NHI Kolibri N-3, scale 1:48, scratch build

[RDN1954]

Merry Christmas and a Happy New year!

 

[Daniel20]

Merry Christmas Johan!

 

[dockattner]

To you as well, Johan! Be blessed this season and into the new year!

 

[Dean62]

Merry Christmas!

 

[shota70]

Merry Christmas my dear friend Johan, and best wishes for 2026.

 

[GrantTyler]

View attachment 565434
Merry Christmas and a Happy New year!

Have a good one Johan. Merry Christmas. Cheers Grant

 

[RDN1954]

I sort of hijacked @Loracs blog on this topic, but I'll contaminate my own blog from now on to elaborate a bit on 3D-modeling and the free 3D-modelling software FreeCad, which I'm currently using to generate a 3D-model of the Kolibrie.
For various reasons I want to have a 3D-model of the Kolibrie, the main one being that I consider building a second model of this helicopter. Another reason is that I'm trying to create a more consistent dataset, instead of having snippets of vaguely related data, like the scans I made back in 2024.
This is where it gets really tricky; the information I'm using to build mtpy 3D-model is exactly the same data I'm now using to build a model from and which I find lacking in usability. Creating a more reliable model from unreliable data, yeah right, you gotta be kidding me.
The main data/dimensions I'm using haven't changed, so there's an awful lot of guestimating going on, at best it will be my interpretation of this helicopter (or at it's worst a nice caricature of said helicopter...).
A very nasty surprise were the scans of the partly dimensioned main geometry drawings of the helicopter. The ratios of the scans in X- and Y-directions are not the same, so I had to check known dimensions in both directions to ensure correct scaling.
I'm not sure where this error originates from, is it in the scanned originals, or is it the scan software of the cellphone, or is there another phenomenon I'm not aware of?
Another issue I encountered was that the models I generated in 2024 in full scale, which I thought usable for any scale, were essentially not scalable. I'm still trying to figure out why not and I have some ideas, but I need a little more experimentation before drawing conclusions.
Also, the FreeCad software frequently crashes and, while there is a restore function, more often than not I end up with corrupt models, taking endless hours to repair.
The best way to split up the helicopter provides me with a nice challenge; I have to consider preliminary build sequence and manufacturing techniques, which dictate the way I have to set up the structure of the models. Since I'm building a scale model, it does not necessarily has the same structure as for its full size namesake.
Below a picture of where I'm currently at, it's slightly more advanced than what I showed in Loracs' blog, lack of progress is mainly due to lack of progress, caused by system crashes and frequent backtracking to correct modelling errors.
It's easy to establish that the free version of Autodesk with its limitation of ten models is not working for me. By the time I'll have the 3D-definition completed, I expect to have 50+ models and that's a conservative figure.

 

[RDN1954]

I sort of hijacked @Loracs blog on this topic, but I'll contaminate my own blog from now on to elaborate a bit on 3D-modeling and the free 3D-modelling software FreeCad, which I'm currently using to generate a 3D-model of the Kolibrie.
For various reasons I want to have a 3D-model of the Kolibrie, the main one being that I consider building a second model of this helicopter. Another reason is that I'm trying to create a more consistent dataset, instead of having snippets of vaguely related data, like the scans I made back in 2024.
This is where it gets really tricky; the information I'm using to build mtpy 3D-model is exactly the same data I'm now using to build a model from and which I find lacking in usability. Creating a more reliable model from unreliable data, yeah right, you gotta be kidding me.
The main data/dimensions I'm using haven't changed, so there's an awful lot of guestimating going on, at best it will be my interpretation of this helicopter (or at it's worst a nice caricature of said helicopter...).
A very nasty surprise were the scans of the partly dimensioned main geometry drawings of the helicopter. The ratios of the scans in X- and Y-directions are not the same, so I had to check known dimensions in both directions to ensure correct scaling.
I'm not sure where this error originates from, is it in the scanned originals, or is it the scan software of the cellphone, or is there another phenomenon I'm not aware of?
Another issue I encountered was that the models I generated in 2024 in full scale, which I thought usable for any scale, were essentially not scalable. I'm still trying to figure out why not and I have some ideas, but I need a little more experimentation before drawing conclusions.
Also, the FreeCad software frequently crashes and, while there is a restore function, more often than not I end up with corrupt models, taking endless hours to repair.
The best way to split up the helicopter provides me with a nice challenge; I have to consider preliminary build sequence and manufacturing techniques, which dictate the way I have to set up the structure of the models. Since I'm building a scale model, it does not necessarily has the same structure as for its full size namesake.
Below a picture of where I'm currently at, it's slightly more advanced than what I showed in Loracs' blog, lack of progress is mainly due to lack of progress, caused by system crashes and frequent backtracking to correct modelling errors.
It's easy to establish that the free version of Autodesk with its limitation of ten models is not working for me. By the time I'll have the 3D-definition completed, I expect to have 50+ models and that's a conservative figure.

