New video: Your CAD should suck!

EvanReese

It really should (sometimes). For a long time I've been wanting to make an explainer video about a concept we use at Ovyl that we call the 3 Levels of CAD. We use it extensively to make sure everyone is putting the appropriate time and effort into each model which changes depending on how much design clarity you have. This isn't a "CAD modeling" so much as a "CAD thinking" tutorial. Hope it's helpful!

Let me know what you think? By the way, an active comments section in Youtube definitely helps with video rankings, so it's my preferred spot to chat about the ideas in the video, but here's good too.

Here's the video.
https://www.youtube.com/watch?v=v6rfw6-Yw5k

S1mon

I haven't even watched this yet, but this intro reminds me of "plan to throw one away", from Fred Brooks' Mythical Man Month.

https://wiki.c2.com/?PlanToThrowOneAway

EvanReese

S1mon said:

I haven't even watched this yet, but this intro reminds me of "plan to throw one away", from Fred Brooks' Mythical Man Month.

https://wiki.c2.com/?PlanToThrowOneAway

I've not read that one (it's in the backlog) but yes, that's a big part of the idea.

steve_shubin

@Evan_Reese

First off I got a thank you for a good laugh. That title is something 🤪😜

I’m guessing that because I’m not signed up with YouTube, that I’m not able to view the video

Here’s what I see when I tap on your link

EvanReese

@steve_shubin try now. I thought I'd made it public but hadn't 

steve_shubin

@Evan_Reese

Yeah it’s working now

I did have to go into the system on my iPhone and clear the history and website data

I just caught the first few words of the video and its already looking real interesting

bryan_lagrange

Great video @Evan_Reese

MichaelPascoe

I've found myself in each of those stages subconsciously. This thought process will help quite a bit.

Thanks Evan! 

SethF

We've talked a lot about the spectrum from quickness to rigor (though not with those exact words). Doing more one-off work, our goal for where to land on the spectrum for a given model is often based on its use more than how far along in the process we are. Configured parts we're going to use internally over and over need to be very rigorous (obviously those could follow your paradigm more - but also those often have design intent relatively pinned down). Models that are going to be built this week and shipped next week and not looked at again, less so.

MichaelPascoe

I think Onshape should include this video at the start of their learning center.

imants_smidchens

it makes me very happy to hear that my design strategy is pretty close to what you suggest. I am 100% sharing this with my team!

S1mon

Finally watched the video. Thanks for putting this out there. I'm definitely in agreement with the concept. There are a few other things I would add to this subject:

• "Level 1" is somewhat like what one might do by hand sketching in 2D. It can be used to capture design concepts and architectural (part break up) ideas. It's there to help find form and structure by lending a little reality to the concept sketches (Is there room for another wall thickness here? Can I fit 2 or 3 battery cells in this envelope, etc).
• At some point it's good to have a clear discussion with the team about where to place the parts/assemblies with respect to the origin and default datums. Maybe this is prerequisite for starting "level 2"? 
• Similarly, assuming that the database is going to have some amount of top-down design, agreeing on the starting approach to how to manage those relationships is important. It can and will change over the course of a project, but have a discussion on a white board where you draw some boxes showing what's driving what.
• Onshape is very freeing coming from other CAD systems because there is infinite micro-undo and it's so easy to dig up where you were yesterday, or last week or whatever. I find that the option-paralysis of wanting to do things the right way the first time around is not as bad with Onshape. If I want to make a change or try something I just do. I can go back and edit/massage/modify it so much more easily without fearing that I will lose work. It's really changed my ability to move forward.
• I would add some distinctions about when not just the CAD niceties happen, but also DFM thinking etc. goes in. If I'm designing something for injection molding, I try to incorporate that thinking very early, at least for any features which would change radically if I didn't start that way. I'm going to try to put in draft on tall features as early as possible, because it might screw me if I leave it until later. Some people design everything in detail as though it would just be 3D printed or machined as late as possible, even if it's going to be injection molded in production. To me this is too much technical debt.
• I get that "Level 1" can be messy, but I do worry about the "hit by a bus" problem even at that stage. Take a few moments to name parts, and put large groups of features into folders with names that help with navigation and understanding.

I could go on for hours about this...

EvanReese

Thanks, @S1mon
I think level 2 is a great spot to agree on lots of things. If we're kicking off work on a complex model with lots of subsystems, we'll have a meeting to cover orientations and major datums, mfg processes (at least a plan A) as well as what in the design should especially be designed for robustness, major variables, and to plan model structure (master studio derived into studios for subsystems, etc).

fnxf

Great video @Evan_Reese ! I have a question about the way you are moving from Exploration to Iteration within Onshape (t=348)-

- do you make a new branch and start over in a new workspace?
- do you stay within the same document, but open a new Part Studio?
- or do you make an entirely new document, and start anew there?
- or something else I haven't thought about?

EvanReese

@fnxfIn Exploration we usually have one exploration doc per person and (to use a fake project) name it something like FTG01 | Crank Toy Level 1 (Evan) or FTG01 | Crank Toy Exploratory (Evan). In those docs I more commonly just duplicate a whole part studio, rather than work on variations in branches. When we're ready to move on to level II we usually make a single document and start sharing it.