Greetings from a newbie - and a question!
-
@stilteg said:
It seems you are considering very small geometries ... You should consider building at 1000:1 scale and reducing it to real scale in the very end )
Sorry, forgot to mention that. I am building in meter scale and scaling with netfabb. So all of the measurements above would be in meters in Sketchup!
-
Same conditions apply whether metric or imperial units.
Sketchup was intentionally designed with a lower limit of about 1mm on face size in order to facilitate other functions in the program behind the curtains.
Did I understand your response?
One technique is to make the geometry or portions into components. Make a copy of the component, which becomes a component instance. Scale the copy up to work on it. Every edit you do to the big copy will be reflected in the small version. When finished, discard the big copy.
I am not familiar with netfabb. However, I assume you could still run the model through netfabb, if needed. -
@iceraptor said:
I'm trying to graduate up to more advanced designs however, and the thing that constantly frustrates me is how to make solid tank-like tracks along a curved path. Because of the constraints of the RP process, minimum wall-lengths are really important, and getting the mesh to be a solid manifold has become my bane.
I'm unclear as to exactly what you want. The sizing issue pointed out is the reason for your broken geometry. Did you want a better way to make tank tracks? Maybe this would help.
Oh, and welcome by the way!
Hope you find this place as helpful as I have. -
Thanks everyone for the offers of help - it's appreciated!
What I'm trying to accomplish is to 'weld' individual treads onto an angled surface, to represent tank treads in a water-tight model. I've a portion of my actual model to illustrate what I'm talking about.
I end up running into two problems. First, most of the time when I'm intersecting a component tread onto the track, the resultant mesh doesn't align with the tread's cross-section properly, and I have to delete the old mesh, the 'underside' of the tread, and then heal any missing nodes manually. I am pretty sure I'm missing something here, as it seems like a pretty normal operation, but I keep having issues with it.
Doing this along a curve is just infuriating as more often than not I simply have to reconstruct the better part of the object. This is where I end up getting bogged down every time I try to solve this problem.
Basically, I'd like to know if there's a more efficient way to interface these types of components onto the model, especially along the curved surfaces. Right now I'm trying to do this by moving the component into place, exploding it, and then intersecting the resultant geometry and manually editing any mesh that doesn't work out. Obviously, that's hideously tedious and any pointers would be appreciated!
@unknownuser said:
Did you want a better way to make tank tracks? Maybe this would help.
I'm looking into this, to see how I can make this work for me. Still playing with it to see if I can get my treads to align along the path properly. Thanks for the reference!
-
See this file for ideas...use the Scene tabs to follow the tutorial. Ask back here if you need more help!
-
Several other observation:
- There will be a relation with the curve length and the number of treads. The number must be an integer unless you will settle for a odd ball at the end and the spacing around the " curves" will need to avoid collision with adjacent parts. Real tank treads show this characteristic.
-
You might also- just to get a little perspective- look on youtube for constructing a wheel or tire. I was thinking if you needed this level of accuracy, using the Rotate/copy method might help solve the problem? Path segment uniformity?
-
@d12dozr said:
See this file for ideas...use the Scene tabs to follow the tutorial. Ask back here if you need more help!
Thanks! I'll take a peek when I get home. Much appreciated!
@mac1 said:
Several other observation:
The number must be an integer unless you will settle for a odd ball at the end and the spacing around the " curves" will need to avoid collision with adjacent parts. Real tank treads show this characteristic.Thanks for this information mac1. I suppose then this means that if my curve radius is 2m, then 1/2 of the curve will have a circumference of 0.5 * pi * 4 = 6.28318531. So I need to have multiples of tracks in the 6 (1m tread) / 12 (0.5m tread) / 18 (0.33m tread) ratio?
@mitcorb said:
You might also- just to get a little perspective- look on youtube for constructing a wheel or tire. I was thinking if you needed this level of accuracy, using the Rotate/copy method might help solve the problem? Path segment uniformity?
I'll take a look and see what I can find. I've been doing copies of the components, but the interfacing of those components down onto the various surfaces of the tank tracks has been that's caused me grief. Thanks for the tip, though!
-
One more tip, keep all of your geometry as groups/components, making sure they are 'solid' as you go. After you have everything placed where you want it, make a copy (in case you need to go back and edit something), then use the Outer Shell command to make a single watertight model.
-
New math huh? c=2pir or c/2 =pi*r
Sorry I mislead you. I was referring to the total track length. For the curves( or at least one of them) I think they will establish your min spacing to avoid collision around it. So it seems to me the curve will establish min spacing and then you can determine the number of treads based on total length unless they have some way to compensate for that.
I can not work on this today but will see what ideas I have this PM
Can you do me a favor and post your model on the 3d warehouse. I have SU7 so have to go through some hand springs to open your SU8 and I hope it does not introduce errors that are leading me to the wrong conclusion or solution.
Thanks -
Here is my first attempt but not happy with this. I redrew the curve for the track with increased arc segments, used the entity info to get the total length. I then calculated the required spacing for some guessed number of track segments until I got to the point of having a reasonable spacing of about 3"( designers choice at this point. spacing = (L-0.15*n)/(n-1). This gave me an answer of 60 segments spaded at .075524m.( 2.97") I laid the track out on linear basis and then laid the track segments on your curve. Unfortunately this gets into a modelling area SU does very poor on, rotating a line to intersect another line at points not a vertex. So for the last segment this resulted in the collision of two segments so had to erase one and change that area spacing a small amount.
Have some ideas on how to do more accurately but that is for tomorrow. http://sketchup.google.com/3dwarehouse/details?mid=198d0f849ff6fc9bcf67f67fff492c0e -
IceRaptor
Tank tread update http://sketchup.google.com/3dwarehouse/details?mid=6359edbb42f97c5d4016a09edf519cd
This model was created much the same as my above post except I used the component stringer plugin. The total tread profile was subdivided using the plugin poly-line segmentor into 60 segments per my above calculation and then the stringer plugin used. Note the stringer plugin does not work on a closed curve so the last one must be placed manually
Look very closely and there is still a slight collision with some piceses( two I think) at the "lug" protrusion on the corner. I will leave it up to you on how you want to solve that issue, wider spacing, lug redesign etc
note I entered the worng length in above rquation of 14.45591m vs sb 13.45591 and gives a cal spacing of .057017 and the design is .056066
update 30 Apr 2011 0919 MST
I did not select the correct scale option and used uniform vs none. Will work on correction. If you want to use in the mean time select each tread and do a reset scale. There are 60 so a short cut key is in order if you can find it, I could not yetUpdate 30 APR 2011 1040 MST model now corrected.
-
mac1,
I wanted to come back and thank you for all of the time you put into helping me. With your examples and the component stringer, plus a bit of planning, I was able to get the effect that I was looking for. Thank you very, very much!
Advertisement