Various interesting snippets from my work at Onshape. Essentially these are single-take rehearsal sessions as part of presentation prep, practice, or just jamming with Onshape. I won't always provide much context to the videos, though they tend to focus on a couple of key aspects and therefore might provide tips/tricks, background, or (hopefully) some entertainment.
In fact it works on all history states that are automatically created every time you do anything. This is the superpower of Onshape, you don't need to explicitly "save" anything. All of the actions you take, edits to a dimension, hiding of an instance, etc etc are transactions to the database. We keep them all.
This is fantastic! I'm currently designing an RC Fokker DViii so I'll definitely try this out . I'm planning on using spar web combs rather than individual webs, any tips on modelling these parametrically might be of interest to others as well as me?
If anyone is interested in the model I build, the doc (public version) is: cad.onshape.com/documents/bdfbbe1b67a9dbde2cacc66b/w/31dbaf9f62cc1ee7d955d808/e/a5219394c2a634aeffcb6199 I did fiddle with it a bit since I recorded the video (fixed a few deliberate errors, etc) but I also enhanced it more rigorously handle specifications such as string width at the nut/bridge. If you do copy it and experiment with it (and like it!) then please feel free to tell people where you found it....
I didn’t quite understand this feature the first time seeing it. But what it looks like is whilst you’re in this repair manager state, changing the ‘broken thing’ changes the downstream appearances of that thing to what you replaced it with?
Onshape has had for some time now a PCB Studio, which is a collaboration tool between ECAD tools (like Altium, Cadence, Eagle, KiCAD etc.) It is *not* a PCB layout tool in itself, it facilitates the collaboration/co-design between teh MCAD and ECAD teams.
Thanks! I have a big backlog of guitar-related things - amps, guitars, effects, midi controllers. I know I will never get to them all, but will try to share things when I can!
Yeah a labor of love for sure - but through the course of it I've managed to test and experiment with a lot of new and upcoming Onshape features. I'm thinking about refactoring it again...
your videos are the most helpful out of what I could find so far! extremely underrated :) could you make a video that goes more in depth on how to strategically place curves when modeling objects (e.g. phone case, game console, car hood, etc.) Whenever I try to surface model complex shapes, I often end up with strange, unwanted creases in certain areas. This happens because I'm not sure how to properly set up the curves or which tool to use-whether to go for a boundary surface or a loft. Seeing it visually and understanding the reasoning behind it would help a lot!
A really good comment - thanks for the suggestion. It is pretty hard to generalize completely the approaches. There are some nice guides (especially the Alias Golden Rules -- this one has been duplicated in various systems but the approaches are all equivalent) but it's going to depend on the part you are designing. A toaster, a cell phone, a car will have different challenges. There are nice "patterns" (approaches) you can utilize to tackle specific challenges - the topology you end up creating will make or break the design. I'll think more on this and see what I can come up with.
Yes there is a world of difference between drawing and designing. For many people this is what CAD still means, and that's fine: Computer Aided Drawing, or Computer Aided Documentation. However there is clearly more to Computer Aided Design... On a side note, it is 2024 (last time I checked), and isn't everything Computer Aided?? Perhaps we should upgrade our acronyms?
why did you use Move Face to finish the fuel tube instead of extrude? same result, but any other benefits of one over the other in that specific situation?
They will/can give different results as they are doing quite different things topologically. In my case an extrude would have worked, since the underlying sweep path was tangent at the ends. However in different situations, such as a simple arc, the move face and extrude will look quite different... The extrude will be by definition a normal direction from the end face, while the move face will keep the curvature of the side faces (thereby extending the arc) Here's a quick example cad.onshape.com/documents/b9b52e11d7266e5ed655ab50/w/77611edca52037d2e35267c5/e/9ed093d8a42d4ae0dbb4e479
about 30 years ago, I watched an interview of the NASA Administrator. The subject was the problem of Math and Science Education in America. He basically said that if, in the 1960s, the entirety of his high-school education was centered around the Chevrolet 350 cubic inch V-8 engine, he himself would have paid more attention to his teachers, who, so often droned on about abstract concepts. I myself have owned several O.S.Max 46 motors (in previous decades), and I would pay for a course that takes me, start-to-finish, through this model, stopping to discuss in-context editing.... and named positions, how to organize the project, and all the sort of tips-and-tricks (why do it this way instead of that way, etc). I think a comprehensive approach is what is missing in OnShape education right now (everything is so compartmentalized it seems). Seriously, people would pay money for such a course, above and beyond all the stuff available on the OnShape Learning Center.
