Tuesday, October 28, 2008

Progress on the manufacturing / process side development

As with the aerodynamics, I am also in a continuous learning mode with composite fabrication and also metalwork. We just purchased a TIG welding machine. That seems like a welding machine that have courage to try out, it is almost like using gas-welding but with a little arc. Hmm. like a little tesla-coil? How the arc behaves seems to be adjustable and e.g. it seems possible to avoid the crater when stopping weld by adjusting the time how long it takes for the arc to diminish. There are many adjustments in the machine and there is lot to learn. And we tried yesterday. Welding aluminum is tricky, it suddenly melts without prior warning. And even after that, it continues to melt more, if I didn't pay attention how long I heated it up. Anyway, seems like a fun challenge to master TIG-welding of aluminum. These things may be obvious for professional welders, but you know, they don't teach welding to engineers. One must start from somewhere. I will use the TIG-welding machine for construction of the big CNC machine that we have been planning with Kate for quite some time by now. A big CNC is needed for creating fuselage and wing plugs. Doing it inaccurately manually seems like great waste of time (have been trying and have found that it does not pay off, a better method is worth to be investigated - I don't take any "facts" for granted, unless I agree with the results and have compared the method to alternatives and found it to be the best for that purpose (by the way, different parts may require different kind of construction method, optimal is not always only one method)).

I have been researching also alternative materials since I obtained the Cozy MKIV plans (which I am not building right now). So I have pretty unused 20 kg can of MGS L285. Nothing wrong with the epoxy, but I just found out a better epoxy: The Hybtonite obviously - the carbon nanotube epoxy. The price seems competitive with the MGS (read: the MGS is overpriced because of shipping costs from Germany) and with about the same amount of money I could as well use this "breakthrough material". It does not change the world by itself, but it can add some welcome stiffness to pieces that might be otherwise too flexible. If I hadn't have this 20 kg unused can of MGS, I would be screaming and ordering a 20 kg can of Hybtonite right now. But having this unused epoxy in the garage a kind of slows the process down since I have lots of money invested in that can and the epoxy has limited shelf life. The Amroy representative is saying that the carbon nanotube material should be as safe as any other composite material (read: not more hazardous than epoxy is already, which is hazardous by definition).

I have been discussing off-line with one UAV/RC-plane designer. He has given me lots of valuable links. I may publish some of them sometime later on this blog, so stay tuned. I am not mentioning his name now, because I am not sure if he wants to be mentioned, but anyway, I find the information found this way quite interesting and helpful. As I have been reading these documents, it has also occurred to me sometimes, that what if the configuration layout would have looks and styling as one major parameter. In my opinion, B2 way the coolest publicly known aircraft out there. So I kind of love flying wings. But I have many reasons to not be thinking of designing a flying wing, for aerodynamic and stability standpoint. But one of the configurations (that I have known before of course, but these documents were kind of reminding me about those, that some find them actually useful over the conventional configuration) - the joint wing. What if you take a B2, use no twist - ie. make a normal main wing, and put a inverted V-tail into it in a box wing configuration so that the inverted V-tail starts from the wing tips, and it avoids yet another intersection by not connecting to the fuselage anywhere. This might make the controls a bit tricky, would mean wire in mechanical control rather than push-rods. Or maybe it could be a hybrid of fly-by-wire and manual control: aileron control could be manual and the elevator and rudder (and the mixing of the two) could be handled with electronics and servo motors would drive these surfaces. Would require very powerful servo motors though (needs to be very fast and very strong). But I have been sometimes kind of thinking this kind of fly-by-wire. Before someone screams that fly-by-wire takes hundreds of years to develop, I would like to remind that it is simple RC-plane technology that people are using all the time in the simplest form - fly-by-wire does not need to mean computerized flight controls and a aerodynamically unstable aircraft by definition. A electric wire weights less than a push-rod anyway when the length is very long (e.g. high aspect ratio wing, and in this case, something that starts from the wing tip). This configuration would make the cockpit very wide and not very tall. The looks would be compromised quite quickly if the cockpit part would protrude significantly from the wing. Obviously the cockpit section would be seamlessly blended into the wing. The interesting challenge here would be: how to make that work okay and minimize associated penalties rather than the motivation to choose this would be some parameter obtained from this configuration. At least it is that way until it is proven that this unorthodox configuration could be any good. At least it could be fun to make a RC-plane like that. And I would paint it to black. Full size plane would be trickier with the color, but there is high Tg Hybtonite available too. A realistic process could be infusion moulding with the hybtonite epoxy (I will investigate this at some point in the future, investing in process can pay back in construction phase significantly, instead of spending 20 years for sanding, I rather think first couple of years and try to optimize the actual construction work to not take 20 years). This would be a kind of alternative for carbon/glass prepregs.

