I am an IT specialist. I live in Poznań, Poland. Aircraft modeling has been my hobby for over 30 years – I had finished my first paper model when I was 8 years old (1976). Since that I managed to get the Master degree in aeronautics engineering on 1994, and specialized in the CAD systems (mid-nineties). Finally I turned my attention to the general IT for the business.

Actually I manage ERP systems of a distribution company. When my children grew up a little bit, I founded a “timeslot” for my old hobby. This site is a result of some hours of my free time – and some suggestions of my dear wife.

Real models

When I was a teenager, I built many models. Most of them were created from the scratch, without any kit. Why? In Europe many things are affected by a History, written by big "H". It also applies to my models. My country, Poland, was in the years 1945-1989 a part of communistic block. It means, that we were forced to have a specific kind of economy, in that time. It was so called "planned economy", adapted from Soviet Union. Planned, but not specially efficient one. From time to time we had been lacking many everyday things, not telling about the model kits. In a small town, where I had lived, any model kit was a rare item. Usually, in cases when they appeared, they were local production, selected from very narrow assortment. How many times you can build a Soviet Yak-1, or Polish PZL-37, or PZL-23? The narrow assortment was an effect of “central planning”: the producer of model kits had just to fulfil the plan for the year, and they did not care about the quality nor customer's needs.

Some scale drawings were more available, in the books or aviation magazines. This situation forced me to learn, how to build model from the scratch, not using any kits, but the drawings, only. I experimented with the Avia B-534, in 1:50 scale, at the beginning. Finally I have decided to make my models in the 1:25 (later 1:24) scale. At the end of 1981 the political and economical situation went worse, and plastic kits finally disappeared from the shops in my town. In 1982 even the aviation magazines had not appeared. They were closed (suspended) by the government. Then, starting from 1983, the situation was slowly improving.

My last model - a 1:24 copy of P-47 Thunerboldt (created in 1985)

I made in that time many models (Il-2, P-61 “Black Widow”, Spitfire, MiG-3, F-4 Phantom,…). They were relatively crude, but sometimes really big: P-61 had the wing span of about 0.8m! None of these models survived. They often sustained damage during cleaning off the dust.

Many of the scale drawings, I used, did not have good quality. To improve my models, I had to collect as many versions of their blueprints, as possible, together with all available photos. Then I started to compare these materials, trying to create from them a scale drawing of better quality. This way – from improved drawings - the “second generation” of my 1:24 models had grown. They were much more exact. They did not survive the room cleaning, either.

The only exception is the last (and most advanced) model, which still stays in my room, today. You can see it on the picture, above. It is the P-47M, from 61th FS/ 56 FG. This airplane was flown by a Polish pilot – Capt. Boleslaw Gladych (also known in 61th FS as "Mike"), on the spring of 1945 (see some details about him).

Some details of the P-47 model - notice the rivets and panel lines

The model was formed as a 1-mm plywood skeleton, covered by the 0.7 mm thick cardboard. The fuselage was created as two symmetrical halves (for better control of the shape), then connected. The surface was covered by the filler and sanded. Imitation of alcald skin - pieces of thick aluminium foil - were glued to it. The rivets and clefts of the original skin were pressed on this aluminium cover. (Take a nail file, put a piece of aluminium foil on top, and sweep over it a needle, along a ruler. It creates straight row of rivets with just one movement). The rivets and the clefts are visible on the second picture. The bubble canopy was a technological problem – I had to try many times, to obtain acceptable bended shape. It was stamped from a hot, flat piece of the celluloid. Pilot, engine carburettor, and some other small pieces, which could not be unfolded, were made from a “modeline”. It is a kind of plasticine, that you can easily form, but it become stiff in high temperatures. So all these small parts, and the pilot, were baked like a cookies, in the oven. Whole thing was painted with nitrocellulose paint (a paint for the metal). All markings were painted manually, with a very small brush. The colors, I have applied, were based on a color scheme published at the beginning of '80 in "Skrzydlata Polska" magazine. Looking at today schemes of the same airplane one will find, that the circle around the penguin should be light blue, not white. Looking carefully, you will find also other minor errors.

