AoI Alternatives

Hi!

(Preface - I have never used AoI)

NopHead's recent post [hydraraptor.blogspot.com] regarding the AoI mesh problem revealed something as problematic as the mesh problem itself:

AoI is not an engineer-oriented software.

I am learning solidworks at the moment and I can see how using an engineering oriented tool would be neccesary in order to produce results with the reprap on a large scale.

Imagine you need to create a disk with 7 evenly spaced holes on it.
with solidworks you create one hole, position it, select "circular patter", select 7 and hit ok.

If I understand correctly, in AoI you would have to calculate (!!!) the position of each hole.

Instead of being smart about and "knowing" that the holes are in a circular pattern AoI only knows you placed the holes in some arbitary position.

Imagine realising you need to move all the holes 5 mm outwards (radially).... what a headache!

Here is a wikipedia list of free CAD software.
[en.wikipedia.org]

some of them are more oriented to graphics... some for for architecture... but at least one looks pretty cool:
Open CASCADE
[en.wikipedia.org]

I'm downloading it now but the screenshots look very promising.

here is the website:
[www.opencascade.org]

if you want to download it to try you can use my login:
user: dudedude
password: dude1234

What do you think?
-Leav

p.s.
PLEASE someone reply to this post! all my other threads are like ghost towns!

Hi Leav,

... thanks for the hints!

I'm downloading open Cascade too and will try the next days and give then some feedback eventually ...

Viktor

Leav Wrote:
-------------------------------------------------------
> AoI is not an engineer-oriented software.

And your point is?

I downloaded Open CASCADE. It had some strange java 'installer' that extracted lots of files. I sifted though the files it extracted and found a folder called 'ros' that had a makefile in, so i tried to build it but I got an error:
../../../drv/Dico/Dico_DictionaryOfInteger_0.cxx:51: error: 'Standard_Transient_Type_' was not declared in this scope
make[3]: *** [Dico_DictionaryOfInteger_0.lo] Error 1
make[3]: Leaving directory `/home/stef/xsb08/downloads/OpenCASCADE_Linux/Linux/ros/adm/make/TKernel'

... i've got OpenCascade working on WinXP.

Most of the executables are samples and demonstration of the capabilities, drawing/constructing seems to be a shell-based task with a lot of keywords to learn

So next is to find some CAD-GUI's - the comercials are in the range of 500 Euros upward ...

Viktor

Forrest Higgs Wrote:
-------------------------------------------------------
> Leav Wrote:
> --------------------------------------------------
> -----
> > AoI is not an engineer-oriented software.
>
> And your point is?

That it seems to be lacking in features that have become a standard in engineering oriented software.
and this (lacking of features) is/will be a limiting factor in designing parts.

if there are better options than AoI for the designing of parts for RepRap printing, shouldn't we at least give them a look?

I'll take a thourugh look at openCASCADE later today and post my impressions.

-Leav

Leav Wrote:
-------------------------------------------------------
> Forrest Higgs Wrote:
> --------------------------------------------------
> -----
> > Leav Wrote:
> >
> --------------------------------------------------
>
> > -----
> > > AoI is not an engineer-oriented software.
> >
> > And your point is?
>
> That it seems to be lacking in features that have
> become a standard in engineering oriented
> software.
> and this (lacking of features) is/will be a
> limiting factor in designing parts.

Do you really imagine that the Reprap project is aimed at making 3D printing convenient for engineers or that we expect our audience to change their behaviour to become more LIKE engineers?

AoI was chosen largely because it has a very short learning curve for newbies. The STL file format was chosen because it meant that you could use whatever 3D modeling software you were comfortable with by virtue of your previous experience in Darwin.

I think that we may well want to add features to AoI to make doing otherwise difficult things with it easier. I can't see Reprap, however, shifting over to some quasi-commercial 3D package on the basis that it is an engineering professional-friendly tool. m

I've been using art of illusion for a while, and I like it. But what annoys me about it is that all operations are absolute once done. i.e. when you join two cubes you get a brand new object, and changing the original cubes has no impact on your new object.

