Dealing with lots and lots of holes- any relief in sight?

edward_petrillo

Our robotics team just finished a model of a robot frame comprised largely of 8020 extrusion.  The main parts studio holds 57 parts and lots of counterbored holes for the fasteners that hold the structure together.  The final result is a parts studio wherein the holes take longer to regenerate than the parts themselves, and the fully featured parts studio is so sluggish that it is nearly impossible to work with.  This situation leads to a catch-22; suppressing the holes to speed performance makes it impossible to create the drawings that are needed for fabrication. At some point in the full design process, there is no way to avoid working with a fully-featured parts studio to finish the overall design. 

Trying to tease out ways to build a feature tree that minimizes the negative impact of hole regeneration takes time and energy that is better spent on actual design.  Are there any transformative improvements regarding hole creation on the horizon?

philip_thomas

@edward_petrillo - Bad news and Good news.
The Bad news is that you have fallen down the slippery slope of "i will just build one more part in this Part Studio".
A Part Studio is for making parts - in general just ONE part. If a second part is DIRECTLY related to the first, then by all means build that second part in the Part Studio, but know now that the any regeneration of the first part will also rebuild the second part and so on. Extrapolating this to many parts, and you now have very long rebuilds.

What's the Good news?
There is a much better way

A typical robot should be built using any number of the top-down workflows within Onshape.
You will likely have a Part Studio with nothing but the layout sketch.
The layout sketch is then derived into multiple other Part Studios to allow the creation of ONE OR TWO parts in each of the derived Part Studios.
The parts from those Part Studios are then assembled in an Assembly.

Here are a couple of good resources;

First a very simple example of a frame built a master sketch/top-down workflow

https://cad.onshape.com/documents/590a2ea57be40f0ffec8ebee/w/83a9ff55ae7c041eaf0383f9/e/433b56a2783cde8768e0256f

(There is also included in that Document, an excellent example of how to use explicit mate connectors to mate entire sub-assemblies with a single mate)

The second resource i am providing for you, is a tech briefing from the learning center written by one of our top CAD stud muffins, in it you will find countless tips on how to manage growing documents.
I will draw your attention to the section relating to how the structure of a document will change over time (very lose at first and then broken into multiple documents later) - search for the text

"Throughout the design process it is likely that a number of the following will be true:"

here is the Tech briefing - 

https://learn.onshape.com/learn/article/product-structure-organization-tips

Here are all the tech briefings - 

https://learn.onshape.com/learn/dashboard?labels=["Knowledge Base"]&values=["Technical Briefings"]

The Top Down one is also very good.

So that is the Good news - You can do it, it does work, all the information you need is cited here

Go First Robotic Team!

edward_petrillo

@philip_thomas Thanks for a very cogent explanation.  I've employed nearly all of those stratagems to varying degrees, but maintaining rigid discipline is difficult and it's easy to fall off the wagon.   Chiral (mirror image) parts are a particular challenge- mirrors are only possible in a parts studio.  I haven't found a scheme for separating a framework across a mirror plane into a left and right half, each half in its own PS.  More than two-thirds of the parts in my example are mirrored, and therein lies the incentive to grow the PS too far.  Any insights on managing mirror symmetry out there?
I'm inclined to create a initial design "any which way", and, once it is possible to see the result as a completed whole, to double back and follow a top-down approach to create a final working model.  More time spent building feature trees, but less time spent waiting for regeneration.
 

MBartlett21

@edward_petrillo

What you could do is derive the part in another part studio, mirror it, then delete the derived part.

edward_petrillo

@mbartlett21, Alas, the sequence < Derive, mirror, delete derive>  fails.  The mirror has no reference once the derive is gone.  

See my post in assembly mirror improvement request thread

michael_mcclain

edward_petrillo said:

@mbartlett21, Alas, the sequence < Derive, mirror, delete derive>  fails.  The mirror has no reference once the derive is gone.  

See my post in assembly mirror improvement request thread

You are correct, but I believe he was referring to the delete part command.

philip_thomas

@edward_petrillo - i think he meant delete the (derived) part (after the mirror) and not delete the derive.

edward_petrillo

<forehead slap> Much better!

MBartlett21

philip_thomas said:

@edward_petrillo - i think he meant delete the (derived) part (after the mirror) and not delete the derive.

@philip_thomas
@edward_petrillo

Yes i did mean delete the part using the Delete part command