Interview on FlightSim.com

While I'm temporary distanced from my PC I got chance to do an interview for FlightSim.com. For those who missed it here it is.

Texturing I: The engine

For several reasons I decided to texture the engine before I move to the cockpit modeling. I never did texturing of 3D model before so I was expecting a lot of learning. Also wanted to check how my way of modeling affects texturing. For example I found that repetitive parts is better to duplicate by Alt+Enter in the blender so later only one instance is required to be textured. All duplicates are automatically textured at the same time. It took me lot of trial and errors to figure out how things are working. One has to switch his mind from technical, modeling thinking to more artistic way :). My goal was to use just 512x512 texture while level of details will be still good. To acheive this I could not simply unwrap all the parts and paint them. I had to think how to reuse particular areas of the texture. With combination of specular and normal map pretty satisfying results were acheived.

      

Some things may still change as I learn some new techniques in future. But for now I'm quite satisfied with the result. I prepared one "trailer" video displaying the textured engine using dramatic soundrack ]:-). Enjoy!

   -Joe

Blender Mdelling VIII: Engine pt. 3

The engine unit is getting close to finish. Though there are still some minor details to be added, it already looks great. I successfully applied main animations related to the engine extension. Most challenging was a properller-stopper assembly. I wanted to make it fully working as in real thing. It is a very little detail however the final result adds a lot to the realism of the whole model. The spring in that assembly is also fully animated. It means it extends/compresses as the lever catches the "pad". I found it very hard to notice, so it is probably just for making me feel better about it :) . FPS are still very good. Just couple less then default X-Plane's ASK21. So no need to worry about over-animating.


Applied animations:
 Engine unit extension
 Doors opening
 Working Propeller-stopper
 Propeller rotation
 Muffler disconnecting

I started to intensively work on the finalization of the DG 1001 M . It means I work on it almost on daily basis. Goal is to deliver one nice X-mas present :). But can't promise anything yet. At least you may be looking for more often updates ;)

See you soon!
-Joe

Blender Modeling VIII: Engine pt.2

Thanks to DG1001M service manual I managed to model the engine unit frame quite precisely. Though it took its time to read all the detail drawings and get all pieces one by one assembled into one big picture.

 

Then I continued adding various sub-assemblies onto it. Interesting one is a propeller-stopper. It blocks propeller from rotating while engine is retracted or in process of extension/retraction. Sure I will animate this later :). I really enjoy getting the things to work as in real. (Part attached to a fuselage are not displayed on the pic)






Another interesting part to model was the propeller. I started with low poly model which did't look good enough for me in closer look so I decided to go for more polys to achieve this...

Finally I was curious to see how it will perform in the xpl. Here is a short clip showing extension of the engine and start up...

And the whole engine as it looks currently:

 Next I will continue adding hoses (hm, how could it work without them :D ), screws and other details to finalize the whole engine assembly.

-Joe

Blender Modeling VIII: Engine pt.1

It's been long time since my last update. Time I can dedicate to the project has reduced recently. But don't worry the project is not dead :) I was working on accurate model of the engine. In first part I modeled the engine itself. In second part I will add propeller and muffler assemblies. The model might look like wasting of polygons but actually each piece I modeled using as few polys as possible. At the end, once the engine will be placed in the glider I will do another polygons reduction (not visible ones).

-Joe

Animation: Exterior animations

All exterior animations were successfully implemented into the model. I was curious how all the things done will look in the x-plane. I did apply basic white painting (let the color dry :) ) and hurray to the runway.

-Joe

Animation: Gear suspension

I managed to animate the spring compression exactly as desired. Final result can be seen in the following video captured in X-plane 10 (with default quick time.)

-Joe

Blender Modeling VII: Ready for animation

I finished with modeling exterior for now. All parts needed to set up external animations (main gear, tail wheel, gear doors, speed brakes, control surfaces,...) are done so the model is ready for animation applying.
Here are some of my render attempts in the blender

Blender modeling VI: Tail wheel

After a little pause I did continue with a tail wheel assembly. The tail wheel is steerable allowing the plane to taxi. It is linked to the rudder control by springs. Modeling included also improvement of fuselage shape around the tail wheel.