Do CAD orientated programs do this? It's really annoying when you want to make a slight change to an object and have to re-do loads of unrelated stuff in order to do it.

The best solution I've found is to use null objects to group entities together and then construct your object out of several key blocks. Then when you change something you just need to do a few boolean operations between the key blocks. It would still be nicer to have the computer do this though.

Yup, once you do a boolean op in AoI you are stuck with the result. What I do is keep the primitives that I used to generate the boolean and then if I need to redo the boolean I work with those.

I believe AOI was chosen over Blender because its user interface is more intuitive:
[blog.reprap.org]

While the interface might be intuitive, how to use it to model certainly is not in my opinion. You have to read Adrian's hints [reprap.org] first and his deep discussion about open and closed sets [reprap.org] otherwise you would never be able to make anything Euler manifold. I showed this stuff to the mechanical engineers at work who have lots of experience modeling complicated things and they had never heard of this problem. The open and closed set maths was way over their heads (their assessment not mine). Do we really expect non technical designers to be able to use this? I couldn't even manage to model a dimensionally correct nut cavity without asking for help.

The only way you get anything done is by using tricks like setting the grid to half the size of your object or making things children of a dummy cylinder to be able to space them round an axis.

It would be nice to be able to use AOI and extend it but it would take an enormous effort to get it anywhere near user friendly for modeling real objects.

[www.hydraraptor.blogspot.com]

One thing that I did find with the AoI boolean ops is that while the union function is unreliable at times the intersection functions seem to almost always work.

@Forest, If you are building a reprap and doing all the circuitry and parts and designing new things (else why have it) then you pretty much are an engineer, or at least a hobby one.

IMO the aoi interface and 3d perspective tools are horrible, at least from what little I tried it. I'm probably not going to bother installing it again because I already have a program I like that works for me. Milkshape 3D (cheap and has stl output w/ plugin).

Maya has some great tools for getting all the angles right and repetitive tasks and getting everything square. There is also a free PLE license.

~Dylan

Dylan Wrote:
-------------------------------------------------------
> @Forest, If you are building a reprap and doing
> all the circuitry and parts and designing new
> things (else why have it) then you pretty much are
> an engineer, or at least a hobby one.

I'm afraid that what what I am doesn't fit into neat categories. For 3D modeling I use AoI. I could as easily use a multitude of other apps or even 3D modeling code that I wrote myself as part of my Doctoral dissertation work.

The point of Reprap is to make the technology accessible to as many people as possible. The STL transfer file format does that.

I agree in keeping the STL. It's a good file size and a standard for 3D RP.

I also think that everyone will use whatever software they want.

What I meant by "you" is the generic hypothetical person doing the project, not you specifically. Even if [an individual] is going to use Darwin to print odd 3D sculptures as an artist, by my defenition, engineer would be partial title of skillset. It doesn't mean (to me) that that's what you are or all you are.
Until we get an all in one kit that people can buy and just put together like ikea furniture or legos, there will be some level of mechanical and electrical skill needed to build the unit.

I couldn't agree more about making it accessible and CHEAP for everyone as well as able to make itself. If we keep making improvements that up the cost more and more it would be easier to just buy a currently available commercial machine. Which is why I like that I see a lot of suggestions for cheaper parts resources and I'm not going to tell people not to use AOI, (it is free after all) but it isn't necceserily for everyone.

@Forrest: integrating Quasi-Commercial software into RepRap makes me feel sick too...

STL is a great format in my opinion. I don't know where it was implied in this post that it should be abandoned... :/

All I am really asking is this: if AOI is not ready to be used as a design program for RapRap parts (second hand information here), what is the project's course of action?

a)Ignore the problem and wait for the problem to solve itself (if at all)?
b)try and solve the problems with the software? (e.g. by encouraging RepRap volunteers to work with the AoI volunteers)
c)move to a different program?

