Calling on Envelope experts (part Deux - transferring info to subassemblies)

Paul started a thread recently asking how folks use Envelope parts, and
it was mentioned that they were used commonly as 'skeleton parts' .
The Envelope contains layout sketches that were used to drive the
assembly, with the benefit of being a self-contained part that would
not appear in the BOM, mass proporties, etc. Cool workaround

I have a question about folks best-practices for using envelope parts
to distribute overall design-intent to subassemblies.

I am using subassemblies more and more for design validation (to check
the real-world assembly of the object) and, of course, when motion is
involved (which seems to come frequently at our business, and is
especially relevent with this second trebuchet I am working on at home)
.. When motion of multiple components is required, subassemblies make
solidworks behave much more robustly ("in my experience")

SO, HOW DO Y'ALL TRANSFER THAT ENVELOPE DATA TO THE SUBASSEMBLIES?

I, as a general rule, never want parts in the subassemblies directly
relating to the main assembly - when it comes to top-down design, I am
a firm believer in 'one-degree of separation'.
I used to start layout sketches in the subassembly, convert entities
from the top-level layout (whetehr envelope or humble layout sketch)
into the subassebmly, then make the parts of the subassembly driven by
those new layout sketches. If, at any time, I need to edit out the
in-context stuff (for instance, when changes will no longer apply) I
only have to nix one relationship. And while I AM working, I only have
to got to one, local spot.

Yesterday, I had another one of those 'duh' moments... I could just
plant the envelope into each one of my subassemblies (or subassembly of
a subassembly), make my relations to that instance of the envelope, and
keep that one-degree of separation that I like.

Is this, in you the experience of the envelope savvy, a good way to go?
Has anyone run into downsides?

Thanks for any insight
Ed

 

Hello Ed,



Trebuchet huh.

As internet savvy as you are I'm sure that the answer is yes, but do you
have ATreb, a trebuchet simulator? If not you should look into it; just
Google "ATreb" its the first to show up.

I made a nice SW model (model as in small as well) unfortunatly it stayed
when I was laid off.



Muggs

 

Forget a simulator - I have a real one, with a six foot high axle and
a 14 foot throwing arm. SWx has been great for working out permutations
FAST and virtually. I am in awe of the foks who build good ones with
pencil and paper. Using a SWx layout sketch and assembly techniques
has allowed me to really try it out virtually and test changes before I
even stepped into the shop. I can even set up an equation to calculate
the theoretical range (so I can see how efficient I can get - treb 1.0
was about 50% efficient before an eye bolt failed and the
tuning/shooting had to stop for the day. Not too good, but it was a
prototype)
I was cocky enough to go to full scale for my first one - and frankly,
lazy (everyone recomends making a scale model first, but that would
have taken as long as it took to build the real one and I just didn't
want to take the time. I wish I could describe my neighbors faces when
I was tuning the release point in my yard with a 25% load)

I am in the process of converting my treb to a whipper-style treb,
using the same frame but a brand new arm. The guy who came up with this
was just inspired. Instead of going through about 90° of rotation this
goes through 270°, and with swx I figured out how to get about 360°
of rotatation and an extra couple of feet of drop from my
counterweight- I built a lego model to check it, and the difference is
amazing.
Oh, don't get mestarted on trebuchets - it has been a passion for six
months now and i could go on and on and on (heck, I just have... must
think about productive stuff, must think about productive stuff. Word
of warning - don't get into trebuchets unless you are prepared to be
consumed by the challenge)

 

Ed,

I do top down with envelopes using the concept of a master model (I
think that is what they call it.)

In one case I had a particularly nasty weldment. Warped plates, nothing
touching, everything slightly skewed but using square cut plates. So a
surface model of the outer boundary was made and everything in the
weldment was mated to it. Before the weldment was even made it was used
in the next level assembly to check fit in the model. And in a part
with warped (lofted) plates it was used as an inserted part to get the
loft right without even having the assembly the loft went into open. In
the top level assembly this master part probably showed a dozen times
in various incarnations. The affect on mate stability was to make this
assembly far more stable than past attempts.

 

Thanks, paul.
So if I read you correctly you dropped that envelope/master model into
the subasemlies (and even as an inserted part - interesting...) and had
stable results?
Did you use this envelope as your primary vehicle for transfering
relations between components, or did you go beyond
one-degree-of-seperation by making relations not only locally, but
having the component of the subassemblies also relate to the next level
or even the toppest level assembly (I know toppest isn't a word, but
darn it, it ought to be)?
BTW - Thanks for the response - my mind has just gotten around to
reusuing the envelope in subassenblies, and now I have to consider the
value of re-using it by even inserting it in to parts! ( a practice I
am initially suspicious of, but frankly I am initially suspicious of
everything - until I try it and it works, then I go full bore. One
nice thing about having an education in science is that testing
theories is in your DNA, and once a theory proves out you can run with
it)

 

Definately off topic now, but what the heck...
Thanks Bob for sharing your experience. I would have LOVED to be there.

Tweaking that release requires a suprising amount of magic - just a
little bend with the vise grips on your pin is all that it takes to go
from a line-drive, to a perfect 45 ° arc, to busting windshields.
However, I am a little suprised that a whipper sent a shot directly
backwards. My layout sketches show that this will be impossible, while
on a regular treb this is just doing business.
AGAIN, a WARNINIG to you all - don't even think about making a
trebuchet unless you are prepared for it to be the focus of your life.
I'm a full-bore treb-u-holic, because, even though I admit I have a
problem, I won't go to meetings because simply don't want to stop.
Ed

 

Ed - please reassure us all that its just a hobby and use inanimate
loads - unlike these guys - and you have no intention of being the
payload.

http://www.extremedreams.co.uk/human catapult/

or unfortunately I fear this might happen.

http://www.guardian.co.uk/uk_news/story/0,3604,1605521,00.html

TTFN


Jonathan

 

I had to get beyond thinking of the part as an envelope because the
"envelope" property seems to be the way assemblies handle parts
designated as envelopes. An envelop is really just a part.

At any rate the construct in my head for all this is that the master
part is just a "fixture", template or measuring device that I can drop
into a part or assembly to establish certain relations that are
parametrically related through the master part.

 

My take, with no actual verification...

It should be OK to put an instance of the the envelope in each
subassembly and have components DRIVEN from it. Probably best to make
sure that any in-context features DRIVING the envelope are in one place
only.

 

Why sketches can't be imported with a part is beyond me. I think they
just use the parasolid and keep away from the relation solver being
involved, hence the stability.

 

FROTFLMAO!

 

Hey Ed,

I tried to email you but the address in the post doesn't work. I wanted to
know what the #'s are for your treb so I can put them into ATreb and run a
simulation and see how it compairs to what you actually got.
I have a jpeg of the dimensions that I would need if your interested.
Change "home" to "comcast" in my address.

Muggs