spline gives you more curvature continuous path. With straight line and arc there is an ugly point where "radius" instantly changes from "infinite" to real value. It is sudden jump, infinite acceleration, jerk, vibrations ... Gradual change in "radius" of curvature in Fitted spline would provide smoother path. Instant improvement of the design. After all, you have used line and arc because that is the easiest shape not necessarily the most perfect shape. A bit of different point of view for your inner perfectionist.
we had a similar problem (making mold for handbag) in our company we tryed for day , a the end we switched for Rhino 3d , so far the best decision we made ,its give so much freedom in the design process
I was told that using solids and bodies to model is like using Window/Mac GUI to use a computer, easy/good enough for most situations. Once you start using surfaces and manipulating faces you're now using the command prompt to navigate the computer, which is needed for a problem like this (complex lofts).
That's an interesting comparison! I can see what you mean. I definitely agree in that sometimes it's necessary to go to the surface level. I always aim to have the least amount of features needed to make something accurate and stable if things need to be edited. I think surfaces are the right choice in this case
If you make a plane perpendicular to the center of the radius you can draw your "j" and extrude it into the pen body ( do not let it combine bodies) then you can use the remove overlapping bodies tool and the slot where the two overlap will be removed. then hide the non overlapped "j" projection.
Model the slot as a body, and combine cut. I use this for like 90% of what I do in CAD. You can also fo things like copy the surface,.offset it, extrude to the offset and then comvine cut, so the bottom of the cut follows the contures of the main surface
In fusion360 I've seen others use the CAM functions to make weird geometry like that, altering the model but I would rather be able to do it in the CAD side of things.
Getting back into SolidWorks after being away from it for years, this reminds me of having bosses that just wanted the design out quick rather than 100% correct as long as the machinist knew what to do. Bugged the hell out of me some times when I could see the interferences like that.
I feel the same way, I understand it in a way, it doesn't matter to them but I feel the urge to make it right! Sometimes you have to let it slide for the sake of time but it's nice when you do something off the clock that you can make it perfect :)
I cant understand why the second option didn't work, it doesn't make sense. I ended up using the sheet metal tood of Fusion to do this, because I gad the same problem with the sweep, but it also colided, but when making a simulation with colisions activated, the pin went smooth, so, I'm not shure what to belive.
That's what's interesting about this problem. It appears like there is a simple solution bit when you try, it doesn't work. Check my latest video to see why ruclips.net/video/TAlYex2l9yI/видео.htmlsi=-2yq-ggusqQ0LYS4 I can almost guarantee the a project will not give you as accurate dimensions of a wrap and extruding it will give you slot edges that are not parallel to each other while also being perpendicular to the cebtelin of the cylinder. Its not as trivial as it seems
@@JamesDeBono Sketchup isn't programmatically, it's pure geometry so making a shape and intersecting it cuts out everything. Using follow me tool all dimensions are kept, in fact it does not work on it's own as the angle change causes the approach angle to change so you have to use a constrained follow me plugin for that. Once you shove the two pieces into each other and cut you get a mostly clean result, but of course with the downside of not being a cad path instruction and just geometry. The difference is in the tool for extruding the cut part, which tries to maintain dimensions, because that's what you need the most in archviz. Of course you could not use that model to program a cnc.
So there is no sweep in solid work? Wrap a sketch and sweep along that line. Then just extruded the ends for the circle bit. Yes the circle won't be perfectly normal to the surface but so won't the drill (this obviously won't work if you require multiple passes of an endmill, but then again you can do it in two parts, drilling the holes and then the rest) Edit, I just watched the rest, I see now .
I wonder if you couldnt just create a surface by wrapping the J on the outside, then wrapping a J scaled to maintain the width of the outer J (ie scale it back to the same width as the outer J to avoid the travezoidal shape of the naive pushed extrusion). But, i cant be bothered checking if that geometrically works, so your solution is it :)
So in normal people words, the sizing of the cut needs to be modeled to scale to the change in radius between the inside wall and the outside wall? And the “normal” way of doing this in Solidworks makes it that it scales linearly, meaning the inside wall and the outside wall have different sizes of cut?
I personally would have wrapped a sketch from outside like you did first, then do it again from the inside of the part, then loft cut between the two sketches. Though, I've heard that wrapping vs projecting, there are pros and cons. This is from a primarily Inventor user though so... 🤷♂️
Check out my second video where I address this suggestion that was also made by others. ruclips.net/video/TAlYex2l9yI/видео.htmlsi=zrWo3We4RPLU1csg the problem is that the inside sketch is impossible to determine from the outside sketch. The outer sketch has a circular arc but the inner needs to have an elliptical arc to match it but there is no way to perfectly determine the shape of that elliptical arc
Would just like to point out something about the wrap feature for the sketch. If your sketch for the channel is lets say 0.5" The projection on the surface is 0.5" for the ARC LENGTH, not the CHORD LENGTH. you'd need to do some math on the sketch to get the chord length, which is the width of the bit, to be correct.