Scale drawings

In 1983 I moved to Warsaw (capital city of Poland). It was possible to find more interesting materials about historical aircraft, there. I collected a lot of detailed drawings and pictures of my favourite airplanes. Using them, I prepared many scale planes – some of them in the ink. The following scenario repeated many times:

  1. I have been interested in a particular construction, and decided to build its model;
  2. Materials have been collected;
  3. Scale drawings have been made;
  4. I have found another interesting construction – return to 1.

As an effect, there were no new models, but many new, improved scale drawings!

During my first year on the Aeronautics faculty (at Warsaw Technical University), I found on a wall a small advertisement, offering interesting one-time job for the students. It was re-drawing in the ink scale planes for a new aviation book. I went to the given address, and met Andrzej Glass – the editor of this book. I had known well his name, because he was the author of a big monographic album “Polish Aviation Engineering, 1893-1939”. (The strange early date was possible because of activities of Polish painter Czeslaw Tanski. He built at the end of XIX century one of the first gliders). My colleague from primary school had this book, and I borrowed it from him many times. It contained detailed drawings of all important Polish aircraft of that period.

At the beginning, Andrzej Glass gave me a simple drawing to make - just to check my skills. Then, when I passed this test, I received my first real job – to draw in the ink the PZL-36 bomber airplane, using original drawings. (These drawings were created by somebody else). When I finished it, Andrzej Glass proposed me another job: to prepare a monograph of an aircraft. It should contain the text, 1:72 scale drawings, the cutaway, cockpit areas, and proposals of color schemes. (The color scheme proposals had to be transferred by a specialized painter into a professional, shaded color silhouettes). All this material had to be published in the first issue of a new aviation magazine.

The decision, which aircraft to describe was left to me – and I chosen McDonell Douglas F-4 “Phantom II”. It took me one year, to finish this article. (In the paralel, I had also to pass an exhausting, most difficult, second year of my studies). My colleagues from aviation faculty helped me a lot, bringing many photos and other materials. (Many of the aeronautics students are also aviation hobbyists). I prepared the drawings on a A3 sheets, in 1:48 scale. In the press their size was reduced to A4, giving the 1:72 scale – a standard practice for that time. At the end of 1988 I started to type the text of my article. It was my first contact with a PC, and a text editor. Then, finally, I directed the whole thing to the publisher at the beginning of 1989.

There were some delays. It was the last year of the old regime in Poland – and there was still the Censor Department! My article also was censored. I remember, that I placed, among other color silhouettes, a “MiG Killer” F-4E from Vietnam, with three small red stars on the air inlet. We discussed with Andrzej Glass, whether the censor would remove it, or not. Finally, after many delays, this new magazine appeared in shops eleven months later – on January 1990. It was called “Aero Hobby”. It had different color silhouettes, not based on my proposals. I do not know, if the original ones were removed by the censor, or simply the editor rearranged whole thing.

I still have the pencil originals, and some photocopies of the ink cutaway:

Original, made in pencil, side view of the F-4, with corresponding page in the magazine
Photocopy of the F-4's cutaway original, with corresponding page in the magazine

Do you know, that 20 years later I have found a "copies from copies" of my drawings, at from unknown source? It seems, that somebody found it still usable. It's nice. You can find my drawings from the original publication here. I have also described there the errors, they contain - after so many years you always know more, than before. Note: I wrote these words in 2008. In 2015, when I visited this site, it seems that my drawings were removed\modified into these drawings. As you can see they are based on my work, but lack many details

Why virtual aircraft?

The monograph of the F-4 required a great deal of precise ink drawing. Have you ever tried this? I think that after such experience I know better, what felt a monk from middle ages, rewriting pages of the manuscripts. Both jobs require similar patience and devotion. What makes them so difficult, is the fact, that you can spoil the whole thing with one movement – just make a splash of the ink on it, or similar error.