View attachment 568914

Note to anyone using 3D modelling software: you can't hit the save-button often enough (spoken from own experiences).

 

[RussF]

When I first got my 3d printer in 2019, I found DesignSpark from RS-Online. I used it to create quite a few models before moving to OpenSCAD, which is open source and totally free. DesignSpark has a free version with no size or quantity limitations. The page I linked shows the differences between versions. If you can live without mirroring and STL reverse engineering, it may very well suit you. I don't remember it crashing often.

Now, I love OpenSCAD, however, unless you have a computer programming background, I wouldn't recommend it for you. Rather than visually creating a model by sketching, extruding, etc., a model is defined in a programming language and appears visually in a workspace as you write the code. Yeah, I know, seems odd and totally counter-intuitive, and at first, I couldn't imagine how you could possibly do that. Once I got the hang of it, however, I fell in love. Again, unless you like writing code, try DesignSpark instead.

 

[Daniel20]

FreeCad software frequently crashes and, while there is a restore function, more often than not I end up with corrupt models, taking endless hours to repair.

Great subject Johan, I use TurboCAD software in my little mapping business and there is nothing more frustrating than battling software crashes, freeze ups, or all the sudden the program just shuts down after a 2-hour design session, and you haven't saved any of your work. I have read many other comments from users of various CAD software apps, and these types of problems seem too common across the board. I do wonder if users of really high-end CAD software have to deal with the same types of problems.

 

[RDN1954]

When I first got my 3d printer in 2019, I found DesignSpark from RS-Online. I used it to create quite a few models before moving to OpenSCAD, which is open source and totally free. DesignSpark has a free version with no size or quantity limitations. The page I linked shows the differences between versions. If you can live without mirroring and STL reverse engineering, it may very well suit you. I don't remember it crashing often.

Now, I love OpenSCAD, however, unless you have a computer programming background, I wouldn't recommend it for you. Rather than visually creating a model by sketching, extruding, etc., a model is defined in a programming language and appears visually in a workspace as you write the code. Yeah, I know, seems odd and totally counter-intuitive, and at first, I couldn't imagine how you could possibly do that. Once I got the hang of it, however, I fell in love. Again, unless you like writing code, try DesignSpark instead.

Sounds a bit like the old CADDS5 software, which also required writing code in Basic. Editing was a freaking nightmare though: one had to go back to the first line of the feature to be edited and take it from there, you couldn't go scrolling up and down the spec tree. That would definitely lead to a corrupted model.

 

[RDN1954]

Great subject Johan, I use TurboCAD software in my little mapping business and there is nothing more frustrating than battling software crashes, freeze ups, or all the sudden the program just shuts down after a 2-hour design session, and you haven't saved any of your work. I have read many other comments from users of various CAD software apps, and these types of problems seem too common across the board. I do wonder if users of really high-end CAD software have to deal with the same types of problems.

Like I mentioned before, professionally I used Dassault Systémes CATIA, lastly version 5. Even with this high-end tool we frequently had issues, although I'm not altogether sure it was always related to CATIA, or that maybe part of the blame fell to either Unix or Windows.
I can't recall whether or not we had an autosave function, native to CATIA, but regular saves are essential.

 

[Dean62]

I will chime in and say Creo (ProEngineer) 3D CAD software is very stable. A crash is a rare occurrence that usually only happens with large assemblies. And even then it’s rare. I use it daily, and I probably average 2 crashes per year. It is similar to Solidworks, which is another great software for mechanical and structural Engineering.
That said, I think your computer is very important and should have a high end graphics card with a lot of RAM. 3D is like gaming and it uses a lot of memory. And your processing speed is very important as well. So bottom line, doing 3D CAD on a low end computer would be a recipe for frustration. I am not implying you are, but you want to make sure your computer can handle the minimum requirements for the software you are using. Hope I’m not being rhetorical, as I know you have CAD experience professionally.

 

[RDN1954]

I will chime in and say Creo (ProEngineer) 3D CAD software is very stable. A crash is a rare occurrence that usually only happens with large assemblies. And even then it’s rare. I use it daily, and I probably average 2 crashes per year. It is similar to Solidworks, which is another great software for mechanical and structural Engineering.
That said, I think your computer is very important and should have a high end graphics card with a lot of RAM. 3D is like gaming and it uses a lot of memory. And your processing speed is very important as well. So bottom line, doing 3D CAD on a low end computer would be a recipe for frustration. I am not implying you are, but you want to make sure your computer can handle the minimum requirements for the software you are using. Hope I’m not being rhetorical, as I know you have CAD experience professionally.

True statements you made about software stability and hardware. I'm using Apple hardware, meaning I am limited in my options for 3D-modeling software. Professional software is beyond reach as well, mainly because of their price tags. Autodesk for instance has a monthly US$85 subscription fee for their Fusion software. For my purposes FreeCad offers the best program, meaning I'll have to live with less stability, but hopefully I'll find one day the reason for the issues I'm experiencing.