@@JamesDeBono It was a little tricky, but I ended up using math/geometry and wrap projected two sketches, one on the outside and one on the inside then did a surface blend between both surfaces for perfect perpendicular walls. Would be nicer if we just had a sweep volume feature like some other CAD software
I know less than nothing abut solidworks but to me this seems like a resolution problem on curved faces where a quad should be a theoretical infinite number of 4 tris going in 4 on each face. This comes form me using blender for many years where you have much better fine control of mesh geometry and vertex mergers into quad faces.
Good video explanation of the problem. Your final solution does work, but the requirement to have such pipeline to get to this point is terrifying - at any change it would be double and triple checking a lot of things to make sure it's still valid. The second approach should be the valid one, but it makes no sense that the edges are cutting into the bit towards the center and not evenly or the other way around, and I would check manually distances there on a cross section just to have some knowledge what is happening, and you didn't show that on a video.
Very interesting. Could this possibly be something solid works could work towards? Implementing some kind of tool to allow this sort of cut to be made more easily?
About 2 years ago just started working as a engineer after graduation and needed to model something very similar and the product even ended up being patented by my company. It literally took me like 10 minutes of googling to find a solution for Siemens NX which has an even smaller community an RUclips etc. What are they doing at LTT 😅
You can also do this in Solidworks. Check out my second video where I address this suggestion that was also made by others. ruclips.net/video/TAlYex2l9yI/видео.htmlsi=zrWo3We4RPLU1csg but then you don't have perfect circular geometry and the swept body still doesn't like the spline
Check out my second video where I address this suggestion that was also made by others. ruclips.net/video/TAlYex2l9yI/видео.htmlsi=zrWo3We4RPLU1csg the issue is when you fold it back into a cylinder the walls of the slot will not be parallel
I only have experience in AutoCad (I know it is not very good in 3D modelling). But when I model there, I would draw the cylinder and a plane surface of the "J". Then I would extrude the J, so it becomes a volume. And after that I would drag the "J" into the cylinder and do a volumetric substraction. I would assume, that this would work for that kind of Problem 🤔 I don't know if that would be a solution for Solidworks tho.
Check out my second video where I address this suggestion that was also made by others. ruclips.net/video/TAlYex2l9yI/видео.htmlsi=zrWo3We4RPLU1csg the issue is when you fold it back into a cylinder the walls of the slot will not be parallel
I'm not sure if this was mentioned or not, bit in the real world nobody should and would mill a groove with a mill bit in the same diameter as the actual groove. You should always use something smaller for inner grooves and that would make all of those problems disappear
It's true that you would use a smaller bit than the slot but that doesn't make all the problems disappear. You still have the same problems you just take a longer time to cut it. The edges of the slot still need to be parallel for the mechanism to work correctly
I have never modelled anything in my life but you still made it very easy for me to understand your thought process
Hey thanks a lot! That's very encouraging for me 😊
@@JamesDeBono I have done VERY minimal modelling in Fusion360 and this helped me understand a few things better! kinda. a little.
@@JamesDeBononot just that, it's such a beautiful, brilliant idea. You've got my nerdy respect and admiration! Subbed my dude
Had a hard time picturing it when they were discussing it. This makes more sense. Thanks RUclips Algorithm for this helpful guide. Great work dude.
All hail the algorithm
Hey fellow engineer! From the „big“ CAD programs (catia, sw, creo, nx, inventor) only NX currently has a feature to accurately sweep volumes (like a milling bit) along more generalized curves. It‘s called „Swept Volume“. And then it only works for convex shapes. All the others only have solutions for special cases like cutting heliocentrical spirals. We developed a different method to generate these surfaced, as we need them to project the path of a more complex body onto another body.
So far NX has been the most robust software I've used
Weird that Inventor wouldn't have it, because Fusion 360 does.
@@StaleReference you got any examples?
@@MnemoLogic ignore me, I was talking about sweeping a sketch in F360. Though I'd think sweeping a volume wouldn't be much more difficult for them to implement.
I'm not a professional designer, but I feel like you could cut the path pretty easily like that though. Punch a hole at either end, then sweap a rectangle as long as the diameter of the bit, and as tall as the material thickness?
Though this vid was about how it's surprisingly tricky, so maybe I shouldn't talk before trying it myself.
Siemens added the functionality to Solid Edge also, like many things Siemens has ported from NX to Solid Edge.
Awesome video! For a more reliable result - just use a projected curve instead of the wrap feature. Wrapped curves do not reliably maintain curve characteristics/constraints, this is why Solidworks isn't recognizing the wrapped path as tangent. If you use a projected curve onto the cylindrical surface, you can do a simple solid body sweep that is aligned to a direction vector (the center axis of the cylinder). This will produce a result that is exactly what will be milled if the endmill is pointing at the center axis of the cylinder throughout the entire path.
