pironman
Guest
Good morning, everyone.
in recent months I have tested the various open-source alternatives regarding modeling 3d, both for having a cad usable at home for personal projects, and to check if there are open-source alternatives to be used in the professional field.
I have decided to share here what I think is to give "talk" to those who want to use one of these software, both (and above all) to have opinions from those who used them and maybe have arguments different from mine.
summarizing what I need:
- geometric verification of the component in terms of masses and inertias;
- verification of any interference between the components;
- display of the component before producing it, to have a clearer idea of how it will be and to find solutions to any problems;
- creation of 2d technical drawings to be able to realize the designed component;
- fem verification
I make a brief review of the pros/against myself found:
brl-cad
pro:
- accurate and fast raytracing;
- realistic materials and geometries, therefore accurate interference, masses and inertias;
- extreme speed both to load and to handle large assemblies;
- import - export with various standard formats (in particular iges and dxf);
- accepts scripting languages, which means that any part or together, if written in the form of scripting, can be parameterized very efficiently;
against:
- the command-line interface requires a certain training period;
- display only wireframe, need raytracing whenever you want to have a realistic view;
- lack of fitting-muses function, must be created with boolean commands;
openscad
pro:
- scripting language easy to learn and very powerful;
against:
- save the file only with "triangular" formats, no nurbs or csg: heavy and little accurate;
solvency
pro:
- extremely intuitive interface;
- excellent management of couplings in assemblies;
- possibility of exporting views and sections 2d (useful for table setting);
against:
- really poor performance with assemblies not even too big;
- lack of pattern for repeated insertion of components (for example screws);
freecad
pro:
- aims to be the most similar to commercial proposals;
- easy and intuitive to use;
against:
- still alpha phase;
- lack of a valid solution for assemblies (but I did not try the last update);
- in my opinion, too much dispersion (I extend better later);
made this brief summary, within a little more in detail:
brl-cad: If geometry is created using its modeling commands, each component must be placed by imposing coordinates. therefore does not allow the creation of constraints between the components themselves, and in case I go to change one, for example the length of a biella, then I will have to manually move all the connected components. But this is also true for software I use at work (creo direct), with the difference that on brl-cad I don't necessarily have to create geometry using its editor, I can create a scripting file! at that point, I can do anything I want, even assign the size of some components according to others, assign any kind of bond, group the variables... once learned to use it, the possibilities are endless.
against this requires more initial work, both to learn scripting language, and to apply it every time. Moreover, if I already have a file created using mged (the geometric brl-cad editor), I cannot automatically transform it into scripting: any modification on that file should be done manually.
I must say, however, that once I learned to use it, the creation of a geometry is much faster than I expected at first. Moreover, I find really intuitive the csg engine (with Boolean logic), because to model the component I have to reason in the same way that if I had the full block and I had to work it on the machine tool.
openscad: just installed, I liked it immediately. it is to say that to me programming likes, in free time programmo with python, and this cad basically is a programming language! same advantages of creating the script file with brl-cad, but simpler implementation.
Conversely, I soon discovered that the possibilities are quite limited, besides the file formats with which it saves have no practical engineering application: you can only print in 3d, but for me the utility is almost nothing, since all the drawing components are obtained by machine tool or carpentry. for this reason I soon stopped using it.
solvespace: the interface is the most intuitive thing I've ever seen, you don't need to read any manual but you can start modeling immediately. each component is saved on separate files, which can be recharged to go to form the axieme. to change one of the components, the axieme automatically updates: works the same way as professional parametric cads. Moreover it has the very useful possibility to save 2d views in dxf format, which can thus be loaded into any 2d cad to realize its constructive designs.
putroppo the initial enthusiasm turned out early when I realized that it is really little enough to slow it down to levels such as to make it almost unusable: has no multithreading, and a simple 400 kb set on a asus rog is enough to block it, sin! Moreover, seeing his github page, he gave me the impression that the developer is very little active on this program, which does not make me think about his future.
freecad: just open, the impression is to use a professional cad system. the beginning is really promising, there are practically all the functions present in any professional cad and also ease of use is exemplary (not at the level of solvespace, which I think the best from this point of view, but it goes very close). putroppo after a little comes bitterness. First of all, only single parts can be created, and not axiemi: even a bic pen is a set composed of 6 parts! a cad that does not allow to create assemblies is, in my view, useless. I know that there is an axiom module still developing, I saw videos and it seems to work well, but the fact that it is not released officially makes me think it has problems. Anyway, I didn't prove it so I can't evaluate it.