I just made a public version available for you! cad.onshape.com/documents/117ebf5924e35f3871c8e6d5/w/6ddb0cd91f2377e06e55ada6/e/665c31713a0909595ff09e24 There are some linked docs that you won't be able to open, but they are just to standard componentry that I don't really want to make public just yet. I think you'll find the Crankcase Part studio the most interesting anyway. If there are confusing things in there, please drop me another comment and I'll be happy to (try to) explain PS in the version I provided the carburetor is not complete. There is a neat sliding/rotating aspect which is only in my private version. I'll hopefully get around to updating this in the public one. (It took a bit of disassembly of the actual OS MAX-46FX which is sitting on my desk here to understand how it works!!)
Tricky, but cool. Thanks Greg. This one was really helpful. I am still looking forward to your forthcoming video on how to plan/organize your thoughts when setting up the model airliner project.
Hey Greg, I like the method you show building the engine from the inside out. I have a question for named positions. After mating your assembly for motion how do you create the named positions with the parts at top and then bottom dead center?
I remember Aeroflyte Greg! South Aussie too. Love your vids. I will move to Onshape one day soon I think. Cloud is still a big issue for my work scenario though. I work freelance in the UK: primarily at film studios with appalling internet connections; sometimes overseas in 2nd world countries with rubbish internet; even at home, it has taken 6 weeks for BT, Openreach and now a different provider to get a landline connection. Shocking. Starling is probably one route I could take to change this sitch. Ok. Now sitting down to enjoy the vid. :)
Great to hear from a fellow South Aussie. There's a lot I miss about the place, having been overseas for decade(s) now(!) There are people in all sorts of out of the way places using various means to get onto the internet, so I'm sure something will work out for you.
Greg. At the end of this wonderful video, you mention something about in the future describing how you plan these projects using a Conceptual Parts Studio, then a Preliminary Parts Studio, then a Detailed Parts Studio. I've learned so much from you about how to approach design problems paramentrically. I'm very much looking forward to this follow-on video.
Thanks for the comment, and the reminder :) I'll get to that for sure - I'l love to work in this full time, but I do have another job in the organization! But I realize (and that's why I'm doing these) by speaking to many customers and prospects on a daily basis that there is still a lot to teach and learn from both sides.
Please, is there way how to flatten the boat parts in onshape? It would be very helpfull to get shape for stitch and glue plywood boat building method. Thank you.
The message I get from this video is that the standard Onshape tools for producing your primary curves, produce curves that are overly complicated for use as primary curves. But you can solve this issue by becoming an expert in FeatureScript, then you can write your own tools that behave as you might expect.