The sketch to wrap can contain multiple, CLOSED contours only. You cannot create a wrap feature from a sketch that contains any open contours. So, how did you get it?...
Projected curves have an issue, and this is, that you can only project over the visible face of the cylinder. How could I fix it?
My first "for fun" solidworks project after my first semester of CAD classes was to replicate a Sten MK2. I figured it was fairly simple and would be fun to model. I ran into this exact problem with the slot in the receiver. I was humbled as I thought my CAD-fu was too weak to replicate this very simple mechanism. Good to know that over a decade later this same problem is still stumping so many engineers!
It's a surprisingly tricky one I think!
As a fellow engineer with some experience in making some wild stuff in CATIA I applaud you for pointing out the logic for the tool path.
I used the offset-method quite a bit, because CATIA loves surfaces and hates boolean operations, plus you have it parametric associative, and everybody loves that.
But a sweep of the tool cross section along a path of the surface, using the path as the spine and normal line to the surface as axis would be usually quicker.
Greetings from Germany 🥳
Hey Greetings from Germany also! Yes I'm Australian but live in Berlin. The sweep would be ideal I think but Solidworks has issues with it
I would have done this in Alias thar has very powerful surfacing & diagnostic tools and forget about the limitations of most solids modelers, but the fact that you solved the feature fully within Solidworks is a testament to the depth of your understanding of it. Congratulations, you should be top hiring material!
I mean it's still a pretty simple geometry, no need for Alias for that kind of stuff. Take a look at what modern solid modelers are able to do with "limited" surfacing tools nowadays :
ruclips.net/video/IjcFotLmCx8/видео.htmlsi=QAdyi9NBdDt4wiXG
I laughed when I saw that LTT episode.
I have done so many cam following grooves in SW using surfacing as it is the best way to do it. (Been there a long time ago and developed something pretty darn close to your third attempt)
Surfaces let you have a lot more freedom in SW and is the best method (in my opinion) to design your parts for tooling and die/mold work.
It's also applicable if you want to do helical slots in a shaft that rotate and change incline angle and taper. (like a mixing head on an extruder screw)
I'm happy to hear that we came up with similar solutions. Hopefully that means I'm doing something right ;)
I'm self taught in SW, and surfacing truly changed my life (and all of the projection tools).
@@mittens9326 indeed!
The only thing I wish is if SW would maintain an intersection constraint through a rebuild.
Struggle with that sometimes.
Fusion let's you do it! But Fusion has issues of it's own 😂
@@bradleynealdaley give Onshape a try, amazingly robust (much more than SW and Inventor in my almost 15 years working with those), great surfacing tools. Here's a great example :
ruclips.net/video/IjcFotLmCx8/видео.htmlsi=QAdyi9NBdDt4wiXG
surfaces are great, no denying that, but this simple shape also could have been easily done with a loft cut with some guide lines... Also, never rely on this shaded view for interferences because your settings could be set to speed and not quality, and it will throw you off...
Sounds like Solidworks need to add a new feature
As an engineer with experience that I still consider humble, my first instict was wrap yes, but as soon as you showed that it extrudes with slanted walls, I immediately thought of surfaces. So I find it epic to know that I cam to the solution that quickly. Seems like 2 years of product development and DFMA work in addition to my other years of SW have done me well.
Hey nice work! it took me 11 years in Solidworks to get here so you're doing great 😉
Well executed, walking through each method with detailed and understandable rational. Does a bidirectional surface thicken work in place of using offsets?
I guess that should work fine as well! It would save a few features. Good call, I'll try it on the model and see if the result is the same
It works for Fusion 360
My exact thoughts ! Thicken-remove after the initial swept surface would save a few steps.
But really, the key is understanding that this geometry needs to be modeled from a center guide.
@@Penofhell very well put!
I love how its all this just so that a piece of computer software can convert it back into a single line for the CNC to move along lol
I had exactly this problem recently modelling a dry break connector with a similar pin fitting. So frustrating how Solidworks kept on distorting the cut no matter how you try to constrain it. As you say, programatically it doesn't matter, as it's easy to do in CAM, but I too am OCD when it comes to modelling, and it drove me insane with the groove being visually inaccurate.
I had to move on to other more pressing jobs (the dry break was just for representation within an assembly, and really no need for that level of accuracy, but I often take the opportunity to model superfluous parts to both expand my skills and produce practical assemblies/realistic renders), but I like your approach and will revisit the dry break now!
Mechanical engineering student here, with 4-5 years of prior experience with CAD before starting. A lot of my experience is in CATIA, and the easy way is to just use generative shape design. Create the cutout with a bit of projection, create the cylinder walls around it, convert to solid. I should try this tomorrow to improve my skills.