I happened several times that the interface changed as I used it, to the point of making it unusable, since they disappeared essential windows for the creation of sketches or components. nothing unresolvable, restoring the configuration file came back as before, but still annoying.
It's the only one of the comparisons to have a component board module, although I found it really limiting, it's far light years from being a professionally usable module.
In fact it is a promising cad, but still very little mature, because of the instability and functions that I consider essential not yet implemented.
I also find it absurd that, after 19 years of development, it is still in phase alpha: Maybe it would have been better to fully develop a module before moving on to the next, instead we find ourselves with a software that does so many things, but it all hurts (or in a non-optimal way).
I deliberately left out of this comparative salome. It is in fact a software cae, more than cad, thought for fem analysis and I must say that it does egregiamente his work. I also use it professionally in the company, where it replaced 100% the simulated creo I used previously. in fact to be powerful is the code_aster solutor, but the salome interface behaves egregiamente.
If someone searches for reliable fem software, then, I can only recommend this. requires a little more time to be learned, compared to calculix (implemented as a freecad module), but it is absolutely worth considering the enormous possibilities it offers.
As for the table, there is no 3d cad with integrated a decent 2d module, librecad use. very similar to autocads, there are some functions present in the latter, but what is more than enough and works nicely. This cad has given me some stability problem in the past, but the last revision is really "upper".
In my opinion, the cad can mature, stable and usable is brl-cad. probably because it had 20 years of development within US Army laboratories, before being released as open source software (to date, it counts 36 years of active development!). all others, even if more intuitive or easy to use, for one reason or another I found them limiting.
My dream, perhaps unrealizable, is this: brl-cad, as it is today, but with addition the possibility to use the solvespace interface, and the implementation of the 2d librecad with direct connection to the 3d.
in recent months I have tested the various open-source alternatives regarding modeling 3d, both for having a cad usable at home for personal projects, and to check if there are open-source alternatives to be used in the professional field.
I have decided to share here what I think is to give "talk" to those who want to use one of these software, both (and above all) to have opinions from those who used them and maybe have arguments different from mine.
summarizing what I need:
- geometric verification of the component in terms of masses and inertias;
- verification of any interference between the components;
- display of the component before producing it, to have a clearer idea of how it will be and to find solutions to any problems;
- creation of 2d technical drawings to be able to realize the designed component;
- fem verification
I make a brief review of the pros/against myself found:
brl-cad
pro:
- accurate and fast raytracing;
- realistic materials and geometries, therefore accurate interference, masses and inertias;
- extreme speed both to load and to handle large assemblies;
- import - export with various standard formats (in particular iges and dxf);
- accepts scripting languages, which means that any part or together, if written in the form of scripting, can be parameterized very efficiently;
against:
- the command-line interface requires a certain training period;
- display only wireframe, need raytracing whenever you want to have a realistic view;
- lack of fitting-muses function, must be created with boolean commands;
openscad
pro:
- scripting language easy to learn and very powerful;
against:
- save the file only with "triangular" formats, no nurbs or csg: heavy and little accurate;
solvency
pro:
- extremely intuitive interface;
- excellent management of couplings in assemblies;
- possibility of exporting views and sections 2d (useful for table setting);
against:
- really poor performance with assemblies not even too big;
- lack of pattern for repeated insertion of components (for example screws);
freecad
pro:
- aims to be the most similar to commercial proposals;
- easy and intuitive to use;
against:
- still alpha phase;
- lack of a valid solution for assemblies (but I did not try the last update);
- in my opinion, too much dispersion (I extend better later);
made this brief summary, within a little more in detail:
brl-cad: If geometry is created using its modeling commands, each component must be placed by imposing coordinates. therefore does not allow the creation of constraints between the components themselves, and in case I go to change one, for example the length of a biella, then I will have to manually move all the connected components. But this is also true for software I use at work (creo direct), with the difference that on brl-cad I don't necessarily have to create geometry using its editor, I can create a scripting file! at that point, I can do anything I want, even assign the size of some components according to others, assign any kind of bond, group the variables... once learned to use it, the possibilities are endless.
against this requires more initial work, both to learn scripting language, and to apply it every time. Moreover, if I already have a file created using mged (the geometric brl-cad editor), I cannot automatically transform it into scripting: any modification on that file should be done manually.