-Leav

Perhaps I just haven't been as ambitious or as exacting as many here in my demands of the AoI 3D modeling software. Frankly though, I've always been able to make it do what I wanted. I've written a few scripts to get me past the difficult bits, but damned few all told. Years ago, I wanted to be able to develop involute profile gears and managed that quite nicely. I also wanted to be able to make polygonal rods and holes and wrote one to do that as well. Looking around in my folders, however, I seem to have mislaid that one and will probably have to write it again.

When I hear Nophead and Enrique talk about AoI falling short I look at what they are actually designing and can see why they feel that way. However, when I see Nop throw up a drawing like this...

[bp2.blogger.com]

...and say that he couldn't see how he could do that in AoI I find myself scratching my head and wondering

a) why he thought he couldn't

and more importantly...

b) why he wanted the mounting to look exactly like that.

Keep in mind that I am not criticising Nophead. I have a profound respect for his work.

Similarly, when I find Enrique, whose work I also deeply respect, writing scripts to make fillets I arrive at the conclusion that their demands for a 3D CAD package is sort of self-referential. They seem to want what they design to look like it came out of an engineering design office. It would follow that to get that "look and feel" you are best served by an engineer-oriented 3D CAD package rather than something like AoI.

I started out at university in aerospace engineering and then drifted into architecture forty-odd years ago. I know what the engineering look-and-feel is like and it really hasn't changed much in all those decades. I guess that I'm not overly impressed with it because I know that the look-and-feel is just that and has, more often than not, very little to do with functionality. You can see that in my work. My stuff, often as not, looks nasty and very much a function of both the materials I used and the tools I used to shape it. It works though.

I feel the same way about AoI. I can make it do what I want, but then what I want is functionality, not a "look". Mind, I doubt that I'll ever win a design award.

a) I don't know how to do blends in AOI, especially when they are in two directions. On an external corner the best I could do would be to Boolean intersection with a cylinder, but I would have to make sure it was not exactly tangential, which of course it should be, otherwise the result would be non manifold. If you look at how many radii there are in the motor mount, each one of those would involve creating a cylinder, somehow positioning it nearly tangential, intersecting, hiding the originals, naming the new part, converting to a mesh.

Unless there is an easier way that I don't know, it is not practical to do this in AOI. How you would put a radius along anything but a straight edge I have no idea.

I suspect this is why if you look at any parts designed in AOI: Adrian's door handle, coat hook and the extruder, they are all very simple angular shapes.

b) It saves build time and material and adds strength. Corners tend to be a bit blobby so if they meet another part it is better to radius them so there is no corner.

The machine shakes the house less if when the corners are rounded!

The rounded top on the coupler is so that it can be pushed into a tube easily, and it could be a taper instead. All the other blends could be removed without affecting the functionality.

When you have a fabrication technology where part complexity comes free it would be a shame to have it limited by the design tool.

[www.hydraraptor.blogspot.com]

nophead Wrote:
-------------------------------------------------------
> I suspect this is why if you look at any parts
> designed in AOI: Adrian's door handle, coat hook
> and the extruder, they are all very simple angular
> shapes.

Yup, I often wonder what the parts set for Darwin would have looked like if Adrian had been more of a purist about using open source tools on this aspect of the design process and designed it using AoI instead of SolidEdge. I think that a very different design might well have emerged.

> The machine shakes the house less if when the
> corners are rounded!

Really?

> When you have a fabrication technology where part
> complexity comes free it would be a shame to have
> it limited by the design tool.

Any product is limited by the tools used to create it. I find, however, that the more familiar I become with a design tool the less important those limitations become.

The fillet script in the skeinforge toolchain rounds all corners slightly, so that the extruder doesn't have make sharp turns and therefore try to do infinite acceleration. Nophead's firmware does something similar and he doesn't need the fillet script, which can be turned off or avoided.

On top of the slight amount of filleting that is done on each corner, including corners of the infill, it is better still if the model has large radius fillets.