Try it! I'm interested to know if you find another viable solution
I'm very pleased with myself. I tried to think how I'd do it in Blender (because that's all I really know) and the method I came up with was pretty much the same as your final method.
I'm feeling real smart right now.
Never used Solidworks but I atleast understood what problem they were talking about. Good job.
Awesome video. Enjoy the RUclips success!
I would have arrived at the swept 3d sketch, but would have stopped after that not trying the surface approach!
Brilliant work mate!
I ran into this problem awhile back but never found a good solution. I am a machinist so it didn't matter too much, but I scratched my head on it a lot
A swepted cut with guide curves on the inside and out should do it as well. You might be able to get away with a single guide curve, but it often distorts a bit so would require testing.
You can also sweep a tool through a path that might work. I had a tricky one years ago where we were making helical cuts with an endmill in the side of a cylinder (bar). Getting the geometry of that cut right was tricky and it mattered because it was a flow passage and we needed to do CFD on it.
Surprising though, that the whole team couldn't figure this out.
Yep! This is how we did it in our Highschool CAD class back in 2012 heh. Pierce & Swept!
You have less control with a solid sweep because it needs to manage more directions/dimensions at once. The advantage of a central surface,which is the correct way to do these geometries if you need good control on the result, is that it allows you to tightly control the part, and then control the volume (thickness) in another operation, which is sometimes not possible with solid sweeps.
Nice Video and nice solution. Often something sounds easye and in the end you sit at it for hours.
Thank you, sir, for that wonderful explanation.
I don't even watch the WAN Show on RUclips. I listen to the podcast. But the algorithm still knew that I needed to see this video.
The way I visualized this in my head was just a bunch of lines projecting out from the center forming the shape of a J and then thickening to take up the diameter of the mill, I had no idea how to translate that into CAD so this was really enjoyable to watch you take this and work it out, thank you.
I imagine this may be very helpful in allot of contexts.
Sadly my brain is too small to be an engineer. But this was easy to follow along. Thanks for the explanation.
Great video! I'm a design engineer too and you're right in saying this problem seems trivial until you get deep into it. That was interesting
well done. i would’ve been stumped and not know to to describe the situation to begin even to google the instructions
Wow, first off all English is not my native language so I apologise for any mistakes in advance. So I actually made pieces like this in 2009 for a company. and although it looks real difficult to draw in 3D. It is actually very simple. You do not wrap the J form around a cilinder but you first unwrap the cilinder to be a flat surface then you place the J shape where it is supposed to go and take out the materiaal (width off the J shape) completely through the dept off the flat model. Now when you are done you wrap the flat surface back to a cilinder, and that is it. On a CNC lathe with the setup as you mentioned (ours was a DMG Sprint42) we actually programmed the line or shape we wanted to cutout on a C-plane as if it was a flat surface. I hope despite my language problem my explanation is understandable. 😉
Your English seems great! Check out my second video where I address this suggestion that was also made by others. ruclips.net/video/TAlYex2l9yI/видео.htmlsi=zrWo3We4RPLU1csg
You end up with a slot where the walls aren't parallel. It might be possible to fix that in CAM but the goal of the exercise was to have CAD that is accurate to the finished product
The RUclips algorithm man😮. I’ve been looking for a guide on how to make this mechanism in CAD. Thanks for the clear instructions.
So glad the algorithm and I could help :)
@@JamesDeBono it was extremely helpful thank you. I searched previous for bolt action and cam and groove, but none showed how to do it around a cylinder. Hopefully you’ll get 1000s more subs off the back of it.
I'm glad this is not just me having issue modeling this,
ill be trying this in Autodesk inventor professional 2024
so I did try this and was not able to replicate it yet
It seems like Inventor handles this process a bit better, to do that I would create my 2d sketch and wrap it as a 3d sketch around the cylinder, create my "tool" and sweep it along the path. I've done some pretty complex splines and it seems happy enough to do it.
The second for the angular control, the other way to do it, would be to use a dimension formula, take the Dia of the OD, turn it into a circumference, then create your angle from the circ/360, The math would look like "Start to End Line Dim=((Dim(OD)*PI)/360)*Angle", as long as it handles dimensions like I think it does all you would need to do to quickly adjust the arc would be to adjust the angle part of the formula.
Wow, I missed this Wan show but I actually had this problem a few months ago as well. It took me a while to figure out how to model it up but it looks like we went through the exact same process.