Tail wheel

Now I will continue with completion of exterior model, adding some details. Adding animations and some basic texturing. Then I'd like to check it in xplane, see how it all works.

-Joe

Blender modeling V: Wings

I wanted to model wings as they are in real. It means that wings can be detached from the fuselage. Thus it required a bit of work also on the fuselage to achieve a nice and smooth transition to the point of connection. Also here I tried to model the rounding as close as possible to reality. I did notice that it is not equal around the whole airfoil shape.

Ailerons are modeled so that one can notice them even in neutral position. A little gaps did the job very well.

Aileron link is modeled as a separated mesh. There is not much rounding visible (on real thing) where it connects to wing/aileron surface. There are some details where more sufficient will be to use normal map or just texture.

Aileron link

I'm thinking about having an option between 18 m and 20 m wingspan. But first I wnat to have done the 20 m one then I'll be playing more with options (e.g a variant w/o engine and 18 m wingspan).

Last thing to model on wings was an opening for airbrakes. The airbrakes will be modeled separately. And here is the whole model for now. Starting to look like a glider :)

Next I will continue with airbrakes modeling. Then adding the main gear assembly and modeling tail wheel (steerable).

-Joe

Blender modeling IV: Main gear

Yes I planned to do wings as next.Wings modeling is already in progress but not done yet (less free time at home last week). Fortunately I had some spare time in my job so I decided to do some modeling there as well (don't tell my boss ;) ) in parallel. Main wheel assembly seemed to be a good choice. I downloaded a service manual for DG 1001M where pretty detailed drawing of the main gear is and stared modeling. And here is the result...

Colors were applied just for better separation of parts. The spring differs a little bit. Original has a more rectangular like profile comparing to the circular I have used. I'm aiming to animate compresion of the spring so the rectangular profile would be a problem then. Also I do intend to use normal maps for texturing so some more details will be added.

Next (hopefully) the wings will be done :)

-Joe

Blender modeling III: Tail

I did continue with the tail area. Modeled horizontal and vertical stabilizer. Moving control surfaces.

And here all together with some mterials applied...

Nex I want to do wings.

-Joe

Blender modeling II: Canopy

Main parts of the canopy are glass and frame. It was quite hard to model the frame (visible from inside) with proportions that will match ones that can be seen on varios pictures. I decided to model the frame in one part together with piece of glass that is not translucent (white) to ease up later texturing.

Canopy frame with part of glass.

It took me three attempts to get about proper proportions. For now I'm not adding any bevels to fine out edges or similar. There are still some details to add.

This is how looks the whole canopy

And some teaser here as well ;)

There are some parts of the cockpit I made before I "oficially" started the project.

-Joe

Blender modeling I: Fuselage

The x-plane's plane maker supports exporting all parts to *.obj format so one can use them later to create more detailed replacements. However I found the mesh density not sufficient for further fine modeling. So I decided to create the fuselage from scratch. Prepared high resolution top and side view pics (from a  pdf downloaded from the DG's homepage). Blender base unit equals to 1 m so scaling and positioning of the background images was easy. To double check and comparsion I put original plane maker parts in other layer.

I decided to use 36 verticies for a fuselage section (compared to 16 from the plane maker). Sections spread is 0,2 m in most cases. Result looks as following:

Scaling the sections by X and Y axes however results in model that follows just max dimensions in the given view. Resulting section shape is more less elliptical. From looking at dozens of photos :) I found the DG1001 fuselage in cockpit area is not narrowing towards the bottom so steep as I get by simple scaling. Looking at several photos I figured out how it should  be approximately...

And here it is. The fuselage...

Next I'll continue with canopy. Some more serious modeling :)

Flight model

Finally! I managed to set the flight model to match the speed polar with acceptable accuracy. What it is acceptable? Well the L/D ratio of the flight model follows the speed polar with max 2.5% error. It means, for example if desired L/D = 43 @ 79 km/h then my flight model is in range +/-1 so the L/D equals say 42. For given speed of 79 km/h the glider should sink about 0.51 m/s. With L/D = 42 it will sink about 0.52 m/s. For now I'm fully satisfied with the result.
All tuning I performed with load of 490 kg (wing load = 28 kg/m2). X-plane flight model calculations seem to do a good job as I found the flight model matches the speed polar also for wing load = 42.8 kg/m2.
Here is the acheived flight model speed polar vs. real life* speed polar for wing load 28kg/m2