With subtractive manufacturing, sawing or cutting operations naturally produce sharp corners. Extra operations have to be done to smooth the resulting corners.

With additive manufacturing, extrusion naturally produces round corners, the rounder the corner the less acceleration there is and therefore the less noise, wear & tear.

The optimum shapes for additive manufacturing have the melted candy look, they look organic. When you see futuristic designs with parts flowing into each other, you know they must be reprapped

Enrique Wrote:
-------------------------------------------------------
> The optimum shapes for additive manufacturing have
> the melted candy look, they look organic. When
> you see futuristic designs with parts flowing into
> each other, you know they must be reprapped

Yes! Exactly!

As I see it, AOI may not be perfect, but it is the easiest 3d modelling program I have learned to use, and you can do pretty complex things with it once you learn the rather roundabout ways of doing them. We know how we would prefer it to work, but I have yet to find another program that is so easy to use and does exactly what we want.

I have experimented with rounding edges in AOI. The (no so) easiest way I found was to create a curve around the edge to be cut off:

____________
|      _____|
|    /
|   /
|  /
| |
|_|

Turn this into a mesh, extrude it, then use boolean operations to cut off the sharp edges. Far from ideal but it works.

Another useful trick I found when designing a stepper motor was to design the central cavity as an object, then just subtract this from two solid blocks to make the motor housing.

Forrest Higgs Wrote:
-------------------------------------------------------

> Do you really imagine that the Reprap project is
> aimed at making 3D printing convenient for
> engineers or that we expect our audience to change
> their behaviour to become more LIKE engineers?
>
>
> AoI was chosen largely because it has a very short
> learning curve for newbies. The STL file format
> was chosen because it meant that you could use
> whatever 3D modeling software you were comfortable
> with by virtue of your previous experience in
> Darwin.
>
> I think that we may well want to add features to
> AoI to make doing otherwise difficult things with
> it easier. I can't see Reprap, however, shifting
> over to some quasi-commercial 3D package on the
> basis that it is an engineering
> professional-friendly tool. m


No one is suggesting that anyone shifts over simply to be more "professional-friendly". However, I would like to say that I am NOT a professional, but I have had the opportunity to work with both Autodesk Inventor and Solidworks independently (outside of a professional setting, mind you), as well as Art of Illusion, through this project. While the former are both MUCH more complex tools than the latter, they also have much shorter learning curves for simple operations. After my first day with Inventor I could easily and intuitively make objects that I still have no idea how to model in AoI, if it is possible at all. Solidworks is if anything even easier. The operations are logical, well documented, parametric, and completely reversible.

I think that this is the consensus I'm seeing here - everyone, including nophead (who as I understand it is using a cheaper program and only began very recently) who has taken the initiative to try out one of these more advanced professional engineering programs is finding it easier to learn and more capable than AoI, professional engineer or no.

I understand that the ethos of the project require an open-source-only design chain. However, I think that maintaining a healthy discussion about potential alternatives to this program is important. At the very least, we must admit that it was not designed for the task we've set to it here.

Kyle Corbitt Wrote:
-------------------------------------------------------
> I understand that the ethos of the project require
> an open-source-only design chain. However, I
> think that maintaining a healthy discussion about
> potential alternatives to this program is
> important. At the very least, we must admit that
> it was not designed for the task we've set to it
> here.

I have no great problem with using non-open source software. For me it's a matter of price and proportion. We've been trying to bring the price of 3D printing down from $30,000 to less than $500 and are making good progress in that regard.

The cheapest student version of your Inventor app is $450 and the usual commercial price is $3,000 or so. I've used it in the past and it's a good app with quite a decent GUI. Mind, I don't agree with you that it is all that easy, compared to AoI, for example, to use though it is undoubtedly much more powerful.

Inventor's price is fairly typical of the really professional 3D CAD packages out there, though I've seen some fairly attractive commercial packages being marketed by startups that range in price from about $600-700.