At the beginning of 1989 I could compare this difficult drawing in ink with the easy typing in a PC text editor. You can always correct your error on the computer. You can also rearrange the whole thing, even at the end of the work. Thinking of it, I was determined to make my next drawings exclusively on the PC. The first laser printers had just appeared, in that time. They gave us the real chance to obtain the printouts with the resolution and contrast, that were required by the publishers.

I started to explore possibilities of popular CAD system – AutoCAD, form Autodesk. In that time it was version 2.62. I found, that it takes a long time, to create on the computer a detailed silhouette of an airplane. On the other hand, the overall amount of work would be similar – because you can make the other things quicker. For example – usually in a monograph you have to create a drawing with small side views of all versions of the aircraft. Could you imagine, how difficult is to draw many identical silhouettes by hand? Every mistake may be interpreted by the readers as the real difference in shape between versions! And, what was also very important, in CAD system you could always to correct the error you have made. General conclusion was that it is possible to create scale drawings in the AutoCAD, in the time comparable to the traditional method. They would be much more exact, especially in the small details.

I also wanted to make cutaways of the airplane on the computer. I found, that it was rather difficult to draw a proper, spatial shape of F-4, viewed by an observer located at arbitrary point. You cannot just use a photo as the template: photos are taken with relatively long lens, proper for a picture of the “postcard” size. They look a little bit unreal on an A3 page, which was a typical format for the large cutaways. I did not dare to think about modeling minor details in 3D. The computing power of PCs was too small, at that time. I just wanted to obtain a realistic “overall shape”, looking natural on a page of magazine format. I thought that the small details one can sketch later on, in the traditional way.

On the field of 3D modeling AutoCAD 2.62 was absolute “newcomer”. That version had just a rudimentary commands to create a simple box or sphere. Of course, you could create arbitrary mesh in it. The problem was, that practically for all its vertices you had to set the location manually. There were no user-friendly tools for the meshes. General conclusion was that AutoCAD 2.62 is not suitable for detailed modeling in 3D.

The idea of 3D modeling on the PC became my “idée fixe”. I have never made another scale drawing – I dug into computers, instead. Trying to find the way into realistic 3D models of airplanes, I chose two directions:

On the first direction – application programming – I soon found myself at coding of standalone 3D editor, which would export finished objects to AutoCAD. This kind of the solution was dictated by poor performance of internal AutoCAD scripting language, dedicated for the extensions (AutoLISP). It was just too slow, and missed tools to create GUI items, that are needed in 3D modeling. So I quickly went through FORTRAN (my first programming language, learned from lectures on my faculty) to Pascal. (More exactly: Turbo Pascal form Borland. I had learned it from the books). Then I had started to use C/C++…

The programming skills, originated from this fierce coding, gave me opportunity to find reasonably paid jobs, even during my students years. Since then I had written many programs in many, very different languages (including high-abstraction object languages, and drivers in low-level Intel 80x86 assembler).

Half of the PW-5 fuselage, waiting to be cut into ribs. The "Junior" canopy was just fitted

On 1992 my 3D editor was used, together with AutoCAD, to create a fuselage for the prototype of PW-5 glider. This glider was designed on my faculty, by the team of engineers and students, led by Roman Świtkiewicz. My task in the whole project was simple: just to create, from designer's sketches, the 3D shape of the fuselage. There was one constraint: it had to fit with a canopy, adapted from another glider (SZD-"Junior"). (It was cheaper to use an existing one for the prototype, than to make a special order for dedicated hood). I draw on the plotter many cross-sections of the fuselage, in 1:1 scale. My colleagues took it, cut from a plywood, and used as ribs, froming the "positive mockup". It was needed to obtain a negative form, which finally hosted the fuselage, made from a glass composite. After prototype flight, PW-5 geometry was converted into CADDS5 system, for serial production. Over 250 PW-5s have been built, since then.