Multiple ways to do it, that's for sure. I've done similar type of paths before back in SW2012 or so when we had a lecture on pierce and cuts that follow a path. So in a similar instance of cutting curves on a round pipe for 4th axis milling, we just drew the curve ( projected it ) onto the cylinder. Then at one end, created a reference plane that was perp to the end of the line. Created a rectangle ( To represent how an endmill is gona cut it ) and performed an swept cut that follows the line. Worked perfectly and accurealy represented what the endmill cuts..Took a second to verify this in SW2023 and got the same result. I guess sometimes you just know useless stuff ( me ) as for me SW is just a tool I dork around with. I just do CNC by trade. So I use solidworks based on how I know tooling works..if that makes sense. I like the 5+ axis CNC machinist for years --> Solidworks/Solidcam --> Engineer direction VS engineering degree --> solidworks --> HowDoMachineWork?? heh
LLT's problem is their asking engineers questions only machinists can answer. James looks like youre one of the rare individuals that understand this. If you were an engineer i had to work with, if gladly work with you man.
Hey that means a lot thank you!! 🙏
Great stuff I was thinking about it when I watched the episode!
Extremely clear explanation. Great stuff.
Great explanation.
I had an issue with this once. A twist and and what not.
Create the Body your cutting out of
create the "milling bit" in Cad as a seperate Body
Rotate it up along a path to cut it. I'll dig out the two features i used when i'm back in the office.
Thats what our tech support was recomending as well.
This is what the solid sweep/solid body sweep function is for. Model the tool as a separate body, have it follow a wrapped curve.
Check out my second video where I talk more about solid sweeps ruclips.net/video/TAlYex2l9yI/видео.htmlsi=-2yq-ggusqQ0LYS4
I‘ve encountered that same problem a few weeks ago when I was designing a bayonet-lock into some part I needed to 3D print and it was extremely frustrating until I came to the same conclusion as you did!
Nice! then I feel more validated :) thanks
This is one of those problems that will create its own subreddit six months down the line.
I am not a solid works person but a 3dsMax one. It is mostly trivial to do in max. You can either do your J curve on the cylinder surface and extrude it flat toward the cylinder center, or make a ton of milling bit copies alogside the J to substract from the cylinder, or many other options. Keep in mind that you will have impresisions in anysoftware anyway so you'll have to build in tolerances in your cut anyway. Regarding your gap it is because of triangulation. A non flat quad polygon can be triangulated either concave or convex but never flat that is just math.
Try it and I think you will find it's not trivial. If you extrude the J towards the cylinder centre you will end up with slot walls that are not parallel. Making a ton of milling bit copies is also equally possible in solid works but would leave you with an equal amount of tiny curved cuts instead of a straight cut so wouldn't satisfy the end goal of having CAD accurate to the final milled piece. Lastly, there are no quads, tris or polygons of any kind in parametric CAD. It's purely vector based
I think the bigger issue is that although yeah, it's totally possible. It doesn't really look like a J because the curve makes it appear basically like a U.
It's just an example 😅
This is one of the few times I'm glad I model in rhino. I'm an architecture 3D modeler, so I've never even touched solid works. Can't understand solid object modeling. Surfaces only for this guy
i don't get the layers
am i supposed to put a sketch here
a surface there
and the resulting solid in the next?
everything in one layer?
just show me the objects
@marc_frank it's like photoshop. You get to decide what you do with the layer
Thid is why solidworks pisses me off so much. This should be so simple, but the fact that SW has an option to keep the cutting face perpendicular to the round face of the tube, but somehow still fucks it up without explanation, is just classic solidworks.
As a machinist who have spent north of 500 hours in CAD (Fusion, Inventor, NX and some SW) the past few years for toolmaking and general quality of life things around the shop, this seems like a simple operation that should be possible in most environments. Just wrap the centerline of the profile and extrude a rectangle along the path
Thats what I thought but then again I started thinking that I‘m just to stupid to understand the problem. Still don‘t know which one it is…
That's exactly what I did in the second solution and the geometry was close but not quite right
I only have a couple day's experience in Solidworks, but if I were making this in NX, I would wrap just the center line of the J shape onto the tube, then use the Law Extension tool, which basically extrudes the line, but the direction can be complex, and if I remember correctly "normal to face" is one of the options, So not I have a sheet going through the tube that is normal to the surface. Then I would thicken that sheet to the thickness of the slot that I wish to cut, and subtract that solid from the tube. Reading the comments, I see there is a simpler way, but that is just using the tools that I have experience with.
Check out my second video ruclips.net/video/TAlYex2l9yI/видео.htmlsi=-2yq-ggusqQ0LYS4
I did essentially what you are describing 😊
Also, for the first attempt you use, there is also a way that is similar to what you did the third try, that is to extrude a curved even surface. I only know how to do it in fusion 360 so I will put the workflow this way:
1. Emboss the J
2. thin extrude the J into a sheet body.
3. Cut the profile with the sheet body you just created.
4. Remove the face you just created, and loft them back together.
I think it will work, and although It did not show clip in fusion, I believe it works.
Are you suggesting doing this on a flat surface and then folding it back into a circle?