Speed polar 28 kg/m2

Center of Gravity II / Airfoil tuning

Last time I came to some wrong conclusions about force position relationship with airfoil shape/type. Moving the wing so that its 25% of chord will match approximately plane's CoG did not help much. Due the wing's shape and x-plane's internal calculations other than lift (induced drag etc) I still had to do some guessing on exact distance to shift the wing, including moment coef. Thus I did a little research and found something (http://wright.nasa.gov/airplane/ac.html). In general ALL airfoils for subsonic speeds are considered to have their liftforce at 25% plus a moment the airfoil produces. It is called Aerodynamic Center. And this is exaclty what x-plane expects user to enter. So I moved the wing back to its original position and continued with tuning the lift, drag and moment characteristics of the airfoil.

Based on wing area, desired speed and sink for every 5 km/h step of the speed polar range I calculated desired lift and drag coeficients I want the x-plane to end up. Then I started adding/reducing lift and drag coeficients of the airfoil based on actual results I got from stable glide at given speed. However this is still too much time consuming and I didn't reach my desired accuracy this way so far. As there are not all internal calculations of x-plane's flight model known (at least for me) I cant use exact math here. But I would be satisfied with some approximation of the unknown making difference between wing's lift and drag coefs vs. x-plane's lift and drag coefs. I assume that these differencies will be always same (approximately) at same angle of attack representing the unknown x-plane flight model calculation. I will play with these differences whether as percentages or constant numbers and hopefully get some expected results :)

DG 1001M video portrait

 Currently I'm deep in process of airfoil characteristics tuning to achieve proper speeds as plotted by polar diagram. DG company treats all data related to the HQK51 airfoil as classified so I did not succeeded with asking them for help. I thought that only airfoil shape itself is being kept as secret however found that  Cl, Cd, Cm characteristics are secret as well. Though most probably I would not use them direclty the data would give me a clue. Now I have to figure the data by reverse engineering along with trial-error method. So its a bit time consuming.

In the mean time here is a nice video portrait of the real DG 1001M. What a beauty...

X-plane 10 demo is out!

Well, that maight change a lot in my further development of the DG 1001M.

From X-Plane Wiki

Wondering what's changed from Version 9 to Version 10? The biggest new features are:

• a brand new system for auto-generating plausible scenery for the entire world,
• a new, highly-detailed cloud and weather rendering system,
• a flight model with improved accuracy and even greater power,
• a completely redone ATC and AI Aircraft system, and
• the ability to more fully utilize multi-core CPUs for enhanced realism.


Center of gravity

I was in process of tuning the glider's flight model to follow speed polar of the real one. However I noticed that model is quite much pitching up. I had to push hard forward to get the nose down. Definitely was something wrong with the balance. I did open the plane maker to double check the CoG position and wings position. Everything seemed to be OK. Somewhere over the x-plne forums I found that x-plane calculates resulting lift forces at 25% of wing's chord for any airfoil. Actually this can be visualized within the x-plane by pressing "/" key.

Clearly it is at 1/4 of the chord lenght


Gliders usually have laminar airfoils that may have resulting force at about 50% of the chord. That means the CoG (defined as in real DG1001M) is too much off of the resulting lift force in the X-plane. Thus a large moment is produced and the aircraft tends to pitch up.

So how to solve this problem?
A) Solution might be to simply change the CoG position in the plane maker. However that breakes designed balance for ground operation as the CoG moves towards the main wheel.
B) Move the wings so that chord's 25% (1/4) will match desired position of the CoG. Then just hide the wings and replace them by ones imported as misc. object that will not be taken into flight model calculation.
C) Calculate airfoil's moment correction.

I'll go for B). This sulution seems to be most accurate at given circumstances

B) solution applied

Edit:
As I found later, there is no need to move the lift force (wings). I misunderstood some things here. So the solution is C). Explanation is here.

Plane Maker model

Following exact dimensions obtained from official DG home page I created a base model in the Plane maker tool. This model will x-plane use for flight model calculations. Also it will be used as a reference for detailed model I'm going to create in Blender. That one then replces visually the simple model from plane maker.