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

[Peter Voogt]

As mentioned before, I am spending quite a few hours in trying to get a more or less realistic1:24 3D-model.
I also hinted earlier that that's easier said than done. Now, a few weeks in the process, I can confidently say that the learning curve is steep, very steep. Moving away from the high-end, professional modeling software to the freeware modeling software remains a huge step. Actions, like projecting 3D elements to sketches, or vice versa became a tedious job. That doesn't mean the freeware is inferior or not good, it's just written differently and one simply cannot expect the same capabilities with freeware.
Still, FreeCad is a very competent piece of software.
Also my starting points are not particularly unambiguous; the few drawings I have were already scanned and sized down from A0 to A4 and scanned again, resulting in al kinds of unwanted deformations and distortions. To make matters worse, the 3 view drawings I use are from three different drawings or even a brochure, meaning the three views don't necessarily agree with one another. The side view shows the least deformation/distortion, leading to the assumption that this view is the most reliable one.
The above reinforces my view on recreating history, based on historical documents, drawings, sketches, etc, namely that what I recreate is not and cannot be historical correct, but merely an expression of my interpretation of the available data.
Still, the emerging 3D-model somewhat start to resemble the Kolibrie and more, I enjoy the results.

The first picture shows the side view, with the 3D-model in front of a scan of an original drawing. As can be seen my 3D-model closely matches the scanned and resized drawing.
View attachment 571754

The next picture shows the top view, taken from a brochure. It's rather evident that there are quite some deviations, especially the windshield frame differs significantly.
View attachment 571755

Then there is the front view. The scan is heavily distorted and even manipulations of the original didn't yield a trustworthy picture:View attachment 571756

The picture below shows the current status of the model, with the spec tree. That spec tree grows like crazy and I haven't even started modeling the complex parts. Yet.
View attachment 571759

And finally two 3D views of the helicopter model. Still a lot of details need to be added, but, once home, my main focus will be on actual building again and 3D-modeling will have to take the back burner.
View attachment 571757

View attachment 571758

That looks very impressive, Johan. The earlier copiers sometimes had significant deviations, especially towards the sides and corners. I remember when we chose a new copier for our office, we discovered that the standard reduction was 0.5%. Some deviations can be explained, but you still have to take them into account.

In appearance, you have a convincing 3D representation of your model. I would say you have a strong knowledge base with Free CAD to come as far as you have. I know with my TurboCAD, every so often I will pick some complex part off the internet such as an engine governor or U-joint PDF that has dimensions and draw it up in 3D just to keep up my drawing skills.

You made a good point, Dan; keeping your 3D-modeling routine up-to-date is essential to stay current. There are many options, hidden in the software, allowing various approaches for modeling specific details. For instance there are various methods creating a solid feature. Developing your own preferred method is quite helpful in getting up to speed.

A while ago, I received a question via PM about how I'd done certain things while developing the line drawing for the Balder. Suddenly, I had to dig really deep again to figure out how I'd done it. It's a volatile medium up there in that brain chamber...
Regards, Peter

 

[Dean62]

Obviously it’s best to have a dimensional drawing so the scale doesn’t matter. Otherwise you find yourself trying to determine a scale with a known dimension, and using the conversion factor derived from that, and measuring everything on the drawing! Not a fun way to work.
I had to do that on the light house I scratch built. I had a 2d drawing I found online and the article mentioned the overall height. That was my only known dimension. But if you have distortion on your hard copy, that’s no good.

 

[RDN1954]

That looks very impressive, Johan. The earlier copiers sometimes had significant deviations, especially towards the sides and corners. I remember when we chose a new copier for our office, we discovered that the standard reduction was 0.5%. Some deviations can be explained, but you still have to take them into account.

In copying a lot can and will go wrong. Normal copying could result in way more than 0,5% and different variations in X- and Y directions could be significantly different.
The only difference for us were the Full Scale (FS) drawings; they were supposed to be accurate.

I highly suspect some of the copies I used were not taken perpendicular to the original, that'll give a person quite the headache and a challenge.

 

[RDN1954]

Obviously it’s best to have a dimensional drawing so the scale doesn’t matter. Otherwise you find yourself trying to determine a scale with a known dimension, and using the conversion factor derived from that, and measuring everything on the drawing! Not a fun way to work.
I had to do that on the light house I scratch built. I had a 2d drawing I found online and the article mentioned the overall height. That was my only known dimension. But if you have distortion on your hard copy, that’s no good.

Ouch, having just the one dimension is more than challenging...
For the current model I made an overview of basic materials and of the dimensions I knew and took it from there.
Once the main structure was complete it became a matter of best educated guesses.

 

[RDN1954]

After returning from our little break in Spain and preparing my fourth article on this build for the NLR-periodic, I had no other choice but to focus on the windshield again. Before we left for Spain I had the basic hardware available and fit checked, so it was just a matter of making things permanent and primer and paint the resulting assembly.
For the top center piece I was able to use thin copper sheet, made available by @Peter Voogt. It was the perfect application for this sheet; easy to cut and easy to apply.
Needless to say I'm rather please with the results sofar:



However, here is the next challenge: installing the transparancies.
Making the templates for the developed acryl and using those to cut the acryl was a breeze. The challenge is in finding the most effective way to bond the windshields to the frame.
A year and a half ago I tried using CA adhesive for this purpose on my 1:48 first try at this model. Needless to say that that didn't end well. Positive control over the addition and flow is very poor with CA and any spill gives a distortion of the transparancy.
I did some research on the all-knowing internet, where I found a tip to use wood glue, since that cures to an almost invisible layer. Tried that and disregarded it almost straight away, the wood glue doesn't provide any adhesion whatsoever, nor to the brass frame, nor to acryl.

