Hi Pilou,

> If you select one arsc the sweep 2 rails works
> If you select the other arc the sweep 2 rails works
>
> If you join the 2 arcs the sweep 2 Rail works
>
> If you select the 2 arcs (not joined) the Sweep 2 rails don't works

Yup, and that's all normal... Anthony, when you do a Sweep each particular "station" along the sweep should be made up of 1 curve - if you have multiple pieces then use the Join command first to join them together into one logical curve so that the sweeper will know what the profile is supposed to be at that station.

The sweeper does not know how to process multiple separate profiles that exist at a single location along the sweep - you need to configure your curves using Join so that you have one single curve at each station.

Either that, or if you want to have multiple curves at a single station then you would need to do Sweep multiple times, each time using just one of the curves at that station.

- Michael

Hi Anthony,

> I should mention all points are in meters.

With small values like this, you'll generally be better off setting units to centimeters or millimeters, rather than meters...

If you have features that approach the fitting tolerance (which in MoI is 0.001 units), it can make for some badly formed results - you want to avoid having entire features of your model that approach that size.

So instead of modeling with units = meters and using numbers like radius = 0.005 and stuff like that, I'd strongly recommend using units = mm and numbers like 5 - that will generally help to keep you away from the problem of having entire features of your model at the fitting tolerance level.

In the future I want to switch some more things to work with an "adaptive" tolerance where it uses a fraction of the size of the smallest feature that is currently being used. Some operations work like that already, but not everything. In the future when the adaptive tolerance system is more fully in place it will probably work better to use things with smaller numeric values but until that time you should avoid it.

- Michael

Hi Anthony, that little tiny arc at the top of your shape up here:



That arc has a radius of 0.00006 units - that's not just approaching the fitting tolerance, it's quite a ways under it. That will cause all kinds of various problems - basically that arc will tend to melt away since various mechanisms will think that a line is within tolerance of it as a proper shape to represent it.

Even when you work at a better scale, that little arc at the top is still really out of proportion with the rest of the shape and you'll probably run into a lot of problems with something that has such a teeny tiny rounded piece in it that is so much smaller than the rest of the surrounding model.

Really you should generally avoid creating models that have that type of way out of proportion feature in it - it throws a monkey wrench into a lot of different kinds of things.

- Michael

Thanks for all of the help everyone. I'll look into this some more. I don't know why my circle would not trim. Both points should have been on the circle. But since you got it to work, that gives me hope. As far as the trailing edge radius being very small unfortunately I can't do anything about that. That is a NACA 65A009 airfoil. I don't specify what it should look like. That is just what it looks like. I can do it in Rhino. I have to change all the default settings to do it though. I wan't to try the suggestion earlier of scaling the model way up then way down and see how that goes. I joined the entire airfoil surface together but it still made the model with errors. I hadn't mentioned it previously. But given everything everyone has said it has to be a tolerancing and dimensioning thing. I'll play around more since now I know at least in theory what I'm trying to do can work. I have the Rhino model as a part of the download to my program on my SourceForge site. I tried to upload the Rhino model here but it was too big.

The rhino model is in the examples/rhino folder within the prop_design.zip file. The reason I was interested in trying MoI is that you do have to fiddle with the Rhino settings a lot and I don't think everyone would know to do that. Also rhino can't do the fillets when you try to blend the blade to a hub. So you'd have to take it into solidworks or something. I don't think everyone has the money to do that. So ideally a low cost, easy to use, program that you can get a detailed fixed pitch aircraft propeller is what I'm looking for. Not for me, but for people that would use my program. I can't whole heartedly recommend rhino as I know it has limitations. Moreover, you can only get so far with it before you have to abandon it for an even more expensive program.

Edit; I have since closed my SourceForge project. You can find my MoI example here; http://propdesign.weebly.com/.

Hi BurrMan,

Thanks, I looked at your model and you nailed it. I will try to do it myself for learning purposes.

Anthony

spoke too soon. when i tried the method burrman used i found the surfaces at the tip near the trailing edge were malformed. so i looked at burrmans model in that region and he had the same issue. i did notice i had to redo the trailing edge arc as one arc rather than two. i had it that way because of how rhino needed it. no big deal to change though.

Changing the mesh angle option to .1 seems to fix this issue. This is a dumb one, I read the help before asking. But I'm on a laptop, I have no middle mouse button. So in the 3d view I can't pan. Is there a way to pan in 3d in this situation?

So I made it as far in MoI as I did in Rhino. Next is trying to blend the blade into the hub, which I can't do in Rhino. I tried a few things in MoI but not having luck. In rhino I can trim the blade with the hub and then do a polar array to at least get a basic idea of what the propeller looks like, but I can't have the hub blend in with the blades. In MoI when I try to trim the blade with the hub it doesn't trim it at the surface of the hub. Not sure how to blend it. Especially if there is no true intersection at the hub.