While that's not the message I was trying to portray, I apologize if that's what came across to you. For the vast majority of mechanical CAD users, the topic I was covering is not so relevant and the complication of projected or intersected curve may not indeed be of any downstream consequence. However as Onshape grows, and the customer base expands, there *are* indeed people to whom we are talking, learning from, and reacting to, that this topic is of great importance. So: 1. it's important to understand that the foundations of Onshape do not preclude such advanced modeling. Under the covers is a fully capable NURBS modeler. It is up to us to decide how and when to present this via features in the UI. 2. FeatureScript is a huge differentiator for Onshape, in that it is the language we have developed ourselves to create the features we ship as part of the standard package. If anyone is interested, the standard library is publicly viewable. If you want to see how we use FeatureScript to deliver things like loft or patterns, or in fact anything in the part studio, then you can do so. The important aspect here is that it allows us to prototype new approaches on top of the foundational capabilities, for example the approximateSpline function has been in the std library for some time now, and opCreateBSplineCurve even longer. With the latter you can completely take control over everything in the NURBS curve you are building. 3. for advanced users, understanding FeatureScript is extremely useful (and I recommend it!) as it helps to know what the part studio is doing. This can lead to a better appreciation of how to approach regular modeling tasks, optimize for performance, and more.. Even if you don't intend to write you own custom features. So I'd prefer the take away to be much more optimistic: that we are rapidly building an extremely powerful application (Onshape) which is applicable to more and more users and use-cases. Along the way, and before we commit to shipping certain things, we are able to experiment, collaborate, iterate, and improve. There is ample evidence of this in all aspects of Onshape, and particularly with surface tools. The last two years have seen unprecedented expansion across wireframe/surface modeling, & diagnostics, with much more to come.
@@gregbrown-onshape7555 Thanks Greg for your considered reply, I appreciate the effort. I should explain I've been using Onshape for about 5 years, and have been a subscriber to your channel for ages (not sure exactly, but have learnt a lot from your videos). I absolutely love the product, and you'll get no argument from me on the unique extensibility offered by FeatureScript. Normally, I find the message in your videos clear and straightforward (I especially like your face-blend tutorial, clear, concise and comprehensive), but I found this one troubling, even after re-watching. Anyway, I won't waste your time by labouring the point, just wanted to say thank you for taking the time to reply. Cheers
I was so excited to see all these awesome features, but a little sad when I saw quite a few of them aren't public yet. But in due time, I hope. You're pushing surfacing in Onshape in directions that I'm really liking! I'm glad that you're part of the Onshape team. I'm really loving Onshape's availability of Feature Scripts. I haven't made any myself, but they make it so easy for users to add extra functionality to the package really easily.
Thanks for the comment! I'll try to make things public when I believe they are ready and robust. Some are undergoing rapid iterations still, so I'd prefer them to settle down a bit first. I'm always thinking about how Custom features can be made more discoverable, ingestible, manageable. Believe me (it's probably clear) that this is a personal interest topic of mine... and I do follow closely the tickets, improvement requests and comments on forums and elsewhere related to this.
Love it! I've definitely run into issues with projected curves causing annoying ripples in my lofts before. The ability to create good quality base curves should be native Onshape functionality so I hope some of this makes it into the official tools (especially the approximation option for projected curves)!
Thanks for the note - yes I guess this demonstrates the power of custom features: that users or developers (internal or external) can experiment, get early feedback, iterate... and make decisions about whether it lives on as a custom feature (which is a perfectly valid outcome) or becomes part of the native code base.
Awesome walkthrough of you tools. It seem like the approximation should really be an option in the "native" projected curve tool! You have mentioned making these tools public, I'd definitely like to try these out as I have run into projected curves causing less than ideal downstream surfaces as I hadn't realized the projected curve would "destroy" the hard work that went into creating clean base curves!
Yes that is my intent - I'll get a couple more sets of eyes on what I've done and then release it in early form. If it is making sense to people I'll "publish" it for easier discoverability. The choice of whether something belongs in the native tools versus remaining as a custom feature is a bit nuanced. The voice the customer plays a bit part of course, so let's see!!
Awesome Greg! Looking forward to using these. Tweaking a bridging curve after the fact seems a great use case for these. Another use case: Last week I created a temporary Loft, created an isocurve off of it ( putting this brand new feature to good use) to then tweak that isocurve to create a guide for a second loft so I could exert some control on it, but make sure this was based on the natural curvature of the loft. What would be the limit of the ‘reduce bezier’ functionality (how many control points does it remove? And why not always use the more comprehensive ‘approximate bezier’ functionality?