@@JamesDeBono I finished the second video right after posting this post. Thanks for addressing!
hire this man
Lol I actually own the reference pen. It's a Kara's custom. The U shape is used a lot in bolt action pens. I have no less than 2 of that style lol
This reminded me of my studies. Don't bring your non-distorted drilling bit to a distorted world. Just imagine you make everything flat again, not just the pencil, also the drilling bit.
came for the CAD, subbed for the rage shirt
Eyyyy! thanks man! nice to find another Aussie here :) at least I assume you're Aussie?
260 followers when you recorded this and 24 days later you have 1.3k, that's a pretty good jump!
Definitely! I was very surprised. It's been fun watching the subs go up :)
Grats m8
I would have tried starting with the tube as two surfaces each with the j cut out with the same width. Created the inside surface between the two Js, capped the end of the tube. And finally knitted it into a solid.
Fun video, glad youtube recommended this to me :) Small detail I noticed, and perhaps this is not important to the functionality, is that the "reference" pens from the forum post have a more angular inner corner instead of the basic round J shape, not sure if this is done for stylistic purposes, ease of manufacturing or are vital to the mechanism...
If you define the center axis of the cylinder, open options in your swept cut dialogue, and choose "Profile Twist" is "Specify Direction Vector" and define the axis as the vector it does what you're looking for.
I've tried this too and it also doesn't work. I wish I could just send a screenshot here so I don't have to make another follow up. The feature totally glitches out even more than it did when set to "tangent to adjacent faces" I made a temporary image link tinypic.host/image/UcZ3d
In reality I believe the first “solution” creates a valid geometry that can be readily machined via 4th axis CNC as long as you use a cutter that is smaller than the width of the slot. This is generally preferred so you have a more consistent finish by only climb cutting or only conventional cutting the outline, rather than full-slot cutting.
That being said, it’s worth considering whether or not you want a slot that’s tapered through the thickness, as it could affect the friction/feel of the bolt-action.
you certainly don't want a tapered slot, the sharp edge will cut into the bolt over time.
when performing the machining operation you specify the edge you want the end mill to follow, so it doesn't matter if it's tapered or not; you can create an offset in the toolpath with stock to leave (which can be a positive or negative value). while this is a "hacky" way to approach it, it doesn't really matter for the end product
@@Pfooh Maybe add a back-side chamfer?
Great work. And also detailed explanation on how you solved the issue. I would probably be to lazy and go "eeh as long as the outside parameter is correct then in the CAM software i will use the outside to set the toolpath
That's fair but if you want to try a 3D print proto or if you want the model to look correct it's nice to make the model perfect. It's definitely not always needed though I just can't help myself 😅
I will admit that I can get frustrated to if something like this happens to me. But it all depends what the customer want. :)
Depending on the cam software you can Export a 3D model of the finished machining simulation. Then clean up the mesh; Not the prettiest method but would work.
Chess Champ!
Beautiful work, sir.
To get such a simple J shape it can be done with 3 extrude-cuts and 4 full face fillets.
At 6:00 instead of offset+loft+knit you could use thicken-cut or thicken+(fillets)+subtract combine.
For this very specific shape maybe, but if you need to make some non-cylindrical shapes and in general anything else than just a straight line and a radius, surfacing is the more flexible tool. The goal of that video I think was to show a foundational method to get more complex results, not one that can barely achieve the result shown here.
There is no need to build a closed surface and knit it though, you can work from the simple center surface and do everything from there with a thicken-remove feature and some radiuses for the end.
This. You're both right. It could be done in this case with extrude but it the j shape changes it could easily become much harder without the surface method
@@JamesDeBono I watched your recent video. Prior to commenting I tried this in Fusion360 and it worked. So I assumed Solidworks can do it too. In Fusion360 it can be spread to about 140° before full face fillet starts to fail. Today I tried in it solidworks and it worked only up to about 45°. I don't know why.
Out of interest I played with Extrude-Bosses instead of cuts (and then subtract combine), and I can make face fillet work up to 140°. But it is very weird and finicky. More problems than it is worth.
In regard to Solid profile cut sweep: something that could be helpful is to make 3d sketch-> convert entities (choose all those lines) -> Fit Spline. This creates one continuous curve, that might work. There is some "error" in the shape of that curve, but you can control that with Tolerance slider to minimize it below micron.
@@primoz6578 oh nice investigation work. Interesting to know the angular limitation of the face fillet feature! good work! It's unfortunate that there isn't one clearly perfect CAD program haha. They all have their issues and they are all expensive!
The fit spline is a nice idea indeed but I think it bugs me knowing that it's not quite perfect. Also would it update if you changed the dimensions?
@@JamesDeBono It does update if you choose Constrained option. As long as converted entities don't break and stay connected in continuous line, it should work.
" it's not quite perfect." No it is not, but you can think that it is better? Fit spline gives you more curvature continuous path. With straight line and arc there is an ugly point where "radius" instantly changes from "infinite" to real value. It is sudden jump, infinite acceleration, jerk, vibrations ... Gradual change in "radius" of curvature in Fitted spline would provide smoother path. Instant improvement of the design. After all, you have used line and arc because that is the easiest shape not necessarily the most perfect shape.