PW-5 prototype. Notice the shape of the "Junior" canopy.
Serial PW-5. It uses different, enhanced canopy (compare the frames behind pilot's head)

I have worked for many years on architecture of “ideal” 3D editor. (of course, from time to time – I had a lot of many other things to do). I wanted use it to make the airplane models, because the price of even cheapest program with 3D capabilities was too high for playing it at home. Joking a little bit: it was programmer’s quest for the Holy Grail of 3D modelling. I was always in details, learning Open GL, or trying to identify the structure of service – oriented approach, or working on other aspects of the Program. I was always far away from the ultimate goal, just looking at the weak light of future possibilities…

The second direction – other CAD systems - led me to an individual path of the university studies: “Computer Aided Systems in Airplane Design”. During this course I learned many industrial CAD/CAM programs of the mid-nineties: Unigraphics, CADDS, Catia, Pro-Enginieer, and the others. After a year it become evident for me, that they are best suited for making thousands of detailed machine part drawings, and cutting them on the CNC machines. For more complicated, overall shapes they were not so sound. I was surprised discovering, that the geometry of new cars was digitalized from clay models, modeled by hand! The vertices, obtained this way, were transformed, with significant amount of work, into NURBS patches.

In the meantime the first programs for 3D animations have appeared. I did not paid too much attention to them. At the beginning, their PC versions could create relatively crude, simple objects, only. I had no possibilities to look at them on more advanced platforms.

Finally I graduated, and accepted a job in an IT company, specialized in the CAD systems. I spent there two years, and then went into another company, dealing with general IT services. (In that times there was an outstanding demand for this kind of specialists). I have never worked as an aeronautics engineer.

For some years after graduation I have tried to continue my quest for “ideal 3D modeler”. I prepared, on 1996, a study of actual state in user interfaces for 3D modeling. It was concentrated on the UI, because it is the essential part of every modeler - the overall program quality relies on it. I had used my comparison of CAD systems from the university studies. The description of 3D Studio Max, known from my professional work, was also included in this comparison, as an example of the 3D animation program. I explored the content of growing rapidly Internet sites, founding there many interesting materials. There were no 3D public-domain programs, in that time. Just some free (to download) software, including the beta version of Rhinoceros. The main conclusions of this study were:

  • CAD systems had the oldest user interface. It was cluttered by many detailed commands. In AutoCAD there were over five ways of creating an arc! Usually you had to answer many questions or fill many fields in a dialog box, before anything was created. I named this approach “the wizard’s apprentice” – you have to issue a complicated spell, then “puff!” – and look at the result. It often was not exactly what you want, so you have to erase it and try again. It is not the most effective way of work. The application of direct manipulation, which is much better GUI solution than the dialog boxes, were dispersed and rare;
  • 3D animation programs or general modelers (like 3D Studio Max, or Rhinoceros) had reduced command set. Their creators more often used the direct manipulation. This made their interface more streamlined and simple;
  • New, better mathematical model was required to describe smooth, curved surfaces with holes and crearses. The NURB splines had been the industrial standard, for over 20 years, in that time. NURBS surfaces were best suited to describe a semi-rectangular, smooth “patches”. All holes, creases, or irregular edges, were serious problem, because they were anomalies for the NURBS mathematics. Existence of such item required the program to triangulate (automatically calculate new, triangular elemental faces) the whole surface. Also the “sewing” two surfaces along common edge, or creating a NURBS fillet could lead to unpredictable result. Why? Because they had to be implemented by a “program” (read: programmers) logic, not by the mathematical model itself;
  • New input and output devices were emerging (it was the time of the “virtual reality” boom). There was no special problem with the visual output. Digital goggle, with stereoscopic view, can be implemented as two-screen system. This can be handled by the Open GL driver, added to the hardware by manufacturer, in a transparent way to the application. Much more promising, but difficult to integrate with existing interfaces, was the force-feedback manipulator. It is a bi-directional device (input: the force vector, output – the reaction vector). None of analyzed programs had a software interface suited for such kind of interaction;

On 1997 I accepted another job, and dived deeply into an ERP systems for general business. I had a lot to do for many years. Finally, I finished the last exhausting implementation on 2006. Since then, I have had some amount of free time at my disposition, like most of the people, at least.