I've been messing around experimenting with this sort of thing for some time, and only now decided to show my work ;) One of my use cases for tweaking a curve, is when building from side/front/top views. The side view can start as a nice clean Bezier on the plane, but then you might want to tweak it from the top/front view. Alternatively if you build clean curves on both side and top, then the Approximated projection can give you a clean curve directly. And then you can further tweak it. Perhaps after Elevating or Reducing it. The point being: I wanted to make myself a toolset where all of these are possible. After all, a hammer, chisel, screwdriver, spanner, hacksaw... etc all have places in a toolbox. Nice to hear you are putting isoparametric curves to good use already. I had one more example to show that took my Approx projection, built boundary surfaces from it, then create isoparametric curves which I used to split the boundary surfs.... and so on and so on. But this was going to make my video too long. If you're interested, let me know. The Reduce feature currently reduces degree by one. You could repeat the feature as many times as you like I guess, until you are down to a degree 1 line... I coded the Elevate to take a target degree as an input parameter, so I guess I could do the same for Reduce, though my own use case was 99% just to bump it down by one. Finally (sorry for such a long reply!) what is the 'approximate Bezier' you mention? We do have (and this is exactly how I implemented the Reapproximation features already described/demoed) an approximateSpline function in FeatureScript (cad.onshape.com/FsDoc/library.html#approximateSpline-Context-map) but this is not Bezier specific.
Yes! Here it is: cad.onshape.com/documents/1bda657a1041f59966643bca/w/ab4191fe18ac7ce49974bfd7/e/d22fe898a6d0bffdcb16564c Note there are two part studios - I just created the second one as we added something last year (in the Sept 1st, 2023 release 1.169) that might be worth looking at. Rather than configuring the text directly (you can see that in the config table for the first part studio), we added a very powerful capability across the product to "convert to expression" (for boolean parameters in features, or text) Thus I could use an expression for the text contents. In this case I used toString(#VALUE) to convert the variable called #VALUE into a text string! For those playing along, yes this means you can open it up to very fun and sophisticated expressions pulling on the full power of FeatureScript. You can encode logic into the text string, just for starters...
Thank you for showing how to use the parametric capabilities of Onshape in a real project. It was most interesting to see the way you linked all the features together with the variables and measured values. Well Done !! Especially your incremental testing....
Greg. Thank you for this. One thing to make this really useful would be a variable to control the wing twist (a.k.a. washout) about the 25% chord point. Is that something perhaps you'd consider addressing? It would make all the difference in the world for me.
Great idea - in fact I've made a bunch of updates and now the wing is looking pretty good. I just experimented with washout and looks like it fit right into my process. I'll cut an updated video tonight.
Here is a beta version, I'll clean it up further and publish shortly - as mentioned, it is pretty simple stuff, but really useful: cad.onshape.com/documents/20afd0071b1fd882855e31f3/v/a861552f5b8dcff963345bfe/e/fc706ce0880ab89b509732af
I was speechless watching your amazing work. As a violin teacher who is currently trying to create violin templates by using Onshape, I'm wondering how I can learn new things from your work.🔥
Thank you so much for the comment. I started playing the violin in elementary school and have been fascinated ever since. Various school projects, university projects, and now tinkering around for work I'm still hooked. If you have any questions or follow ups for your own work I'll be happy to try and help.
Learned a lot in this one. Thanks. But would love to see all the nuances of the circular pattern steps because I can't get that aspect to work when I try to recreate this.
Specifically, I can't get the circular pattern of scallops to adjust correctly and move in synchronicity when the number of sides variable changes. Seems like it might involve a constraints issue, but for life of me I can't figure it out.
You have really helped elevate my modeling skills. Huge thanks man. I have 2 products being injection molded as we speak. Only started modeling last October. Thanks a ton man.
Thanks for the comment. The Face blend is designed primarily for "smooth blending" between faces, but in theory there is no reason why it could not support a convex profile. We've implemented "chamfer" profile (linear blend) so one might argue this is not exactly "smooth blending" ;)
Thanks for the comment. We are certainly working on many fronts, and have an incredibly deep bench of VR/AR/MR technology here at PTC that spans many product lines and includes many acquisitions. There's a photo of me somewhere demoing in DIVISION/MockUp way before RU-vid even existed...