A bit of different point of view for your inner perfectionist.
Also use at looking control sketches being the initial sweep path
we had a similar problem (making mold for handbag) in our company we tryed for day , a the end we switched for Rhino 3d , so far the best decision we made ,its give so much freedom in the design process
I was told that using solids and bodies to model is like using Window/Mac GUI to use a computer, easy/good enough for most situations. Once you start using surfaces and manipulating faces you're now using the command prompt to navigate the computer, which is needed for a problem like this (complex lofts).
That's an interesting comparison! I can see what you mean. I definitely agree in that sometimes it's necessary to go to the surface level. I always aim to have the least amount of features needed to make something accurate and stable if things need to be edited. I think surfaces are the right choice in this case
If you make a plane perpendicular to the center of the radius you can draw your "j" and extrude it into the pen body ( do not let it combine bodies) then you can use the remove overlapping bodies tool and the slot where the two overlap will be removed. then hide the non overlapped "j" projection.
There are some really nice bolt action type pens out there
Dan the man is bound to have seen this by now?
What happens if instead of offsetting the 2 surfaces on both sides, you use thicken (midplane) for the surface? Does it still distort?
This is the content I want to see
This looks pretty simple on something like cadquery or openscad. In fact I might just give it a try.
Try it and let me know how it goes!
Model the slot as a body, and combine cut. I use this for like 90% of what I do in CAD. You can also fo things like copy the surface,.offset it, extrude to the offset and then comvine cut, so the bottom of the cut follows the contures of the main surface
I did model the slot and combine cut it! haha
Funny. I was just modeling Bolt Carrier cam path and I ran into same problem.
If you're like me you probably never imagined it would be as complicated as it is haha
I have money on the LTT team trying to revolve their way around this problem
Amazing Explanation
In fusion360 I've seen others use the CAM functions to make weird geometry like that, altering the model but I would rather be able to do it in the CAD side of things.
Getting back into SolidWorks after being away from it for years, this reminds me of having bosses that just wanted the design out quick rather than 100% correct as long as the machinist knew what to do. Bugged the hell out of me some times when I could see the interferences like that.
I feel the same way, I understand it in a way, it doesn't matter to them but I feel the urge to make it right! Sometimes you have to let it slide for the sake of time but it's nice when you do something off the clock that you can make it perfect :)
Excellent. Maybe I will try this in Fusion.
I cant understand why the second option didn't work, it doesn't make sense. I ended up using the sheet metal tood of Fusion to do this, because I gad the same problem with the sweep, but it also colided, but when making a simulation with colisions activated, the pin went smooth, so, I'm not shure what to belive.
I work at a cnc manufacturing plant in the US. I wish you were an engineer at my company. Ours suck. Lol.
7:00 testing bit should be a little bigger in diameter and then if collisions be rectangle than geometry is proper.
Good idea, I will try that to verify 👍
Nice ❤
Sketchup user here, this is hilarious on how complicated this is when it could be as trivial as project intersect cut in Sketchup to get that result.
That's what's interesting about this problem. It appears like there is a simple solution bit when you try, it doesn't work. Check my latest video to see why ruclips.net/video/TAlYex2l9yI/видео.htmlsi=-2yq-ggusqQ0LYS4
I can almost guarantee the a project will not give you as accurate dimensions of a wrap and extruding it will give you slot edges that are not parallel to each other while also being perpendicular to the cebtelin of the cylinder. Its not as trivial as it seems
@@JamesDeBono Sketchup isn't programmatically, it's pure geometry so making a shape and intersecting it cuts out everything. Using follow me tool all dimensions are kept, in fact it does not work on it's own as the angle change causes the approach angle to change so you have to use a constrained follow me plugin for that. Once you shove the two pieces into each other and cut you get a mostly clean result, but of course with the downside of not being a cad path instruction and just geometry.
The difference is in the tool for extruding the cut part, which tries to maintain dimensions, because that's what you need the most in archviz. Of course you could not use that model to program a cnc.
That’s why you 3D print it 😂😂
So there is no sweep in solid work? Wrap a sketch and sweep along that line. Then just extruded the ends for the circle bit. Yes the circle won't be perfectly normal to the surface but so won't the drill (this obviously won't work if you require multiple passes of an endmill, but then again you can do it in two parts, drilling the holes and then the rest)
Edit, I just watched the rest, I see now .
I wonder if you couldnt just create a surface by wrapping the J on the outside, then wrapping a J scaled to maintain the width of the outer J (ie scale it back to the same width as the outer J to avoid the travezoidal shape of the naive pushed extrusion).