You talk about "healthy discussion". When you are targeting, as I am, an audience of kids and non-professionals I don't there is anything healthy about promoting the use of software that is going to put much of that audience in the position of either not being able to use Reprap or heading over to Pirate Bay. That most definitely NOT a choice that I want my user demograpic to have to make.

What this "healthy discussion" is doing as a practical matter is either making John Walker, founder of AutoDesk and now resident of Neuchatel in Switzerland either richer than he already is or giving kids a hard shove towards becoming big-time software pirates. That ain't healthy, imo at least.



Edited 1 time(s). Last edit at 07/04/2008 06:23AM by Forrest Higgs.

Hi Forrest,

what's with the free MLCad ( [www.lm-software.com] ) from Michael Lachmann?

He developed it for building LEGO-models, but you can make your own parts and macros and it's exporting into other 3D-formats and the BOM too.

As Lego-models are built (brick- and) sheetwise, this could be an interesting variant for designing and fabbing ...

Viktor

VDX Wrote:
-------------------------------------------------------
> Hi Forrest,
>
> what's with the free MLCad (
> [www.lm-software.com] ) from Michael
> Lachmann?

Is there a question here?

Forrest Higgs Wrote:
-------------------------------------------------------

> You talk about "healthy discussion". When you are
> targeting, as I am, an audience of kids and
> non-professionals I don't there is anything
> healthy about promoting the use of software that
> is going to put much of that audience in the
> position of either not being able to use Reprap or
> heading over to Pirate Bay. That most definitely
> NOT a choice that I want my user demograpic to
> have to make.
>
> What this "healthy discussion" is doing as a
> practical matter is either making John Walker,
> founder of AutoDesk and now resident of Neuchatel
> in Switzerland either richer than he already is or
> giving kids a hard shove towards becoming big-time
> software pirates. That ain't healthy, imo at
> least.


Surely it is beneficial to know that there is something better out there, at least. And you don't need to break the bank to get a decent CAD package - I've heard good things about Alibre Design, which has a free version that has full functionality for creating parts and even lets you make assemblies with up to ten unique parts, enough for many small projects.

I think that even if such a discussion must be confined to open source projects, it is still worth having. There are various alternatives to AoI that haven't been explored such as Open Cascade, and having a place where people can share their experiences with these free tools is valuable.

What seems least healthy to me is treating AoI as a deity above criticism. While it is useful in some applications, it is far from ideal.

Kyle Corbitt Wrote:
-------------------------------------------------------
> Surely it is beneficial to know that there is
> something better out there, at least.

Knowing it's out there is one thing. Suggesting that it is realistic to include it as part of Reprap is quite another. Right now, you can plunk down $30,000 and get an entry level Dimension printer that will run circles around Darwin. Does that mean that promoting Dimension printers on these forums is reasonable?

> And you
> don't need to break the bank to get a decent CAD
> package - I've heard good things about Alibre
> Design, which has a free version that has full
> functionality for creating parts and even lets you
> make assemblies with up to ten unique parts,
> enough for many small projects.

Yeah, and the moment you outgrow that limitation you are either going to have to plunk down $1,000 or head over to Pirate Bay. I downloaded and worked with Alibre. It looks like a good package for all that it is apparently written in Java. The price jump from the limited use free package to the entry level commercial package is, however, a tad steep.
>
> I think that even if such a discussion must be
> confined to open source projects, it is still
> worth having. There are various alternatives to
> AoI that haven't been explored such as Open
> Cascade, and having a place where people can share
> their experiences with these free tools is
> valuable.

There are a variety of open source 3D cad packages out there. Blender is the most obvious. What we need is for people to get out there and use them and then share their experiences here so that other folks can benefit from their trail-breaking work and get up to speed with them faster.

> What seems least healthy to me is treating AoI as
> a deity above criticism.

I don't think that anybody here can be properly accused of considering AoI as being "above criticism". That's a ridiculous and utterly hyperbolic charge to make.