Monday, October 27, 2008

Xfoil for Ubuntu Intrepid

You may find out that Xfoil is not in the Ubuntu repositories. For compiling the source package, you need g77 compiler, but that is not included to Ubuntu Intrepid repositories right now and getting it to work from the source seems to be a lot of trouble. Here is what I found, after some digging, a ready made package which installed on Ubuntu Intrepid fine:


Download the file and install with

sudo dpkg -i xfoil_6.97-1_i386.deb

If some of the dependending libraries are missing, just install those from Intrepid repository and it works fine without problems. Have fun!

For more reading about XFoil, please see:
Xfoil manual
Xfoil tutorial with illustrations
Terrabreak.org XFoil tutorial
More on Xfoil at mh-aerotools

Sunday, October 26, 2008

Fuselage cross sections with iRhino

I described earlier on one blog entry how to make fuselage cross sections with iRhino easily. Here is the illustration of the lofted object cut to cross sections:

Cross sections, perspective:

Cross sections from front (simplified):

The source model:

The lofting capabilities of iRhino are awesome, it is easy to create shapes that would be impossible to come up with 2D cad models. I continue to be amazed with the quality of Rhino and I am also more and more convinced that there is no need for a 2D drawing program, I can do everything with Rhino. We are going to model our house next (which will not be discussed on this blog because it is not on-topic), while it is excellent for uniform 3D-shapes, it works so nicely with 2D shapes as well that it would be quite lame to spend all the time for nothing with AutoCAD (I am still getting shivers about the bad user interface, how simple things could take enormous amount of time to do and how innovation could get killed by the tool, we used to use that program when I was doing my studies, it is like completely from different planet than Rhino, and it is not a compliment for AutoCAD) where the work can be completed in matter of minutes in the Rhino..

Thursday, October 23, 2008

Feasibility of air travel on short distances

I, like many others, have been thinking the state of the current air travel system. With all the security checks and check-ins, the travel time becomes long. And easily using car, train or bus wins the passenger plane in the spent time for traveling from place A to B.

In other words, if you fly from Helsinki to Tampere, you can expect the check-in etc. to take at least 1.5 hours prior to the flight and then the flight takes maybe 0.5 hours. It might be also late, cancelled etc., and if this is a connecting flight, you may need to wait for your flight for another 5 hours sitting at the airport. On the other side you are waiting for your luggage to come, and it may take easily 0.5 hours.

So the shortest time with check-in luggage to travel from Helsinki to Tampere is maybe 1.5+0.5+0.5 hours = 2.5 hours. You may have also spent 60-70 euros for taxi from home to the airport, and on the other side the same amount of money to taxi. This makes moving from Helsinki to Tampere to be 180/2.5 = 72 km/h.

What about if this was a connecting flight that you waited for 5 hours and which was one hour late. That makes 5+1+0.5+0.5 = 7 hours. This makes the speed 180/7.0 = 25 km/h. You could beat the plane with a bicycle!

How about if you used a light aircraft to fly by yourself:
- Getting to the airport takes the same time, although it is easier to get to the Malmi airport than to Helsinki-Vantaa with public transportation without paying the large Taxi fee.
- Doing pre-flight check for the plane takes 0.5 hours. If you are quick, you have filed flight plan etc. during this time too. You may be able to speed this up if you are not flying alone.
- If the plane flies 222 km/h on average (includes takeoff and landing), taxi tie down etc. time is accounted with +0.5 hours (includes both airports), the total time to fly to Tampere would be: 0.5 + 0.5 + 180/222.0 = 1.81 hours. This is 99 km/h.

So the slow light general aviation plane is faster than the airliner on this trip. It doesn't win use of car though, getting to the airport and from the another airport takes time. But it wins train, because in case of train, you would have to get to the train station, and get from the train station to your destination with public transportation, which adds easily 0.5 hours on both ends, even Pendolino is therefore slower than the private car on this distance, so it is not better than using the personal aircraft.

How about if the light plane was a bit faster. It would travel 300 km/h. The travel time would become 1.6 hours. This is 112.5 km/h. Actually you can save fuel on Toyota Prius if you travel 112.5 km/h. And you are sooner directly at your destination. Not bad though for the plane. Wins the airliner hands down even in the best case.

If the distance was a bit longer, it would change the other way. The private plane would be a lot faster way to travel than car. And the airline would still have the overhead associated with security checks, package check-ins, package claims etc. For example, already if you would be going to Kuopio or Jyväskylä, the personal aircraft would be faster than the private car. And still the airliner would be the loser in the speed.

If you would go to e.g. Tallinn, Estonia, then the private plane would be excellent choice. That is because you can't go there by car, you have to use some transportation in between (e.g. boat). Even fastest boats are slow compared to even small ultralight aircraft. You could take the airliner, but it would take very long to get to the destination because of the overhead taking place at the airport. Instead if you took off with private plane, the overhead can be made smaller.

This of course requires that the plane could be flown in all weather and it would be simple to operate, with no need to do complicated tasks prior to flight in the pre-flight check. Something that would be for personal travel like a family car rather than for "flying sport". And the plane should be very low drag and very high efficiency design to make it compete in the fuel burn with the car (competing with e.g. Toyota Prius with a plane is very tough - the fuel budget for 100 km would be about 5 liters to be equal). Many current aircraft are not like that. But I feel that there would be use for that kind of planes, and this would not be impossible.

Here is by the way a video I recorded last May in the California trip. This video is about flight from Mojave Space Port to Palo Alto.

Cirrus SR20 flight from Mojave to Palo Alto (raw footage) from Karoliina Salminen on Vimeo.

Tuesday, October 21, 2008

Eclipse ECJ

Here is Gizmag's article about Eclipse EJC:


Pictures on Airliners.net:

Airliners.net: Eclipse ECJ

Saturday, October 18, 2008

Suction stabilization for low fineness ratio pusher engine pod

I got an idea how to achieve suction to the read of the engine pod.
* The prop is located just after the laminar-turbulent transition to the pod and the remainings of the pod is a very large spinner which is open from the center.
* The air tunnel inside the pod has venturi-shape.
* There are tunnels that connect the venturi tube and the ring that is supposed to have suction.
* Airflow (which is used to engine cooling) inside the venturi (helped with the propeller part that is inside the pod) causes suction to the rear of the pod. The air exits at the end of the venturi tube, which happens to be the center of the spinner.
* The exhaust in the center makes the cut aft end of the spinner to still maintain low drag, it functions in the same way as the rear cut fuselages in jets

I have not tested this idea and don't know it it would work, but I think it would be pretty easy to try out in the model scale, even with an electric motor. This interests me enough that I think I am going to try it out of someone doesn't tell me (with better knowledge, as a fact that has been proven and tested) that it is not gonna work.

I hereby license this invention under the terms and conditions of GNU General Public License, version 3, or any later version. (C) 2008 Karoliina Salminen. All rights reserved. By reading this text, you aknowledge this and agree with the terms and conditions of the GPL license.

Friday, October 17, 2008

iRhino learnings of Today

Today a colleague (Jani Ylinen, a graphics designer who knows everything about Rhino and Maya) from Nokia helped me out with Rhino a bit. So here is what I learned today.

If there is need to make 2D cross sections of for example of the fuselage, it can be done as follows:

1. Create a rectangular surface.
2. Make rectangular array of it. Adjust proper step to proper direction and use appropriate number of copies. E.g. 100 cross sections, one per each 5 cm for example.
3. Then choose Object intersection. Select all items (the rectangles plus the 3D model that you are going to cut apart)
4. Hit enter and wait that iRhino does the processing. It is slow in the current alpha-version.
5. Move the cross sections to another layer
6. Hide the 3D object layer
7. And you have cross sections. You can export these to in dxf format to for example to Qcad and process them further there. E.g. you can plot them to paper. Printing from Rhino is possible as well, but it prints the zoom level of a view that is currently present, and the scale you get can be about anything (not something that you can repeat for each model and do exactly the same scale drawings on paper each time, does not succeed with Rhino printing capabilities).

Jani also showed how to do radius. Select radius tool, select surfaces and type the radius and hit enter. Magically the radius appears to the piece with amazing accuracy.

There is also a silhouette function that makes a 2D projection out of the wireframe. You can propably utilize that in a 2D Cad, e.g. QCad (or Autocad if you are wealthy enough to have the overpriced licence to that outdated software).

I learned today that actually Rhino can be used for technical drawing without using more traditional technical drawing programs. You need to keep your model history with layers manually (if you change some cross section for example, you need to loft again), but with some work, it seems to be all you need. Also measurements can be handled, but you need to maintain them manually too, if you change some shape, you may need to update your dimension as well. With cutaways with the cross sections and the silhouette function, it seems that all sorts of technical drawing can be done with Rhino. It is different and some things are very manual, but on the other hand, as a bonus, the 3D side is so blazingly good that there is nothing that compares with it in user friendliness and expressivity. You can really create with this tool and about everything is there, you just need to discover all the functions.

Seems like the price-value ratio of Rhino is exceptionally good. With one thousand you can get so nice tool that it actually is better and especially a lot easier to use than overpriced Autodesk tools. This is how the design is done in the future for sure.

I am also downloading the Maya personal edition for Mac now. I plan to try it out for rendering models modeled in iRhino.

Many thanks to Jani for guidance with graphics software. It is very much fun to learn new things.

RC Advisor

Carlos from RCAdvisor commented my one post and I decided to check out his site. I created user account there etc. I was really amazed the RC Calculator, it not only has quite amazing features for model makers, but it also seems to have quite interesting animated UI, I didn't know that this kind of stuff can be nowadays done with Java (or is it flash?). I haven't had time to yet surf what all is on this site, but it looks quite comprehensive and promising and I will for sure look further into it. Indeed, maybe I find some tips for the twin concept RC-scale model I am going to do. Thanks Carlos for your link!

the link to the RCAdvisor

Thursday, October 16, 2008

Trying out different configuration layout for the twin concept

I decided to post here a snapshot of one of my Rhino-models. It has now struts which hold the engines out of the wing surface. The canopy was also replaced with windows.

Monday, October 13, 2008

Modern and even future concepts from over 70 years ago

There are interesting similarities in the old Luftwaffe aircraft concepts to the modern aircraft flying now:

For example:
Similar to Rutan Boomerang
Similar to Adam A700 (Originally designed by Burt Rutan)
Similar to NASA Oblique wing (which was by the way done by Scaled Composites / Burt Rutan)
Almost like the Rutan SpaceShipOne
Some features of Rutan WhiteKnight 1
Wing dihedral and anhedral similar to Rutan Proteus
From up, this could be mistaken to Rutan Long-Ez
Hey I can find similarities to Rutan Vari-Viggen here
Here is the WhiteKnight 2 configuration obviously
Almost like Adam A500 (originally designed by Burt Rutan)

Ok, then what about these:

RAM-jet, back in 1946
First commercial jet aircraft was DeHaviland Comet. But was it invented there? Doesn't this have quite recognizable look. The airliners still have this configuration and look.

And finally, but not least, this one:

Governments are still flying people into orbit with less modern hardware than this, and the idea to this one was from 1929 originally!. Think of it - first flights to space were super-ancient designs (non-aerodynamic rockets) instead of what was thought tens of years earlier already. Even Space Shuttle is quite clumsy compared to what this could have been. I would not be surprised to see someday a rendering on a page of a science magazine, which would look exactly like this and have for example the Northrop-Grumman -logo on it. The sad part was that this was only considered as a bomber, everything was some sort of warplane, it somehow didn't occur to people back then that they could have done the first human space flight earlier than it was done. I wonder why wasn't Dr. Säger utilized in the space programs which followed couple of tens of years later. I am thinking what could be done if the rocket monorail-train was replaced with a MAGLEV-train (and the track would slope upwards inside a mountain to the altitude of couple of kilometers in a tunnel). Wow. If my hair was short, it would be most likely pointing to the ceiling now.

Seems like to design a novel plane, there are examples of about every possible configuration layout, which have been long forgotten already and nobody has maybe utilized it, it just is waiting for someone to find it. The history of unfinished aircraft concepts seems to be an interesting source for inspiration.

So, if there could have been a orbital space plane already in 1930s or so, and ramjets were thought about already back in the 1940s, how much then air travel has advanced in something like 70 years? Not at all, it seems. With modern materials and tools, the feasibility of these designs have increased, but the idea still is very old. And most modern designs are just copies of each other without anything new and creative.

Wednesday, October 8, 2008

Wing droops on laminar flow section

If you have wondered why Cirrus has the discontinuity on the wings. This may answer to that to some extent. I have not found any factual information about the airfoil section used on the Cirrus other than that it is a natural laminar flow section. Cirrus VK-30 used the Jeff Viken NLF414F airfoil. I don't know if the SR20/SR22 uses the same airfoil or a different NLF section.

Anyway in this NASA tech paper it is explained how the stall resistance can be made better with the wing droop. The wing droop on the NASA test C210 actually indeed resembles the discontinuity on the Cirrus SR20/SR22 wing. Please have a look:
Wind tunnel results of the low-speed NLF(1)-0414F airfoil

Notable thing is that the Vmax-probe did not have this wing droop or any other means to prevent tip stall. And it crashed on landing possibly according to NTSB report and Bruce Carmichael's book, because of unfavorable stalling charasteristics at low Re of the airfoil caused a hard landing (which the pilot did not survive). NLF414F is not to be used without some means to prevent tip stall and to soften the otherwise very sharp stall at low Re.

Monday, October 6, 2008

Conceptual design, design requirements, high efficiency twin

Here are set of requirements I have combined for an aircraft suitable for my use case. I have been collecting these things for quite long time now, and have changed them back and forth. However, it seems like they are becoming more stable now:

- Two engines. Rotax 914 (preferably fuel injected) or similar (912 turbo conversion). Alternate engines: HKS700T (the speed may not be achieved with the HKS option). (low power engines which run on autogas are mandatory requirement)
- Range 1000 nm with three on board (mandatory requirement)
- at least 3 places (mandatory requirement, long range flights, third seat is needed for baggage and rescue equipment)
- Designed for IFR flying (mandatory requirement)
- statically stable, dynamically stable behavior (mandatory requirement)
- gentle stall (mandatory requirement (for safety))
- Cruise speed > 200 kts @ 80% power (mandatory requirement for both range and usefulness)
- Stall speed max 55 kts (mandatory requirement, for safety)
- High altitude capable (cruise at 24000 feet) (optional requirement)
- Pressurization as an option (optional requirement)
- Lightning strike protection (mandatory requirement)
- Positive climb rate with one engine out (mandatory requirement)
- Spin recovery possible (mandatory requirement)
- Very high glide ratio and long glide range when both engines out (mandatory requirement)
- BRS system (mandatory requirement, for safety)
- Spin recovery parachute (mandatory requirement, for testing safety)
- Tri-gear possibly with RG, at least the nosegear with RG mechanism (Trigear mandatory, RG optional)
- At least normal category (mandatory requirement)
- Utility category (optional requirement)
- Reasonable cost to build a prototype

Means how to achieve this:
- Selection of efficient NLF airfoils
- By minimizing fuselage and engine pod wetted area
- By minimizing skin friction drag (smooth surface, gelcoat on top of laminate and polyurethane paint on top of gelcoat)
- By utilizing laminar flow over wings and fuselage as much as possible
- By using wing geometry that has higher effective aspect ratio than actual AR
- Turbocharged engines
- Lightweight molded composite structure manufactured from carbon fiber prepregs, foam.
- By minimizing intersections and protruding elements. As clean fuselage and wing as possible. Known limitations - double slotted flaps do require external mechanism.
- By use of double slotted flaps for high Clmax.
- By use of either T-tail or V-tail for good spin recovery.
- Large fuel tanks in engine pods
- For cost effectivity, a pair of midtime Rotax 912ULs equipped with e.g. VEMS fuel injection and Garrett turbocharger is more reasonable cost than pair of stock Rotax 914s. Downside: ease of installation is lost when the engine requires more work than usual for Rotax installations. However, fuel injection is essential for safety.
- Negative sweep on main wings
- Glass cockpit (IFR requirement)

Possible configurations to achieve this:
- Twin with tractor propellers on each wing (known limitation: the propeller causes turbulent flow behind it which increases drag over the engine pod and the wing behind the propeller arc)
- Twin with pusher propellers on each wing (known limitation: the pod on front of propeller decreases the propeller efficiency)
- Push-pull configuration with twin boom tail (known limitation: front propeller disturbs the airflow to the rear propeller and the efficiency of the rear propeller decreases)

How to verify the effectiveness of each parameter:
- Calculate basic parameters for each combination where only one parameter is altered in each.
- This concept generates several different designs and each parameter is justified if it produces verifiable benefit.
- The design points that are proven to produce positive results with large enough margin are incorporated into the design if it does not overly complicate the manufacturing.

Low hanging fruits (design points known to have been succesfull in other designs):
- Double slotted flaps with a mechanism similar to Dynaero. Proven on Dynaero MCR.
- NLF-airfoils. Proven on Cirrus, Lancair and Columbia (Cessna 350 and Cessna 400 nowadays) high performance aircraft.
- Molded composite structures. Industry standard nowadays in most new aircraft designs.
- Tractor twin. Proven on most twins around.
- Pusher twin. Proven on a Polish design called Orca.
- V-tail. Proven on Beech Bonanza and later Cirrus Jet and Eclipse Jet.
- T-tail. Proven on many aircraft designs to date
- High aspect ratio on twin engine propeller driven aircraft. Proven on Diamond DA42.
- Negative sweep (moderate) found on many dual seat gliders. Small amount of negative sweep can also be found from Diamond aircraft.
- Trapezoid on wings. Easy to manufacture from composite materials, the shape is not limited by manufacturing process.
- Prepreg composites. Proven on Cirrus, Lancair, Diamond and Columbia aircraft.
- Rotax 912 series engines. Proven to be highly reliable and simple workhorse on almost all new ultralight and LSA aircraft. HKS is slowly gaining some share, but Rotax rules so far. From personal experience I also know that the Rotax engines meet their TBO, our flying club frequently runs Rotax engines to their TBO without problem without major overhauls. Claims that TBO of Rotax is just marketing and that overhaul is required well before 1000 hours to that is simply not true, this has been proven by experience, the aircraft we used to own has flown about 1000 hours, and the Rotax is the original one and runs nicely without problems and it has been in hard use because the plane was used for the first 683 hours to train student pilots. Rotax can also run on autogas (actually the preferred fuel is autogas, not 100LL).
- Glass cockpit can be made a lot simpler than tradtional gauges. Wiring behind traditional gauges is a mess and takes a lots of handwork to accomplish. Glass cockpit wiring can be made very simple and it can be highly integrated where most of the tasks are done in software rather than mechanics.

Criteria for defining success and failure
- Minimum acceptable range is 800 nm.
- Minimum acceptable payload with full fuel is three persons with no baggage.
- Minimum acceptable cruise speed at 80% is 160...170 kts on Rotax 912/100 hp. However, when comparing to Rotax 912 twins (coming and existing), the speed is no longer in the "top class", but rather below the top.
- Minimum acceptable glide ratio is 1:15. Target is more than that.
- Maximum acceptable stall speed is 55 kts. Design requires changes, if it is more than that.