But, i cant be bothered checking if that geometrically works, so your solution is it :)
Luckily I just uploaded another video where I tried something like that! ruclips.net/video/TAlYex2l9yI/видео.htmlsi=-2yq-ggusqQ0LYS4
So in normal people words, the sizing of the cut needs to be modeled to scale to the change in radius between the inside wall and the outside wall? And the “normal” way of doing this in Solidworks makes it that it scales linearly, meaning the inside wall and the outside wall have different sizes of cut?
Precisely!
I understood nothing here, but it looks like a great solution
I personally would have wrapped a sketch from outside like you did first, then do it again from the inside of the part, then loft cut between the two sketches. Though, I've heard that wrapping vs projecting, there are pros and cons. This is from a primarily Inventor user though so... 🤷♂️
Check out my second video where I address this suggestion that was also made by others. ruclips.net/video/TAlYex2l9yI/видео.htmlsi=zrWo3We4RPLU1csg the problem is that the inside sketch is impossible to determine from the outside sketch. The outer sketch has a circular arc but the inner needs to have an elliptical arc to match it but there is no way to perfectly determine the shape of that elliptical arc
Would just like to point out something about the wrap feature for the sketch. If your sketch for the channel is lets say 0.5" The projection on the surface is 0.5" for the ARC LENGTH, not the CHORD LENGTH. you'd need to do some math on the sketch to get the chord length, which is the width of the bit, to be correct.
That's true but since we only use the center line in the last solution anyway it doesn't affect the end result
@@JamesDeBono It was a little tricky, but I ended up using math/geometry and wrap projected two sketches, one on the outside and one on the inside then did a surface blend between both surfaces for perfect perpendicular walls.
Would be nicer if we just had a sweep volume feature like some other CAD software
@@neondreams2337 There is Swept-Cut with body but it struggles to follow the wrapped path
I know less than nothing abut solidworks but to me this seems like a resolution problem on curved faces where a quad should be a theoretical infinite number of 4 tris going in 4 on each face. This comes form me using blender for many years where you have much better fine control of mesh geometry and vertex mergers into quad faces.
There are no quads, tris or resolution in parametric cad. Only maths :)
Does it really matter though. Even if it wouldn't be accurate in CAD it would still be accurate in the product.
It matters if you want to do a 3d print prototype or an accurate render. Or just to satisfy OCD tendencies 😅
Good video explanation of the problem. Your final solution does work, but the requirement to have such pipeline to get to this point is terrifying - at any change it would be double and triple checking a lot of things to make sure it's still valid. The second approach should be the valid one, but it makes no sense that the edges are cutting into the bit towards the center and not evenly or the other way around, and I would check manually distances there on a cross section just to have some knowledge what is happening, and you didn't show that on a video.
Very interesting. Could this possibly be something solid works could work towards? Implementing some kind of tool to allow this sort of cut to be made more easily?
It actually does allow this with swept bodies it just didn't manage to interpret the path as being continuous so it errored out
About 2 years ago just started working as a engineer after graduation and needed to model something very similar and the product even ended up being patented by my company. It literally took me like 10 minutes of googling to find a solution for Siemens NX which has an even smaller community an RUclips etc. What are they doing at LTT 😅
In ptc Creo Parametric you could have createt Points and then a spline through them. After that just the sweeping.
You can also do this in Solidworks. Check out my second video where I address this suggestion that was also made by others. ruclips.net/video/TAlYex2l9yI/видео.htmlsi=zrWo3We4RPLU1csg but then you don't have perfect circular geometry and the swept body still doesn't like the spline
How about, unwrap the profile into a sheet metal, do the cut and re-wrap the metal?I think in that way it will also generate even offsets.
Check out my second video where I address this suggestion that was also made by others. ruclips.net/video/TAlYex2l9yI/видео.htmlsi=zrWo3We4RPLU1csg the issue is when you fold it back into a cylinder the walls of the slot will not be parallel
I only have experience in AutoCad (I know it is not very good in 3D modelling). But when I model there, I would draw the cylinder and a plane surface of the "J". Then I would extrude the J, so it becomes a volume. And after that I would drag the "J" into the cylinder and do a volumetric substraction. I would assume, that this would work for that kind of Problem 🤔
I don't know if that would be a solution for Solidworks tho.
Check out my second video where I address this suggestion that was also made by others. ruclips.net/video/TAlYex2l9yI/видео.htmlsi=zrWo3We4RPLU1csg the issue is when you fold it back into a cylinder the walls of the slot will not be parallel
Do more of these
name is bono...
I'm not sure if this was mentioned or not, bit in the real world nobody should and would mill a groove with a mill bit in the same diameter as the actual groove. You should always use something smaller for inner grooves and that would make all of those problems disappear
It's true that you would use a smaller bit than the slot but that doesn't make all the problems disappear. You still have the same problems you just take a longer time to cut it. The edges of the slot still need to be parallel for the mechanism to work correctly