> While it is useful in
> some applications, it is far from ideal.

No doubt about it. I guess what annoys the hell out of me with this discussion is folks like yourself pretending that commercial software like Inventor and similar are realistic choices for a low cost, open source project like Reprap when you know very well that most of the people who are already struggling just to afford Darwin are going to have to go out and literally steal the software you are promoting to stay in the game.

Imagine how someone who came into the forums promoting Dimension printers would be received when the subtext of their promotion was that in order to play the game people were going to have to literally hijack such printers to stay in the game. IMO, promoting expensive commercial software is morally equivalent to that.



Edited 1 time(s). Last edit at 07/04/2008 07:40AM by Forrest Higgs.

Forrest Higgs Wrote:
-------------------------------------------------------

> Yeah, and the moment you outgrow that limitation [>10pc assemblies]
> you are either going to have to plunk down $1,000
> or head over to Pirate Bay. I downloaded and
> worked with Alibre. It looks like a good package
> for all that it is apparently written in Java.
> The price jump from the limited use free package
> to the entry level commercial package is, however,
> a tad steep.

With AoI, there are much greater limitations that cannot be overcome with all the money in the world. You're holding these equally priced (free) programs to wildly different standards and implying that the existence of an optional upgrade path is a detriment.

> I don't think that anybody here can be properly
> accused of considering AoI as being "above
> criticism". That's a ridiculous and utterly
> hyperbolic charge to make.

Good. Then I'm criticizing it. But we all know that a critic who doesn't offer any solutions is useless, so I and others here are also offering potential alternatives.

> > While it is useful in
> > some applications, it is far from ideal.
>
> No doubt about it. I guess what annoys the hell
> out of me with this discussion is folks like
> yourself pretending that commercial software like
> Inventor and similar are realistic choices for a
> low cost, open source project like Reprap when you
> know very well that most of the people who are
> already struggling just to afford Darwin are going
> to have to go out and literally steal the software
> you are promoting to stay in the game.

Perhaps bringing Inventor into it was a mistake. I understand that a hobbyist cannot afford software like this (for the record, I never use pirated software and received both Inventor and Solidworks as promotions through an engineering competition I'm involved with). I only used it as an example of the capabilities of this class of software because it is the program that I'm most familiar with. However, free or low-cost solutions such as Alibre Design exist in the same category.

I'm not advocating that anyone go out and buy Inventor. I'm only saying that it is well worth our time to talk about the various alternatives. I also am not advocating that anyone go out and buy a Dimension, but we've had plenty of threads on this forum discussing features that that machine has that we don't, including some that seem near impossible for this machine even in the future, such as a heated build chamber. Are those threads implying that you should go steal a Dimension? Maybe so, but I don't see it that way.

Kyle Corbitt Wrote:
-------------------------------------------------------

> I'm not advocating that anyone go out and buy
> Inventor.

You're not the only one who's done this. Ed and Adrian, after making a big to do about sticking with open source and freeware what with SDCC, Java and Linux did the parts for both ARNIE and Darwin using SolidEdge, a very good 3D CAD package that is DEFINITELY not a bargain price-wise. The parts, iirc, had to be migrated into AoI from scratch afterwards.

> I'm only saying that it is well worth
> our time to talk about the various alternatives.

Like for instance? I've heard you talk about Inventor (~$3,000) and Alibre (~$1,000-1,500) so far. Pardon me if I've got a creepy idea about what you think "alternatives" are.

Forrest Higgs Wrote:
-------------------------------------------------------
> Like for instance? I've heard you talk about
> Inventor (~$3,000) and Alibre (~$1,000-1,500) so
> far. Pardon me if I've got a creepy idea about
> what you think "alternatives" are.


Well like I've said, you've got a fair point with Inventor and I already apologized for that.

But I'm still curious what you see as the problem with Alibre Design Xpress, which is crippled compared to the full version but will design individual components just fine and outclasses AoI on nearly every front.