Hey everyone!

I'm going to attempt to build an ornithopter. I've already been through a couple of websites. But, like it is with hovercrafts, there is no place that tells you exactly how to build an ornithopter step-by-step.

Here are the points/ questions I've decided to consider before building:

Wing Design:
1)   How flexible should the leading edge spar be?
2)   Triangular vs. elliptical wings.
3)   Is there a need to set an angle of attack (and how much) or does the flexing of flapping wings automatically handle it?
4)   How to set the up and down flapping angles?

Gearbox and power system design:
1)   Low kv BLDC with low gearing ratio vs. high kv BLDC with high gearing ratio.
2)   Clustered spur gears vs. worm-wheel gears.
3)   Circular to linear motion mechanism.

Please add any important points I've missed and share your thoughts about the above.

I request everyone to refrain from pointing to websites as I've already been through a lot of them. I need answers from personal experience and knowledge.

From the various sources I’ve been through, the following seem to be the optimal parameters/ mechanisms for a normal sized orni:

1)   Wingspan of 36 to 42 inches.
2)   Flapping frequency of 4 to 5 Hz (flaps per second).
3)   Weight between 350 to 450 grams.
4)   Up-flap angle of 30 degrees and down-flap angle of 25 degrees.
5)   Outboard wing hinge flapping mechanism.

Hmm…no orni enthusiasts. Not unexpected.

Anyway, as there is so less straightforward info available here, I'm going to post basic concepts and a detailed build log so that someone new to ornithopters won’t have to go through scores of papers on bird flight.


1] Ornithopter flight mechanism:
I used to wonder how the simple up-down motion of flapping wings could produce lift and thrust. Shouldn’t the two motions cancel-out each other’s effects?
Then I read (and re-read multiple times) Robyn Lynn Harmon’s paper on flapping wing motion tracking experiments.
Here’s an excerpt:

For steady level flight conditions the ornithopters flap their wings three to six times per second. There are two spars, one at the leading edge and another placed diagonally from the leading edge to the rear of the fuselage. This spar arrangement creates two regions in the wing, the triangular “luff” region, which is a loose membrane, and the “flap” region which is kept taught by a series of fingers that run from the diagonal spar to the trailing edge. The flexible skeleton-membrane structure allows for highly dynamic passive shape change as the wing moves through the air.
The large degree of bending in the wing is a result of the membrane adjusting its camber and pitch to maintain tension equilibrium throughout its surface. At the beginning of down-stroke and upstroke the inertial acceleration of the wing causes the leading edge spar to bend significantly. This results in a variation of the local stroke angle along the span and therefore a phase-lag between the wing root and wing tip during the stroke period.
Additionally, since the flap region is essentially hinged about the diagonal spar, it experiences a large deflection. A consequence of the flap deflection is that the flap’s force loading exerts a moment on the wing that increases the pitch into the flapping motion, so if the wing is in down-stroke, it will have downward or negative pitch. This pitch adjustment is important to maintain a relative angle of attack with minimal stall, whereas an untwisted rigid wing would experience accelerated flow separation due to the large inflow angles.


The wings have a triangular support structure. A main spar runs along the leading edge of the wing and a strut connects from the rear of the ornithopter's body to a point near the tip of the main spar. From this strut there are several smaller carbon rods that project to the edge of the wing which are somewhat free to move. This results in a fanning motion from the trailing edge of the wing that produces thrust while the leading edge is flapping up and down which directly contributes to a part of the lift in addition to the conventional lift coming from airflow over the wing.

Morals of the story:
1) In order to produce lift and thrust, the leading edge spar of the wing must be stiff (or rather have very little flexibility) while the rest of the wing should be relatively more flexible. 
2) The up-flap angle at wing-root should be a couple of degrees greater than the down flap angle.