The latest tests involve a Revell plastic adhesive with a very low viscosity, almost waterlike.
The initial result is promising, but need more testing before I'm convinced, plus I might be tempted to try other adhesives as well.
The picture below shows the Revell adhesive used and the very simple test specimen:

 

[Dean62]

After returning from our little break in Spain and preparing my fourth article on this build for the NLR-periodic, I had no other choice but to focus on the windshield again. Before we left for Spain I had the basic hardware available and fit checked, so it was just a matter of making things permanent and primer and paint the resulting assembly.
For the top center piece I was able to use thin copper sheet, made available by @Peter Voogt. It was the perfect application for this sheet; easy to cut and easy to apply.
Needless to say I'm rather please with the results sofar:

View attachment 575619

However, here is the next challenge: installing the transparancies.
Making the templates for the developed acryl and using those to cut the acryl was a breeze. The challenge is in finding the most effective way to bond the windshields to the frame.
A year and a half ago I tried using CA adhesive for this purpose on my 1:48 first try at this model. Needless to say that that didn't end well. Positive control over the addition and flow is very poor with CA and any spill gives a distortion of the transparancy.
I did some research on the all-knowing internet, where I found a tip to use wood glue, since that cures to an almost invisible layer. Tried that and disregarded it almost straight away, the wood glue doesn't provide any adhesion whatsoever, nor to the brass frame, nor to acryl.

The latest tests involve a Revell plastic adhesive with a very low viscosity, almost waterlike.
The initial result is promising, but need more testing before I'm convinced, plus I might be tempted to try other adhesives as well.
The picture below shows the Revell adhesive used and the very simple test specimen:

View attachment 575622

View attachment 575620

View attachment 575621

Looking great Johan.
Finding the right glue for a job can be frustrating. I know when building plastic models, the use of PE parts and clear plastic parts can call for different applications.
I know when bonding PE or metal on plastic, I can use CA or wood glue. When bonding plastic to plastic I use the Revell liquid glue for tiny parts and the tube glue for large areas that need a strong bond. On clear parts (like canopies) I use wood glue because it dries clear and doesn’t melt the plastic like CA or Plastic model glue. But you are correct the initial bond of wood glue is not strong and needs to be clamped or held in place.
Your biggest obstacle is if you are going to be bending the acrylic and using the glue to overcome the tension of bending, that is going to be problematic.
I would suggest trying to bend a piece of acrylic over the frame and heat it with a hair dryer to try to get it to remain in the correct shape prior to gluing. Then you can use wood glue like Weldbond. Similar to pre bending wood on a hull. The other obstacle is you would need a second frame just to form your acrylic on.
In conclusion, trying to glue straight pieces on a curve will be a challenge.
Best of luck in finding a solution.

 

[Dean62]

One last thing, you could try wrapping the acrylic around a coke can or something similar and heating it with a hair dryer or heat gun to get it to stay curved. Just a thought. Short of using a vacuform machine, you will have to be creative.

 

[Daniel20]

It would seem as long as the glue bonds to the paint and plastic window and the paint stays bonded to the brass/copper frame you would achieve success. Have you considered some form of mechanical coupling devices to positively clamp the poly to the frame. I know that is an idea way easier said than done.

 

[Peter Voogt]

After returning from our little break in Spain and preparing my fourth article on this build for the NLR-periodic, I had no other choice but to focus on the windshield again. Before we left for Spain I had the basic hardware available and fit checked, so it was just a matter of making things permanent and primer and paint the resulting assembly.
For the top center piece I was able to use thin copper sheet, made available by @Peter Voogt. It was the perfect application for this sheet; easy to cut and easy to apply.
Needless to say I'm rather please with the results sofar:

View attachment 575619

However, here is the next challenge: installing the transparancies.
Making the templates for the developed acryl and using those to cut the acryl was a breeze. The challenge is in finding the most effective way to bond the windshields to the frame.
A year and a half ago I tried using CA adhesive for this purpose on my 1:48 first try at this model. Needless to say that that didn't end well. Positive control over the addition and flow is very poor with CA and any spill gives a distortion of the transparancy.
I did some research on the all-knowing internet, where I found a tip to use wood glue, since that cures to an almost invisible layer. Tried that and disregarded it almost straight away, the wood glue doesn't provide any adhesion whatsoever, nor to the brass frame, nor to acryl.

The latest tests involve a Revell plastic adhesive with a very low viscosity, almost waterlike.
The initial result is promising, but need more testing before I'm convinced, plus I might be tempted to try other adhesives as well.
The picture below shows the Revell adhesive used and the very simple test specimen:

View attachment 575622

View attachment 575620

View attachment 575621

Painted in color it’s looking very nice, Johan. Be sure the transparent parts are in complete shape. Because you will glue on painted parts. With a bit of tension it will rip of the paint.
Regards, Peter