It was rather interesting experience: to come back to the same matters after nearly 10 years of hibernation! How the world of 3D modeling has changed during this time? Looking around and at the Internet, I have found that:

  • Home PC become a powerful graphic workstation. The requirements of computer games forced it to evolve this way; Now it has graphic accelerators and enough RAM to keep a very large textures. (The large textures – about 2048x2048 – are required for realistic bump maps of small details, like clefts or rivets);
  • High demand on 3D animation market created pressure for rapid development of advanced software, that could support the film and ads spots;
  • Plenty of good software has been released under GPL license, which makes them available to all;
  • New mathematical model – subdivision surfaces – has been utilized, especially by 3D animation software. Some of its “schemas” (Catmull-Clark) can be treated as more general version of the old NURB splines (for a rectangular patch of control points both models will return the same faces). Subdivision surfaces can describe very complicated shapes, with many crease edges, fillets and holes. You can also “sew” them along any edge, without a problem. From the CAD/CAM “old timers” viewpoint, the fault of subdivision surfaces is that they have no classic, analytic representation. You can always get their vertices, with any required density, but there is nothing like “equation of subdivision surface”. To calculate their derivates, you have to utilize the calculus of the infinites. It is an "ideological", not practical flaw. From programmer’s point of view, obtaining any tangent direction is not any problem – you have to pass the input vertices to the "black box" of mathematic model, and receive the result (some vectors). Which model in inside this "black box" - NURBS or subdivision surfaces - does not really matter. Both have such standard function. Subdivision "black box" offers just more functions than NURBS;
  • New input and output devices have not emerged into wide market. They remain as a niche product, especially the force feedback pointers (I had many hopes in them, 10 years ago)

So, like usual, there were bad and good news. I almost instantly started to like the subdivision surfaces, and the idea of the GPL software. The lack of development on the field of input devices was a disappointment.

In the middle of 2006, I finally stopped my approach to create a 3D modeler. At this moment, as more experienced programmer, I know how enormous is such work. Maybe authors of the other GPL systems have not always created them in a way I would wish, but they have offered the real thing, instead of my bunch of projects and ideas. So I decided to find a GPL program that would be closest to my needs. I started with careful look at the available 3D modelers on sourcefroge.net. Finally I have decided to use Blender. The detailed explanation, why I have chosen this program is available here.

On Spetember 2006 I started to experiment with Blender. I had only a very limited experience with the 3D animation software, that time. I did know nothing about the lights, textures, and other things.

My tactics was to find the proper ways for modeling in Blender at the very beginning. I have selected the La-5 fighter airplane for my first virtual model. For over one month I had just created a propeller and its spinner. I exprimented on them many methods of working with materials, textures and lights. During this process I selected two helper programs for dealing with texture images. First of them was Inkscape - for geometric, vector shapes. Second was popular GIMP - for raster images I did not try the Photoshop. When I had found that its GPL counterparts are good enough, I decided to complete whole thing using the public domain software, exclusively.

Finally, when the propeller was textured and rendered from many sides, the overall process of creating new parts was cast. It was right time to apply it to the rest of the model. I started to think, that this new kind of hobby can be signalized to other modellers – “hey, there is a new world for your aircraft!”. It was also very important that they can be created using the public domain software, only. There are many of 3D models created in many commercial programs, on the net. The examples of complete virtual aircraft created with Blender or other GPL program are extremaly rare.

Finishing the first virtual model, I have reached my goal, set about 15 years ago. When I think about it, I really appreciate all the programmers, that took part in creating the public-domain software, especially the Blender team. This site is my contribution to their effort. All models, tutorials and scripts, published here, are available under Blender Artistic License.

I will be happy, seeing pictures of your models, made in Blender - let me know about them!