Umm, i am by no means even entitled to speak on this topic but i did watch a TED talk a while back where they showcased an ornithopter built by the company Festo called the SmartBird. In the talk they demonstrated its working and the approach was a bit different than what you described above. The wing was divided into two sections by a joint, and during the downward motion the wing would be at the maximum extension thus exposing more surface area and generating lift. But during the upward motion the wing would fold in such a way that the surface area would lower and hence the lift was more and the bird could fly. I think video would explain it a lot better than i can.

I know you asked not to post any links to websites and i apologise in advance if you have already seen this.

http://www.ted.com/talks/a_robot_that_flies_like_a_bird

Quote from: ayu135 on April 06, 2014, 10:59:39 PM
... called the SmartBird. In the talk they demonstrated its working and the approach was a bit different than what you described above...

Thanks for joining in man!

I did see that video a few days after it was released. I have a habit of downloading TED talks. Got a huge collection now.

The mechanism you described is possible due to use of articulated wings. Most beginner ornithopters do not use articulated wings due to their inherent mechanical complexity. Not to mention, those guys have amazing resources (machines and materials) at their disposal. I haven't completed my education yet, so I'm gonna have to stick with whatever build material I can find in my recycle bin. 

Quote from: ayu135 on April 06, 2014, 10:59:39 PM
...
I know you asked not to post any links...

Well, now you know why!    

Just kidding! Don't go thinking that I'm conceited.

Swapnil sir, am keenly following this thread of yours, as am also planning to build a micro ornithopter for a while now, if possible gimme your email, I'll share the data that i have regarding them

regards
Bilal

Thank you! Micro ornithopters have a totally different set of parameters and mechanisms. I do have some material on them.

And don't call me sir, man! I'm your age (maybe even younger)!

2] Gearbox:
The gear box is needed to get the 4-5 flaps per second from high kv BLDC motors. The following equation is used to calculate the gearing ratio.
Max. flapping frequency = (kv of BLDC X LiPo voltage)/ (gearing ratio x 60)

For a 1800 kv BLDC with a 2s LiPo and 5 Hz frequency,
Gearing ratio = (1200 x 8 ) / ( 5 x 60) = 32

For such high gearing ratios more than one stage of gearing is required. Then one has to choose between a few stages of large clustered-gears and many stages of small clustered-gears.
The worm gears seem to be an obvious alternative as they provide similar gear reduction in a single stage. But, their frictional losses are extremely high and should be avoided. However, some people seem to have successfully used worm gears.

3] Flapping Mechanism:

The design of a gearbox is nowhere near as complex as that of the flapping mechanism. There are loads of options. Some are simple but inefficient while others are efficient but really complex and heavy. Here's a discussion thread on the various flapping mechanisms generally used.

http://www.rcgroups.com/forums/showthread.php?t=1842514

I decided to start with the gearbox.

I realized it would be easier to use the gearbox of the geared motors used in robotics.

Below is the picture of the first gearbox I tried (resting on my first sweepstake  ).

The problem with it is that it came with a 100 RPM motor and has a very high gearing ratio (50+) and that I don't currently have a high kv BLDC. It is , however, a very simple and lightweight option. If anyone wants to try this method, go with a 1800+ kv BLDC and the gearbox from a 300 RPM geared motor.

Initially I thought about designing and fabricating acrylic gears. It's really easy considering that there's an awesome gears plugin available for sketchup.

However, I happened to visit my favourite electronics hobby shop (Sharma Electronics) and to my surprise they had 3 boxes full of gears. I had to go through a lot of them to find the perfect ones.
Finally I came home with those shown in the pics below. The big black ones fit perfectly on the 6mm shaft of a geared motor. The smaller black one (in the 2nd pic, clustered with the bigger one) had a bit smaller shaft-hole and needed some modification.
I got the pinion gear from the final stage of a 24V 100RPM geared motor. Removing the pinion from the shaft was a bit tricky. I had to take care not to harm the gear teeth while using the hammer. The 3rd and 4th pics show the axle and pinion. 

Fitting the pinion on the BLDC was a bit complex as they had a diameter difference of 2 mm.

