efficient way to model and/or assemble roller chain

[Snoopy]

Forgive me if this has been asked before (I did check google):

I am looking to assemble a good length of #40 chain. 78" between sprocket
centers. It rides inside a plastic chain guide with lugs attached through
its rivets every 5 inches. I know I can just assemble it, but it would be
locked in position? what's the best way to keep the assembly flexible so
maybe I could animate it? Or is this asking too much? I suppose I could hide
a large portion on the details....

Does anyone have some comments?

Thanks


--
Jeff
Design Engineer

[kellnerp]

I have attempted something like this. It will likely bring SW to its knees
computationaly. It you assembled individual pieces it could be flexible. It
you used subassemblies you would have to set up a configuration for each
instance and this is very inefficient. The mates for going around the
sprockets would be a real trick. Perhaps you could mate the rollers to an
envelope part and make the envelope invisible in order to fake the sprocket
engagement.

There are some kinematic packages available from Solution Partners that can
simulate chain including dynamics.

Post by Snoopy
I am looking to assemble a good length of #40 chain. 78" between sprocket
centers. It rides inside a plastic chain guide with lugs attached through
its rivets every 5 inches. I know I can just assemble it, but it would be
locked in position? what's the best way to keep the assembly flexible so
maybe I could animate it? Or is this asking too much? I suppose I could
hide a large portion on the details....
Does anyone have some comments?
Thanks
--
Jeff
Design Engineer

[Sporkman]

I've also tried this, with rather mixed results. I was never able to
show motion (too much computational time), and I found that the most
efficient way to deal was to make the entire chain one Part with Feature
Patterns (both linear and circular) and with Mirrored Features. You can
also make Component Patterns for a subassembly. Regardless of how you
do it, it's mostly for show, unless you're trying to figure out sprocket
tooth position (relative rotational position). I also suggest creating
a top level configuration in which the chain is suppressed entirely to
speed up work on the rest of the assembly.

'Spork'

Post by kellnerp
I have attempted something like this. It will likely bring SW to its knees
computationaly. It you assembled individual pieces it could be flexible. It
you used subassemblies you would have to set up a configuration for each
instance and this is very inefficient. The mates for going around the
sprockets would be a real trick. Perhaps you could mate the rollers to an
envelope part and make the envelope invisible in order to fake the sprocket
engagement.
There are some kinematic packages available from Solution Partners that can
simulate chain including dynamics.

[Wayne Tiffany]

What we usually do with chains is to model the links on the ends and then
use a sketch for the length in between. That way you have correct looking
connections, but not the immense overhead of modeling the whole chain.

If you needed the positioning side of it, use an equation driven sketch to
evaluate.

WT

Post by Sporkman
I've also tried this, with rather mixed results. I was never able to
show motion (too much computational time), and I found that the most
efficient way to deal was to make the entire chain one Part with Feature
Patterns (both linear and circular) and with Mirrored Features. You can
also make Component Patterns for a subassembly. Regardless of how you
do it, it's mostly for show, unless you're trying to figure out sprocket
tooth position (relative rotational position). I also suggest creating
a top level configuration in which the chain is suppressed entirely to
speed up work on the rest of the assembly.
'Spork'

Post by kellnerp
I have attempted something like this. It will likely bring SW to its knees
computationaly. It you assembled individual pieces it could be flexible. It
you used subassemblies you would have to set up a configuration for each
instance and this is very inefficient. The mates for going around the
sprockets would be a real trick. Perhaps you could mate the rollers to an
envelope part and make the envelope invisible in order to fake the sprocket
engagement.
There are some kinematic packages available from Solution Partners that can
simulate chain including dynamics.

[Merry Owen]

You can now do a 'curve driven' pattern in a part - therefore you can create
a multi-body part and do an equal space pattern around a spline that you
have created to represent the path (this can be made up from lines and arcs
then use the 'fit spline' function.

Merry :-)

Post by Sporkman
I've also tried this, with rather mixed results. I was never able to
show motion (too much computational time), and I found that the most
efficient way to deal was to make the entire chain one Part with Feature
Patterns (both linear and circular) and with Mirrored Features. You can
also make Component Patterns for a subassembly. Regardless of how you
do it, it's mostly for show, unless you're trying to figure out sprocket
tooth position (relative rotational position). I also suggest creating
a top level configuration in which the chain is suppressed entirely to
speed up work on the rest of the assembly.
'Spork'

Post by kellnerp
I have attempted something like this. It will likely bring SW to its knees
computationaly. It you assembled individual pieces it could be flexible. It
you used subassemblies you would have to set up a configuration for each
instance and this is very inefficient. The mates for going around the
sprockets would be a real trick. Perhaps you could mate the rollers to an
envelope part and make the envelope invisible in order to fake the sprocket
engagement.
There are some kinematic packages available from Solution Partners that can
simulate chain including dynamics.