I must say, however, that once I learned to use it, the creation of a geometry is much faster than I expected at first. Moreover, I find really intuitive the csg engine (with Boolean logic), because to model the component I have to reason in the same way that if I had the full block and I had to work it on the machine tool.
openscad: just installed, I liked it immediately. it is to say that to me programming likes, in free time programmo with python, and this cad basically is a programming language! same advantages of creating the script file with brl-cad, but simpler implementation.
Conversely, I soon discovered that the possibilities are quite limited, besides the file formats with which it saves have no practical engineering application: you can only print in 3d, but for me the utility is almost nothing, since all the drawing components are obtained by machine tool or carpentry. for this reason I soon stopped using it.
solvespace: the interface is the most intuitive thing I've ever seen, you don't need to read any manual but you can start modeling immediately. each component is saved on separate files, which can be recharged to go to form the axieme. to change one of the components, the axieme automatically updates: works the same way as professional parametric cads. Moreover it has the very useful possibility to save 2d views in dxf format, which can thus be loaded into any 2d cad to realize its constructive designs.
putroppo the initial enthusiasm turned out early when I realized that it is really little enough to slow it down to levels such as to make it almost unusable: has no multithreading, and a simple 400 kb set on a asus rog is enough to block it, sin! Moreover, seeing his github page, he gave me the impression that the developer is very little active on this program, which does not make me think about his future.
freecad: just open, the impression is to use a professional cad system. the beginning is really promising, there are practically all the functions present in any professional cad and also ease of use is exemplary (not at the level of solvespace, which I think the best from this point of view, but it goes very close). putroppo after a little comes bitterness. First of all, only single parts can be created, and not axiemi: even a bic pen is a set composed of 6 parts! a cad that does not allow to create assemblies is, in my view, useless. I know that there is an axiom module still developing, I saw videos and it seems to work well, but the fact that it is not released officially makes me think it has problems. Anyway, I didn't prove it so I can't evaluate it.
I happened several times that the interface changed as I used it, to the point of making it unusable, since they disappeared essential windows for the creation of sketches or components. nothing unresolvable, restoring the configuration file came back as before, but still annoying.
It's the only one of the comparisons to have a component board module, although I found it really limiting, it's far light years from being a professionally usable module.
In fact it is a promising cad, but still very little mature, because of the instability and functions that I consider essential not yet implemented.
I also find it absurd that, after 19 years of development, it is still in phase alpha: Maybe it would have been better to fully develop a module before moving on to the next, instead we find ourselves with a software that does so many things, but it all hurts (or in a non-optimal way).
I deliberately left out of this comparative salome. It is in fact a software cae, more than cad, thought for fem analysis and I must say that it does egregiamente his work. I also use it professionally in the company, where it replaced 100% the simulated creo I used previously. in fact to be powerful is the code_aster solutor, but the salome interface behaves egregiamente.
If someone searches for reliable fem software, then, I can only recommend this. requires a little more time to be learned, compared to calculix (implemented as a freecad module), but it is absolutely worth considering the enormous possibilities it offers.
As for the table, there is no 3d cad with integrated a decent 2d module, librecad use. very similar to autocads, there are some functions present in the latter, but what is more than enough and works nicely. This cad has given me some stability problem in the past, but the last revision is really "upper".
In my opinion, the cad can mature, stable and usable is brl-cad. probably because it had 20 years of development within US Army laboratories, before being released as open source software (to date, it counts 36 years of active development!). all others, even if more intuitive or easy to use, for one reason or another I found them limiting.
My dream, perhaps unrealizable, is this: brl-cad, as it is today, but with addition the possibility to use the solvespace interface, and the implementation of the 2d librecad with direct connection to the 3d.