I had to first superglue a shaft-adapter ring (from the set that comes with an electric prop) on the BLDC shaft. Then, with the BLDC running at full throttle, I held a sand-paper against the plastic ring and trimmed it till it was 6 mm in diameter.

I also used the plastic rings as spacers as can be seen in the pics below.

I designed a rough CAD model of the gearbox before trying to implement it. Then I printed the design and cut the parts out of a copper clad PCB. I used plastic 'stand-offs' to hold everything together.

The PCB gearbox helped a lot in determining the perfect distances between the gear shafts and gearbox plates (I did do the measurements and the math, but it's the feel that really tells you).

However, this gearbox was flimsy and I needed accurately spaced shaft holes. This meant fabricating acrylic parts with a laser cutter.

After refining the 1st design and fabricating the acrylic parts (3rd pic) I realized the pinion and 1st gear were a bit too close (0.5mm). The 4th pic shows the corrected design.

Hahaha! This thread is turning out to be a lengthy n lonesome monologue.   

I'm elaborating everything as I'm not using off-the-shelf or ready-made parts. Hope someone finds this useful someday.

@swapnil
Great job. Keep it up.
Took a while to respond, as i got interested enough to go through the thread again from the beginning.
I am a great believer in jugaad, and i feel proud of you.
How much does the gearbox weigh, and what is the final ratio?

Just saw the acrylic pic.
Excellent 

Iyer sir, thank you for your kind and encouraging words! 

I haven't weighed the gearbox yet but it's a bit on the heavier side. I'll have to do some more 'jugaad' to trim its weight.

I've attached the gearbox specs. Thanks for reminding me, I almost forgot! The gearing ratio is a bit low, but that's the most I could get without risking greater size and weight. I'll be using a 2s LiPo at half throttle.

Swapnil, I am following this thread with a fair bit of interest. Of course, not knowing much about Ornithopters, I chose to lay low and say nuffin, but learn something.

Please do continue, it is fascinating!

Only comment I have is I don't trust that plastic ring superglued on! Please look for a better solution, depending on what you have available  

Regarding 'monologue', i was about to say 'wait till Sanjay sir sees this' !

Quote from: sanjayrai55 on April 11, 2014, 05:42:30 PM
...
Only comment I have is I don't trust that plastic ring superglued on! Please look for a better solution, depending on what you have available 

Sanjay sir, actually I've used 5min epoxy. I've also epoxy-glued the metal pinion to the shaft using some elaborate 'jugaad'.
And the first stage of gearing doesn't undergo really high stress. Now all I can do is hope it doesn't give up.

Like I mentioned before, selecting and designing the flapping mechanism is much more difficult than the gearbox. The Scotch Yoke mechanism seemed stable enough and I decided to try it. It turned out to be really challenging to design and put together.



Also check out this vid: http://www.ornithopter-pilot.com/images/N1FlapMechTest10ScotchYoke.mpeg

First test video:

www.youtube.com/watch?v=2ZKjw-8TRS8&feature=youtu.be

Swapnil bhai, thanx for all those lengthy elaborations, am learning a lot along they way.
besides, the scotch yolk mechanism is really tricky to design, the forces at TDC and BDC are exactly equal and opposite so they cancel each other leaving the mechanism with no force acting on it, this sometimes leads to a 'stuck' gearbox. To overcome this generally an offset of like 0.5 degree is included in the design, but "kehna asaan hai", so better watch for that.

one more thing, how are you gonna obtain differential flap angle with this mechanism??

regards
Bilal

why not use something like this??

Quote from: Bilal on April 11, 2014, 06:29:07 PM
... the forces at TDC and BDC are exactly equal and opposite so they cancel each other leaving the mechanism with no force acting on it, this sometimes leads to a 'stuck' gearbox. To overcome this generally an offset of like 0.5 degree is included in the design, but "kehna asaan hai", so better watch for that....

Could you please elaborate that for me? I am an electronics engineer, don't know jack about mechanical stuff.

I chose the scotch yoke as it seemed much more elegant than the others.