Right now the MoI model is scaled up 1,000,000 times its true size with default settings, so that the trailing edge radius comes out right. A polar array about zero using 8 blades would give the finally propeller.

hi there

..regarding PAN in 3D-window:



best wishes,
marco

Hi Marco,

I press that button but still can't pan. I only have a right and left trackpad button since its a laptop. Holding down the right button rotates even if I press the pan button.

Oh duh I just figured it out. If you hold them down and move it works.

When you draw a curve at the intersection of the blade and hub the curve is not at the intersection. Same goes for the trim. The orange and green surfaces are the result of trimming the blade with hub (shown in blue). Before trimming I also tried having a curve at the intersection of the blade and hub and it comes out exactly where the trim does. My main goal is to smoothly blend the blade into the hub. The exact geometry and dimensions aren't too important so long as it looks right. Then trim the remaining blade and do a polar array of that geometry to make the whole propeller.

Edit; Attachments have been removed to save server space.

Hi prop_design - re: Pan - yeah clicking and then dragging on that button at the bottom of the viewport will do it, or another alternative is to hold down Shift while you drag the right mouse button inside the viewport, that will also do the same as the middle button pan.

- Michael

Hi Anthony,

> I don't know why my circle would not trim. Both points should
> have been on the circle.

Can you please post the 3DM file with the circle in it so I can take a look at it?

That's the only way that I will be able to tell you for sure what is going on with the circle trimming for your particular case.

- Michael

Hi Anthony,

> spoke too soon. when i tried the method burrman used
> i found the surfaces at the tip near the trailing edge were
> malformed.

It looks like most likely your actual surface are ok there - that's just a display artifact.

In order to make the display MoI has to convert your surfaces into triangles, because your video card only knows how to display triangles. If some areas do not get enough triangles created for them it can end up looking like what you show there.

Really the best thing to do for a situation like that is to just ignore that display glitch.


> Changing the mesh angle option to .1 seems to fix this issue.

That's ok for just checking it out, but I strongly do not recommend leaving it set at such a tight angle, because a tight setting there will cause a display mesh of enormous density to be created in other areas of the model. That will tend to cause an excessive amount of memory consumption once you construct a model that has more pieces in it.

So once you are satisfied that it was just a display glitch, you should change the mesh angle setting back to the default 10 degrees for regular use.

- Michael

Hi Anthony,

> Right now the MoI model is scaled up 1,000,000 times its true
> size with default settings, so that the trailing edge radius comes
> out right.

This is kind of a sign that your project is just not going to work very well - why is it that you need to create such an extremely small detail in your object?

If you are modeling the object for rendering, that is so extremely small that you won't be able to see it.

If you are modeling for manufacturing, that is so extremely small that there isn't any manufacturing process that will be able to generate such a high level of detail.

Why is it that you're trying to create a model that has such an incredibly tiny little arc as part of it?

You would really be much better off not having things like that extremely tiny arc in there in the first place - just have the sides come to a sharp point instead of putting such a tiny arc at the top and your entire project will probably be much easier to handle.

If you do need to create a model with such a huge variation in scale within the same profile, you probably will need to use some kind of specialized CAD program to do it and not MoI - from MoI's perspective that is a kind of strange and not well suited shape to try and build in MoI.

- Michael

Hi Anthony, I loaded your MOI_EXAMPLE.zip attachment from above.

So now you have the opposite problem - your model is now too extremely huge numerically and that will cause other kinds of problems.

When I select your objects, I can see that the bounding box around them is of this size:



A bounding box width of 4330590 units is extremely huge - that's so far out of scale from the regular fitting tolerance of 0.001 units that it will cause an enormous amount of detail in various calculations, like very heavy control point counts and probably hitting recursion limits in various algorithms.

You should generally target an object size of something like less than 5000 units in overall size with individual features no smaller than 0.05 units, or something similar to that.

If you have features of an extremely small micro size approaching or lower than 0.001 units that will cause a variety of problems, and if you have objects of an exceedingly gargantuan numerical size like more than 6 orders of magnitude above 0.001 units, that will also cause a variety of other kinds of problems.

If you have a project that is unable to work within those limits, you will probably need to use some kind of specialized CAD software to do it and not MoI.

If you want to complete your current project in MoI, my best recommendation is to scale your model to have an overall size of about 10 units across and eliminate the little teeny tiny micro arcs that you have in your profile shapes and instead make your profiles just come to a sharp corner instead of a micro arc curved tip. If you do that you should likely be able to finish your model without issue.

You will be spinning your wheels and running into no end of problems if you try to use either a huge scale or micro sized features as you've been doing so far.

- Michael

thanks for the help. ill have to stick with rhino.

Here's the model back with a fillet at the join, in case anyone else was following.. I scaled the model down by 100 times, twice, then put a .005 m fillet on the join, then scaled it back up to ludicrous! I also put a blend at the tip for grins. (Though I ended up with that last little non-continuous patch, for obvious reasons).

The OP should really just use a point as the edge up there. If a machine cut the edge sharp, I could probably apply the radius by licking it!!

Here it is with a 5 meter G1 fillet on it: