Well... I must say my family thought I was out of my mind when I started this new hobby of mine - the model planes !
This was one of my favorite dreams at 14 years of age - to build a BIG radio controlled model plane. Now the timing seemed to be just perfect for it, I had the pilot license to know how the real planes behaved, I had the skill to build for example a sail boat, so why not a model plane and as the most important fact I finally had the money to make it to fly !
The idea struck to me when I visited the Smithsonian Udvar Hazy flight museum in Washington in June 2004. I saw the most beautiful (or cute) airplane there was - the Grumman Goose. It was the perfect consensus between flying and sailing, a flying boat. I photographed many of the objects in that impressive show but this plane occupied the majority of my taken photos. Something I always wanted to possess.
The idea was put aside for a month or two but then I received email from Derek Linder, a person who wanted to keep the memory of Grumman Goose alive and publish a magazine about the 345 Geese made ever. I admitted to be more than interested in the subject and enquired about possible drawings of the Goose, to be able to build a model plane of it, the size to be about seven feet wing span (a bit over two meters). He struck me by a great surprise by sending me all these constructional drawings I needed. To save your download time I am not going to add them here but whenever you feel the urge for them, just send me an email and you got them !
I stared at the drawings for a few weeks, slowly creating a set of my own to the scale of 1:7. Once the design seemed to be thorough enough I walked to the local hobby shop to enquire about the parts, radio gear, balsa etc. They were most friendly but guided me to another shop that had the best knowledge about electric motors and drivers that I was after. I had some very bad experiences about diesel model motors in my youth and that added to my background of electronics guided me towards the electrical solution.
The surprises were not over ! When I called this shop it turned out that the owners were the previous owners of my wrecked sailboat ! We spent a good time together sharing memories and photos about the Snipe I had renovated. After all, the world is quite small ! Once they saw my Goose drawings they wanted to ask their friend for an optimal wing profile that I had nothing but just copied from the original drawings. The same profile does not work properly in the small dimensions of a model plane. Lucky me !
I received a truckload of information about model planes that I had never even imagined about. Loaded with that I came home and started to turn my plans into reality. One month later I had the hull more less constructed, waiting for the specially wide balsa I had ordered by internet from a Lone Star Balsa Company in Texas.
Well, a bit over one month later there still was no balsa in sight! I contacted the company several times with no result. On the right the hull is ready and awaiting for the final covers... Once I had the time I decided to add some more material for improved ruggedness (Compare to the first version on the left). Hope it still is able to fly ! So far the hull alone without the outer covers weighs a bit over 1 kg.
I reached the next level of existence in a short while. This was not only to be the case of nice and easy woodwork that I already had gained practice for with my sailboat. This was a challenge for TOTAL construction design of an airplane ! I had to renew my original thoughts about the schedule of this project. It was not going to be a matter of a month or two but merely a two year project !
Once I had the basic hull constructed it became the turn of the dirty details. What about the landing gear ? Two in front and one in the rear. To be able to land this thing both on the earth and in the water the landing gear had to be retractable - and the way it was done in the original Goose. From the drawings I got it seemed that also the tail wheel was reeled in but I decided to skip that because of excess complexity. I could make it to turn and be watertight but having it also pulled in was just too much for the benefit it would make. Out will it hang ! On the left you can find the first version of the front landing gear attachments. In theory they seem to work all right.
Next it became really interesting ! I got the profiles for the wing. This gentleman had gone through quite an amount of work to figure out profiles that would look as the originals but operate properly in this scale. According to him the original wing profiles (NACA 23015 in the center and NACA 23009 at the tip) would work quite all right but induce unnecessarily high resistance. He therefore suggested to use Ritz 1-30-15 in the center and S3021 at the tip, but slightly modified to move out the 55% wing length stalling point. When the Ritz 1-30-15 leading edge was sharpened to 70 % of its original value and the S3021 leading edge was rounded to 150 % and its thickness increased by 10 % the stalling behavior became proper. The aerodynamic zero twist required the tip to be tilted down by 2 degrees. The results of his considerations are here. Thank you once more.
Now it was my turn to make the plans into reality. It took a couple of evenings to draw down the profiles and interpolate the profiles in between. Then I draw them all on one paper including the 2 degree tip tilt. Copy, cut out and draw on balsa. Next weekend I spent cutting the shapes out of wood. To make it a bit easier a used double-sided tape to attach two balsa sheets together and cut out both wing profiles in one action. I also added the "studs" shown in the picture below to ease the alignment of the construction. While they carried the information of vertical position and tilt angle of each profile the final assembly was going to be really easy. Once the wing was done they are just removed.
Well, I had to give it some more time. Considering the details how to make the flaps and ailerons to operate took more time that I had thought. Meanwhile the long expected balsa suddenly appeared 8 weeks after I had ordered them. It drew my attention back to the hull for a while. On the right the bottom is covered and lots of epoxy glue spread all over the seams.
Working with balsa was great fun. The wood is so soft and easy to shape. However, I decided to make some parts of the plane out of mahogany, for durability and most of all for the looks of it. On the left the mahogany nose piece is being glued in place. The newspaper is put below for splashing glue.
Then came the side walls. Suddenly the whole project started to change from a skeletal structure to something that looked real. It also became much harder because I had decided not to paint it but let the nicely finished wood to show. Lacquered and polished to shine of course. Every bump and scratch I made would show and thus every operation needed extra care.
The windows required a bit more constructional thinking. First I embedded a veneer frame to the arcs that had a bit smaller openings in it than the outer cover would have. Next I glued the balsa cover sheet with the larger openings on. As result the window openings now had a narrow border around them against which the window pieces could be glued from outside.
Before I could close the hull by adding the top pieces some wiring preparations had to be made. While the plane would be radio controlled it was of greatest importance to prevent its own electronics to disturb the radio transmission. Therefore all the cabling had to be shielded properly. On the right you can see the 12mm aluminum tubing fitted on the tail. It will carry the wiring to the two rear servos and to the tail lights. Another tube is fitted vertically next to the front end of this tube to bring the wires down to the bottom and join there the wiring of the tail wheel servo. All these shield tubes are connected together with a piece of wire and finally center grounded to the negative potential of the system. On the right you can see the tail wheel servo fitted close to the bottom with two springs and steering pins hanging loose while the tail wheel is currently removed. The two servos that operate the rudder and the elevator ( the name of the device to control the altitude) can barely be seen at the top right corner.
Well, on the right those two servos are now in a closer view for better visibility. The shield tube brings the wires as close to them as possible.
On the left is one of the on-the-go modifications shown. Finally I made up my mind and decided to add that front cargo lid to the model. If I had chosen a different set of profiles (odd) instead of the even this stage would have been much easier. Now I had to split the profile No 4 and add pieces of profiles 3 and 5 to create the cargo hole.
Then it became the moment of truth. As I already had one of the wings built up I had to take it all apart again. Let it go unattended I had skipped the wise advise of my tutors - not to build up the wing before making all the necessary openings for the wiring etc. ! They also did advice me to change the 5 x 5 mm pine rods to 10 x 5 mm ones for enough strength. On the left the result of all this is shown - I took the whole thing apart and reshaped all the wing arcs, now including also the aluminum tubes for the wiring and arcs modified for the larger pine rods.
Had to give it a break ! While it seemed to be just too much I went to the rudder department for a while. Especially that my aerodynamic friend sent me the optimal profiles for it ! He explained me that the elevator profile cannot be (vertically) symmetrical or it would be unstable. Instead, he recommended the enclosed (profile) that would have a suitable "lift " downwards.
Then I had to stop altogether. I took the whole project to my country house not to mess up our home for a prolonged period of time. I tried to work in the basement in order to keep the normal life at pace. After three weekends I finally understood my stupidity. I had been constructing the plane in a below ground level, non ventilated basement using cyanoacrylate to glue the parts together. It took me all these three weekends to figure out that the Saturday afternoon headaches that continued all day Sundays, developing themselves as a Monday morning dizzy feeling with blood running out of the nose and finally making me cough blood out of my lungs was all about the glue. That stuff is really dangerous - watch for it ! Now I have taken a pause of a couple weeks and hope to be able to continue shortly. However, the front part of the hull became extremely difficult to bend balsa over. Four times after I still have nothing but shattered pieces of balsa in my hands. Maybe, if I am heroic enough I'll enclose some photos of it here - later.
To make my life easier I brought my attention to the elevator wing for a while. It didn't take more than a couple of hours to finish the frame and cover it with a 2mm balsa sheet. On the right the whole stationary part of the wing is shown, the moving aft ailerons and tip portions still missing. Note the darker mahogany front edge strip.
On the left the internal structures of the elevator are shown. First the whole thing was glued together with bottom skins and veneer enforcements at each hinge station. Next the hinge slots were cut open and finally the top skin was glued in place. In theory.
In practice I managed to glue the right side top skin in place without cutting the slots first open :-( I then opened the left side slots, made some very careful measurements and made very narrow slots to the right side, widening them up carefully by filing. Hours later it was all done, huh. Think before acting !
The rear edge was so thin that I decided to cut the 2 mm balsa skins short and glue a readily available triangular balsa strip to make the final edge. That worked out well.
Once the parts were all in place it was the time to give them the final shape. With a grinder (used for my sail boat) I got the shape done in five seconds. On the left the elevator is finally complete. Now let's get to the rudder part to get the tail done...
I decided to give myself a bit relief. Instead of drawing everything on paper and scaling it with ruler and calculator I adapted the computing power available and draw the plans into Adobe Illustrator. How beautiful a tool it turned out to be ! I snapped the profile shape from the original plane drawings, redraw it carefully and started to copy and resize it until I had the complete rudder designed. On the right the rudder design with location of the elevator wing in red. I still don't stop wondering the weird shape of the bottom part of it. It is so thick and clumsy. All the pictures I've seen of other model planes of Goose have it so slim. Nevertheless, this was taken from the original profiles - so be it.
And here we go. The rudder frame is finally building up on the left. With the hinges and all. Holes have already been made to the arcs to allow electric wires to be lead to the navigation lights - a red blinking one on the top of the rudder and a continuously illuminated white light at the rear. Illumination angle of it to be 140 degrees as the regulations define for the rear navigation light.
The construction had to be different from the elevator because the moving ailerons were crossing in the middle. The elevator hinge was fixed to the front edge that formed its main structure. In the rudder the front edge had to be split to allow the elevator to go through it and thus another support system had to be thought of. So, I made up a rigid beam out of two 3mm pine sticks 2 mm veneers glued on both sides of it. The front edge could be made of balsa now because of no need to carry loads. On the right the construction is shown just before gluing the other side panel in place. As you may be able to see, the panel rear edges are thinned from inside to make one nice sharp edge. Shielded wiring is in place for the top beacon and a service hatch has been provided for later connections and possible wiring repairs. The wires are moving when the rudder is turned and will wear out some day, doesn't matter how well you make them to last.
The plane tail starts to be in shape, some more skin work to be done, the servo arms to be constructed and I can move forwards for the wings. All the design of details seems to be taking quite a lot of time. Anyway, it is interesting and it is why I started this project in the first place.
On the right the final version is shown, all wiring and servos enclosed in a grounded copper sheet shield that minimizes the electromagnetic interference. Note the flexible silicon wiring towards rear that will power up the rudder navigation lights.
Well, maybe I should start to put up some dates here, just to keep us all synchronized. This whole model plane project of mine started in August 2004 and today it is May 8th, 2005. I used the last month or so to concentrate on the electric motor drivers and thus did next to nothing on the wood work. That's why there has been no updates. The schematics will follow later, once they are proven to be functional. I also gave some thought to the navigation lights of the Goose. Starting from the aft a red beacon needs to be fitted on top of the rudder, throwing its light mostly horizontally. Another white navigation light must be fitted in the rear with a radiation angle of at least 140 degrees.
When I was told by my tutors that these lights are no good for outdoors and only visible in dimmed halls I decided to make the difference. I went through some LED specs and decided to use special Lumiled diodes for this purpose. One green and one red Lambertian (evenly radiating point source) emitting diodes for the wing navigation lights, one red side emitting diode for the beacon and three white Lambertian emitting ones for the rear navigation light and for the two strobes in the wing tips.
On the left is the beacon light. Check the curious lens design - first normal lens and then on top of it a mirror system that reflects the light mostly to horizontal angles - very effectively according to my experience. I still need to add a transparent hood around it to shield the mirror from dirt.
On the right is the rear navigation light. Here the lens looks more normal. The copper studs of both of the lamps are there for heat dissipation. These LED's are very powerful and need to be adequately cooled not to overheat. The beacon is only pulsed while the rear light is on all the time, that's why the size of the coolers is different.
Just to give you an idea how they perform I took a photo shown on the right. The LED's were driven at less than half of their power not to saturate the camera and still all the lighting in the room is from the LED's alone. Really bright, must be clearly visible also in bright daylight ! (Don't pay any attention to my opened up lab power supply, those 30 years old solder joints start to oxidize and need continuous service).
Actually, finally now in mid May 2005 I drew the Goose Main Diagram.pdf of my model plane. I was surprised by the complexity of it - can't I make anything simple any more ? Seems to become too professional. All the ideas of devices to be controlled are shown but still some special option control buttons need to added. (one is floating in air there waiting for its purpose)
I wanted to gain some progress on the hull finally. Boldly I decided to wrap skins around the rudder base. (see how complex that piece is from above, double curved surfaces) After one evening of trial and error I just gave it up. Too demanding, I could almost bend the balsa into shape and just when I was there it was shattered... again.
Fortunately I have kept this web page to share all the Grumman Goose enthusiasm with anybody interested in the same. One friend from southern Spain shared my interests and helped me out on this. He sent me a copy of an article describing the striping technique. Split the balsa first in narrow stripes with similarly angled edges and then glue them in place one by one. While the stripes are relatively narrow they allow themselves to be bent in two ways, allowing for almost any shape to be made.
This article was based on a misused "commonly known balsa striper" that I simply could not get from anywhere. Next to nobody is using balsa techniques today, utilizing plastics, styrofoam and polyurethane instead. My local tutors had an old striper lying in their drawers that they promised to lend. I decided not to, too much trouble to return after every use. Nobody knew how long this project of mine is going to take.
Now, two months after I was given the idea of striping I was sitting in the basement of my country house, trying to think how to proceed. Then the idea struck to me. I had it all available, it just needed to be put together properly. I took a disposable knife, a Stanley plane, an adjustable clamp, a piece of mahogany and combined them with my carpenters bench. On the left is the result.
I removed the blade from the plane and screwed a spare blade of the knife there instead, pointing well out through the bottom. I then tightened the plane upside down to the bench and fixed the piece of mahogany on it with the adjustable clamp. Requiring some carefulness I started to cut the balsa into stripes. The position of the mahogany block defined the width of the stripe and being long enough it also guided the balsa sheet to be parallel to the blade. The angle of the blade of the plane served as a good cutting angle. Not adjustable but served it purpose well. Gluing angled stripes together and smoothing them later off with sandpaper makes the seams to practically disappear. Seams in right angles are almost impossible to make vanish.
Instead of the tail I decided to take the challenge in full and move to nose again, left as it was almost six months ago when nothing seemed to solve the problem. On the right the outcome is shown - with some minor tuning the balsa finally allowed itself to be twisted to the shape ! No steaming, swearing or other magic, just apply them on in narrow enough stripes.
Something else for a change. Some electronics instead of wood work. It was June 9th that I received the circuit boards for my motor inverters. Dimensions 40 x 87 mm. A bit of soldering and software still need to be added to make the propellers move. According to my calculations these drivers will be good for 60 Amp current and up to 15 V operating voltage without overheating (900 W). Want to see the rear side ? Or the schematics ? Software listings I will keep to myself, once they exist.
Now, that I finally started to get hold of the project again (June 20th) I decided to allow my concentration to turn back to the wings again. I had taken the wing constructions apart several months ago, drilled the missing holes necessary for the wiring shields but then left the bunch of arcs intact, waiting for the becoming ideas on how to construct the flaps and ailerons. I got it all together now and decided to go for it.
Actually it took only a couple of hours for the whole wing to be in shape. Then another evening or two were required to build up the final construction with all the necessary slabs to build up the rigidity of the wings. Above are the achievements of yesterday shown, both wings done, only the center connection rods, flaps and ailerons missing. But all designed how to now.
Summer vacation started (June 27), more time for hobbies ! As gardening and sailing ;-) Some time also spared for this project. First I went to the ever-so-difficult nose and had it done with. With the cargo lid and all. Hull is now almost finished, just some skin at the rear still missing.
So, back to the wings again. Next thing was to make the veneer supports that connect the upper and lower longitudinal rods together, forming a rigid "tube".
Using the cyanoacrylate glue only outdoors turned out to be still insufficient. After revisiting the doctors I went to the hardware store and bought a gas mask that can be seen on the table at right. Finally the 'flu' started to release its grip. It was the glue all the time that made me sick for several months.
My daughter Sara wanted to make a permanent memory of my hobby by drawing a picture of her father, constructing the Goose wings at our country house balcony table. Actually the pictures on left and right show almost the same moment, just from opposite directions.
Well, it is July 24 today and my summer vacation has come to its end. Most of the time went unfortunately to gardening and sailing as I was afraid of. Nice time, though, but this project did not proceed that much. Maybe I needed a pause because it starts to resemble more and more hard labor than a hobby. Some proceeding, however, can be noted down. I shaped and glued 88 small veneer plates to connect the wing middle and rear upper and lower main support rods vertically together to form rigid beams that can take the torque applied to the wing.
Next I managed to connect the wing halves together. For this purpose I had bought some special gear - 15 x 2 mm square steel rods and a tightly tolerated square brass tubing to fit on. Never knew these things even existed before I started this hobby ! First I consulted my friends to understand how long these support rods had to be to be sufficient. Then I cut them in proper lengths and positioned both wing halves together exactly in line and fixed the brass tubing temporarily in place with the cyanoacrylate glue (wearing the gas mask, of course). Picture on the left.
Next I combined some knowledge of my sailboat renovation to the model plane building. The brass tubing had to be somehow fixed very rigidly to the wing frame. I decided to cast the brass tubing in place by epoxy glue I used to rebuild the rotted mahogany rear corners of the sailboat. For less weight the epoxy glue was filled in with plastic micro-balls to make a light but rigid gray mixture. At up right the cast wing supports are shown. Next the top veneer plates were glued in place to complete the beam structure. It turned out to be a very good and rigid construction.
Continuing steadily forwards. It is the 3rd of August, almost one year since I begun this project. The wings are getting more details on them now. On the left is one complete wing frame shown upside down where the ailerons and flaps are growing on. While I never make anything the easy way, I decided to design their operation optimal.
(By the way, note the white piece of paper 5 arcs up from bottom. It is modeling the coming carbon fiber part where the wing tip float supports are going to be fixed.)
The ailerons are not hinged at the center of the wing profile but at the top surface instead. This makes their operation asymmetrical, moving them less down than upwards. This compensates for the braking effect that is much larger when the aileron is moved downwards and leads to less compensation required by the rudder and thus to more efficient flying.
The flaps instead are hinged on a completely opposite way, pivot point being well outside and below the wing profile. Detail shown on the right (upside down). The chosen construction makes the flaps extend back while their angle is increased. This preserves the wing profile more optimal, reducing the resistance and increasing the wing lift remarkably compared to a center profile pivoted solution.
Now it was again the time for a change. While I was seeking for a proper space to machine the carbon fiber parts - a really messy job - for float supports, aileron and flap servo reinforcements - I went back to the electronics for a while. On the left one motor inverter is shown with the components soldered on. The output flywheel Schottky diodes on the left are still missing, as well as the copper beams that fit on the silvery rails to act both as heat sinks and power conductors.
At the first glimpse you may get the impression that the circuit board is pretty empty. Wrong ! To witness this see the enlarged picture on the right. There are so many components on it but - boy - so small. It took me several evenings, wearing reading glasses and a magnifying glass to build these boards. The soldering iron tip was like a needle and the solder itself resembled hair. The nowadays surface mount technology is not exactly meant for manual assembly. Look at them, some of the tiny resistors just dropped right through the 1 mm diameter component holes when slipped from my pliers ! But finally I got them done. Now only the software is missing ;-)
A couple of days later the first software saw daylight. Version 0.01 started to run on one of the boards, blinking an LED. Wow ;-) Unfortunately the other driver PCB is refusing to wake up. Have to hook up an oscilloscope to it to find the problem. Bad joint maybe.
A week later I had to give in. The only way to get the required carbon fiber parts was by the hard way. Wearing a dust mask and barely anything else I went outdoors just before the rain was to fall. I then cut the pieces into shape by a grinding wheel that threw most of the dust on me and the rest all over to the nature. Meanwhile my daughter heated up the sauna that I could wash the dust out really well. While scrubbing it off the nature washed itself. Its only carbon, after all.
Made some progress. On the left all the various parts are creeping together to form an airplane at one of these days. On the top of the cabin I have fitted two plywood parts that will be fixed to the wings as adapters. In a week or two I hope to be able to assemble them together and then finally close the tail roof. I left it open to make it easier to make the wing adaptation. The last part to make will be the cockpit, once I see how it is going to fit in.
Next I started to put it all together - finally. Made some extra supports to the wing center to bear the twist and prepared the rear fittings out of stainless steel. The wing front is fixed in place with two 6 mm plastic bolts - that give in in case of a crash down :) Weighed it by the way. So far the hull was 2,2 kg and the wing exactly the same - 4,4 kg altogether, not so bad when 7 kg is the goal.
Mid September I had struggled down the basic software stuff, getting the microprocessor in its senses and doing also something else than just blinking the LED. Both boards running now fine, it was just a matter of a cold joint. Soldering these tiny pieces is somewhat challenging. Bought a new pair of glasses, what an enormous relief ! So, got the driver electronics running, first slow and then full speed, up to 60'000 rpm ! Being an LRK-motor the actual spin will be just around 6´500 rpm that is enough for the speed required. Once I had the waveforms checked I wanted to give it a challenge and connect the motor, no prop installed so far to save my fingers.
Suddenly a nasty crack-swoosh could be heard, some smoke exiting, followed by a scent of burning electronics - and it shut down. Oh my :-/
The following inspection showed that it was my fault. Having hooked up my amp-meter between the power supply and the driver PCB gave a hard time to the tiny onboard capacitors, one of them giving up and letting its operative smoke out. On the left, just below the red wire end. Pulled it out and the operation recovered. Must be more careful next time. Once I am sure that the system will hold I'll add an accumulator as a power supply. Then a short will make everything just carbonized :)
In all my enthusiasm once I finally got the propeller running so fast that it blast my hair back I wanted to squeeze all out of it. Of course with my current limited power supply that could only deliver 6 Amps, about 10 % of the current required for full power. Trying to push it to its extremities I turned the wrong switch - and gave the system double voltage :( It blew up. And I decided to go back to the woodwork for a while.
It is now the end of October 2005. I already built three more inverters but have not powered them up yet. On weekends I have been building the ailerons and flaps in my country house basement. On the left both of the wing halves are equipped with the moving fins, servos in place and tested. For testing purposes I modified one of my inverters to read a potentiometer and deliver respective servo control pulses. With the aid of it it was easy to check the proper movements.
On the right is a close up of the flap construction, turned 40 degrees down. Wanting to make everything as good as possible I did not satisfy to just some flaps but made them "semi - Fowler". When turned down they also extend back because the hinge point is brought below the wing as I already explained earlier. Unfortunately that added quite a bit complexity to the construction. To make them fit properly I had to fit and shape every arc individually.
The servos I assembled in carbon fiber cradles that are fixed with two screws to the wing. The actuator arms connect to L-shaped carbon fiber pieces that are in turn screwed to long carbon fiber profiles that are glued along the flap and aileron rods. By the chosen construction I aim to avoid large point loads that could break the wooden construction.
Well, it is mid November and I haven't visited my wood-work-room for a while. Too busy at the office. Slowly I have made some progress with the propeller driver software. Now the inverter already runs the prop smoothly and doesn't heat up too much. The speed can be adjusted with servo control signals. On the left you can see my test setup. On top of the oscilloscope is my power supply and on it is the servo signal generator that is supplying the speed commands to the inverter. That in turn can barely be seen if you follow the red and black wires hanging left from the power supply. It is just next below the the blue LRK-motor. The three blade propeller gives a faint reflection while running quite fast.
Check out my dynamometer. A few blocks of wood nailed together, the vertical part been hinged to the base and the motor fixed on top. A fish-scale attached on top of the base block to measure the pulling force of the prop (3:1). Whole thing is attached to the table with a clamp that is also acting as my solder iron support.
While I am still refining the algorithms I haven't hooked up an accumulator to the system yet. Therefore the power is limited to 6 Amps and the maximal achieved force is not more than about 1 kg. It already creates quite a blow.
It is the beginning of December now. Because of the Finnish independence day (December 6th) we had a longer weekend to spend for example to model plane building ;-) I started by gluing two magnets to two aluminum pulse wheels, manufactured by my workmate who knows how to operate the thread mill. The two magnets were oriented in reverse, one having its P-pole up and the other down. This was to ensure their operation, whichever way the detector wanted to be sensitive. Two magnets were needed for balance. I then screwed one of the wheels down to the motor back and attached a Hall-detector next to it. It worked out just fine, right from the first time. I had to borrow a better LeCroy oscilloscope from the office to be able to really see what was going on in my inverter design. Something did not operate as it should and the maximum speed I could achieve was only about 10 % of the designed maximum. My 30-years old Trio oscilloscope was just too poor to give me the details I needed to see.
The secrets started to reveal themselves. The operating voltage was not stable enough and sudden current peaks taken by the motor caused it to drop all the way down, making the microprocessor to reset itself. No sweat, adding another large capacitor next to the driver PCB rectified the problem. On the right you can see the latest setup and on the left is the plotted current curve (red) and the rotation pulse signal picked up by the Hall detector. Here it can be clearly seen that the LRK-motor makes only one turn for seven magnetic field revolutions. With this setup the maximum current to the motor is about 50 Amps.
However, the maximal speed I was able to reach was about 180 Hz, only 20 % of the designed. As you may be able to see on the picture above right I had already added a battery to the system to give more power to the driver. It is located on top of the pile, just left from the soldering iron on top. Yes, I took my better soldering iron in use because the other one blew up - or burned down to be exact. A few pictures earlier you could see it hanging down from the clamp. I forgot it on like that and the handle melted down. Luckily the inside wiring was soldered and melted open, not burning up our home. Thus I can continue the story.
So, the system is still not able to run the propeller fast enough. This time I have a clue why not. The speed input is the servo signal and its small variations cause the driver to vary the prop speed faster than it can follow. And it drops out of sync. I need to add a speed derivate limiter to the input signal. Well, that's only software :)
For a change I did some woodwork again. Paneled the right wing bottom surface to make some progress to the project. Easier than I had thought, took only 15 minutes. Of course I forgot to use my gas mask while gluing it in the basement. The next three days went with a running nose - oh when do I learn ?
Happy New Year 2006 !
Spent my Christmas time relaxing. Unfortunately the calendar was not so favorable this year, leaving practically no extra holidays. I spent some time analyzing and optimizing the software with not so much luck. Bought a new camera. Hopefully the pictures will appear better from now on.
After performing some very thorough software runtime analysis it became clear that my problem was not in too fast speed control variations neither in the lack of computing power. I reorganized some functions and optimized critical paths by assembler patches but the behavior remained the same. Had to borrow the better oscilloscope again to find out the cause. So stupid. The current measurement resistor - made of a middle layer copper area - was not even close to its tolerance. The current measurement was 3 times too high and thus the overcurrent limit turned the system off at 20 Amps instead of the designed 60 A. It also turned out that the linear VHz compensation I had used was not suitable for a propeller drive. At about 180 Hz the system started to oscillate, finally pulling out of sync that can be seen on the left as a sudden increase in the overall current consumption (red). Back to the drawing board. I have to educate myself by some knowledge about propeller drives.
While wrestling with the software algorithms we spent a long weekend at the country house. Had a look on my flap designs and decided to do the whole thing again. The hinges were far too weak, not lined up properly and the overall construction seemed not to be rigid enough. At first I spent a day installing a new lighting system to the basement to make it easier to work down there. Then I designed a hinge guide shown on the right that I was able to line up the hinges properly with the aid of a 1 mm drill bit.
Next I started to add more support to the hinge points as shown on the left. While the flaps will carry enormous weights during landing I decided to support the hinge points with diagonal rods to the main beams to prevent the wing from twisting.
Another weekend brought the project a bit forward. Managed to complete the flap of the right wing. The new sturdy hinges work just fine.
(Most of my time went trying to figure out how to make the flap up picture to change to a flap down one either by bringing the mouse on the picture or just by itself. Could not and so they are there both side by side on the left. If anyone reading this can give me some assistance please do so. I am building this page up with Microsoft FrontPage.... )
A month later I figured it out by myself. Look at the motion ! With the aid of Windows Movie Maker I managed to turn the subsequent pictures to a movie.
End of January 2006. I was trying to fit the wing tip float supports on. Too late ! I had not thought about those when I glued the right wing bottom balsa cover on. The right support did not fit into its place any more. Oh face ! Whatever it was to become I decided to finalize the supports during the next week at home and try them on at next weekend when back at the country house.
On the right the construction is shown more in detail. On the left both sides are complete, waiting for final outline to be shaped. As it can be seen at right, the float support bars are fitted in square shaped holes, lined in with carbon fiber surfaces for improved durability.
I was lucky ! By filing a millimeter off from one of the wooden main beams I managed to squeeze the support on its place. Lots of epoxy and they were fixed in place.
February 4, 2006. I finalized the construction of the ailerons. While I had not completed the design the time I cut the arcs out of the balsa wood I had to do it the hard way. First I added curved extensions to the aileron arcs after which I cut out the rear ends of the wing arcs to make room for the ailerons to move in. Basic construction shown not so clearly on the left and the final result shows on the right. Only the part of the aileron inserting the wing is covered, the rest will be done the next time.
Note the two square holes for the wing tip float supports.
And the fact that I managed to reutilize my newly gained ability to make movies ! Now ( a month later) also the aileron is moving.
A break for a change. Thanks to my friend Cosme Salas from Spain I got a glimpse how it might look like when I make the first attempt to fly the Goose. Hopefully not... To see the catastrophic trial of a full size Goose takeoff, please move your mouse on the frame at left.
I managed to spend some more time with the project. As a result of a couple of days continuous effort I almost completed the right wing construction. Epoxy hardening took the major part of it.
The flap and aileron mechanisms are complete. Servos are in place and operating. Servo wells have their cover lids fitted on. The wing tip float support bars are in place but are of course detachable for easier transport.
What is still missing?
The aileron and flap ends have to be finalized with a thin veneer coating and the wing tip must be cut out from thick balsa block and fitted in place. With navigation lights of course. The float bars must be supported with four wires and their fixing positions have to be prepared to withstand the forces when the float hits the water. And of course the engine nacelles have to be prepared.
The wing top skin cannot be fitted on before those. Actually I will not fit it on before the left wing has been brought up to the same point. Easier to copy.
Now, in early March 2006 it came clear to me that the Goose will fly next summer. However, before that I had to learn myself how to.
Following the advice of my tutors I acquired a simple plane to start with, an Estarter. The assembly that was supposed to take a couple of hours took me two evenings - plus one night. Not so bad, after all. That time included getting familiar with RC control systems and how to program them for the specific application. The last night went to build skis under the plane. Once I got to my country house it came clear to me that the ice surface was too rough for wheeled takeoff but required skis instead. 3 am they were ready ;-\
In the morning I went to the ice (at left) for my first ever trial to fly an RC plane. Turned it on and rechecked the functions - good for me. The ailerons were working the wrong way round. I had ruined all three servos of the plane by moving them by hand when building it. Their gear cracked even so careful I was. Purchased new servos and replaced them not noticing that there is no standardized way they operate. This one new servo was just running the other way round. Well, it was quite easy to reverse its operation at the transmitter once I had figured it out...
Boy was it difficult ! The first flight attempt lasted two seconds. The next one took about three seconds. I was so lucky to try it in this nice calm weather with 30 cm (12") of soft snow on the ice. One wing support bar broke off at second crash but no real harm done. Glued it back and continued after one hour. Trimmed the flaps more to left because the plane seemed to twist right. This trial was much better, lasting well over one minute ! Unfortunately my coordinates got mixed while the plane was about 300m (1000 feet) out on the lake at 10 m height. It dived in the snow again with nothing broken but I had to work my way through the snow for it. Next trial was a partial success - after a two minute flight I managed to land it to my snow-ploughed skating rink. Unfortunately quite roughly, though. Another wing support broke off. Glued it back and recharged the battery from my car.
After another hour I decided to try it again. It got out of my control, flying towards the woods and I had to cut the power. Nothing happened ! I had to guide it down for another crash-landing on the thick snow. It kept on digging itself to the snow ! Switched the transmitter off and on. Tried all the tricks I knew. Finally I ran to it and pulled the plug. Silence, at last.
It turned out that the motor speed controller had let its smoke out that made it inoperative. As we all know, the electronics operate by smoke, once let out the operation stops. The reason had to be overheating and thus I had to position it for better cooling the next time. Further investigation showed that it had also failed to manage the receiver operating voltage at 5 V but delivered 9.6 V instead ! I have to do my own power regulator system even to this small practice plane. Who can you trust nowadays !
Next week I visited the hobby shop. To my surprise I was given a new, more rugged motor driver as a warranty replacement. The receiver was luckily also intact and thus I could continue at no cost. On top of all this they volunteered to test fly and trim my plane. That's what we did after their working hours. The first time I could try to fly my plane relaxed. It was first elevated high enough to cope with my steering errors and trimmed to fly straight before the controls were given over to me. It was really fun ! After getting some touch to its behavior I gave the controls back and asked him to land it for me. Watched close to be able to do it by myself the next time.
Tried it alone a few days later. It lasted for about five seconds after which the plane had circled around, landing not exactly on me but on an icy snow pile next to me instead. One propeller blade broke, the skis cracked and the wing broke off with large styrofoam parts breaking off from it. Oh my ! Collected the remaining items to my car and drove to a nearby hardware store for more epoxy glue...
March 26th. Got some good advice from Cosme again. Purchased an USB interface cable to my RC-controller and started to practice flying a simulator. After a crash no epoxy was required, just start from beginning again with a push of a button. After some time it felt boring. Needed the real action.
After a couple of hours of simulator training I was confident enough to try the real plane again. Just as with the simulator the takeoff was perfect (including rudder compensation for full throttle) and flying around showed problems only when flying towards me and inverted controls were required. Moved the controls wrong way round and the plane flew direct to ashore and heading for trees. Cut the power in panic. Still it managed to hit the trees, banging down in pieces. The whole tail was broken off. Glued it back for another try.
Did some more simulator training but the landings were still troublesome. Tried it again in real life. Otherwise fine but when it was the time to land the plane the problems started. It had never succeeded with the simulator. Neither did it succeed in real life. The plane crashed down with power still on. Cut the power and started to walk to the plane in snow. The walking turned into a run when I saw the flames ! The motor controller had let the smoke out again but this time it also caught in fire burning the styrofoam structures around as shown on the left. A larger renovation work ahead !
Well, I went back to the Goose construction again. On the right the wing tip float support wire fixing points are done. The four support wires are snapped onto the carbon fiber flanges that were glued to the rigid wing structures.
Some more constructional design was required for the engine nacelles. On the left I went back to the drawing board. At first I was in the belief that I could finish the design in one evening. No way ! I had detailed drawings for the rear portion only and the rest had to be estimated from the quite small 3D drawings I had. Plus the construction for motor assembly was to be thought rigid enough to withstand all the forces, including a crash down.
Some weeks later and much more experienced I finally started to have the nacelles designed. They turned out to be really difficult because of various versions of the Goose that I had drawings for plus the altered wing profile. Finally I think I found the design that fits my version.
The drawing is too big to fit here conveniently and therefore I have to ask you to click the enclosed link nacelle. to see the profile. Next week I spent an evening cutting out the carbon fiber arcs and the next weekend I made the rest of the arcs out of balsa.
Well, the spring takes its time, gardening and all. A bit at a time I have been gathering the nacelles together, not succeeding that well after all. End of May I had to give in that my Goose will most probably not fly this summer.
On the left the rear arcs of the right nacelle are already glued in place but the front end is still seeking for its position.
Unfortunately I got completely engaged by my Boat House project for a while. Once I have it up and warmed I will move the Goose there to be continued.
It did not go exactly that way. Waiting for the sixth carpenter to begin with the Boat House I had a couple weekends to spare. Thus I came back to the Goose for a while. It was almost a year ago that I left the project waiting for a better workshop. Now I took the wing back to town and continued where I had it started - at the dining room table.
I replaced the nacelle balsa arcs by veneer ones for better durability and fine tuned them all to fit in place. At right the nacelle frame front part is basically in place. Next it needs to be positioned exactly and glued up.
To be continued as I progress.