Quick Silver (The Electric Bat)

better suggestions for a name (still) considered



Notes on the video: It was taken off VHS hence the quality. No one expected the model to fly (I showed them) hence the lack of take off shots. The ending is rather dramatic but on closer examination of the raw footage the model was on an even keel before leaving the shot. And in reality it did just keep flying. Boing!!

The project began after considering building a really small machine much like Matt Keennon's Microbat and deciding a better bet might be to use an existing rubber design as a basis for a larger model to start with. It seems great minds think a like as you will see in this thread here.

After seeing that I sat down at the PC, made some design decisions and drew a few parts in Rhino to be cut out on my little CNC. That sounded straight forward didn't it. OK it took longer than an afternoon. With flight tests only once a month and some missed flight sessions the development speed has been a little on the geological side.

The basic spec:

Wingspan: 12"
Weight: 17g
Motor: KPOO, or M20LV a 10mm diameter hot brushed motor.
Radio: JMP actuator combo.
Battery: 1 X 145mAh LiPoly cell.
Gear ratio: 33:1 using Didel gears in a custom gearbox. Bushes are Didels and shafts are hardened drill blanks.
"Crank ratio": 4:1 as per the FreeBird this model is based on. The crank geometry is a little different.

The picture above shows the overview of the bird (or bat). The wings are aluminumised mylar (from Sams models) and the leading edges are carbon fibre. They were made by splitting a 2mm carbon rod down the middle with a sharp knife like bamboo. The two halves were sanded and tapered to the same weight. The flat of the now half round rod is an excellent gluing surface. The glue used was 3M77 spray adhesive, it sticks like you know what. Occasionally tiny squares of DS tape were used in the field if edges began to peel. The membrane stiffeners are balsa and seem to help although I would like to test that out further. The pivots for the wings are simply brass tube and the carbon rods have piano wire pivots bound to them with Kevar thread (from FreeFlight Supplies). The bumper is made from 3mm wide carbon capping strip (also from FF supplies) Three layers were glued together on a former to make the main vertical bumper, the horizontal one is from two layers added later (now removed). This one has been repaired twice and kevlar added to prevent delamination. The model really bounces with this on and I recommend something like this or similar on any Ornithopter design for use indoors.
The side view shows the more of the basic structure. The fuselage tube is carbon braid based tube from Westech. The tail is a 0.5mm carbon frame with 0.5mm Depron covering except for the rudder and stabilizer which is 2mm depron. The frame on the tail is again for protection in the event of crashes, 17g of carbon can fall very heavily indeed. It also protects the actuator nicely. The ball joint that allows the whole tail to be cranked up for trimming purposes can also be seen. This was made from polymorph which is a low temp polymer available from Maplin amongst other places. The ball was rolled in the fingers, allowed to cool and the socket formed over it using more soft polymorph.. The battery has a length of teflon tube taped to it and this slides onto a 1mm carbon rod, this allows CoG adjustment and also helps in crashes.

Here is an overview of the flapping mechanism. The first thing to notice is that it is scaled down in comparison to the FreeBird2, the crank is much shorter and everything is more compact. This may not have been a good idea in terms of bearing load but I like the look of the model. The motor is a push fit in a piece of custom made carbon tubing and this in turn is glued into two CF plates that each hold the brass bushes. The two Didel gears then form a 33:1 gear ratio to the crank output. The crank is a double crank as per the original FF rubber design. This was perhaps not the best way to go about things but at the time my knowledge of flappers was limited. What I did know was that I did not want the meshing of the gears to change because of the force applied to the crank. To prevent this I used drill blanks for the shafts, this barely bends and never breaks. However you can't bend a crank from it!! The solution suggested by Peter Frostick was to use a technique sometimes used by FF rubber guys to drive props. The crank was made with a spring at one end such that the force applied to the crank tightened the spring to grip the shaft. A small tab on the end of the spring would also engage with the gear. Unfortunately the gear is a little too thin for such engaging tabs and the spring effect not quite up to the task especially in the early days with a heavy throttle thumb.

To beef it up I borrowed an idea from the Didel prop hubs. These engage with the small gear in the centre of the spur to transfer the power. Mine is made from CF and details of its manufacture can be seen here. To make it even beefier the end of the spring goes through the crank driver and engages with the gear as well. The cranks are CNC cut 0.8mm thick CF sheet bushed with brass tube. All these parts are retained with cable outer which is also dabbed with 5min epoxy as these things know how to work things loose. The gearbox is removable and is a push fit on to the sub fuselage, this joint is also 5min epoxied but can be removed when required. The sub fusy hangs on a couple of 0.8mm CF stand offs. One other modification is the dropped wing pivots, the point at which the conrods connect to the wing spar has been lowered to avoid a nasty toggling action at the suggestion of Peter again. This required a change in conrod lengths and this was done by trial and error while trying to get a fair amount of dihedral in the flap.

More pics of the mech. You need to build these like a heli in terms of resilience etc. But that is no so easy at these sizes. Another modification made after the original build was to put 1mm carbon rods between the two plates that make up the gearbox around the locations of the bushes. This prevents the plates pushing together in a crash causing problems with bearing alignment. You can see the double crank quite well here, it allows you to adjust the flaps of each wing but to be honest I left it as I originally made it and used a bit of bluetac on the wing tip to stop the model turning sharply, a suggestion of John Mack.
Before trying more exotic control or going for a servo I decided to try a rudder. I knew Matt had success with is machine using rudder and elevator control so I unplugged the fixed V-tail (used in the partial FF trials) and plugged this one in. It uses an actuator kindly made for me by Peter Frostick that weighs just 1g and is meaty in the torque department. The "long" coil is also said to give more linear torque than some. What I wanted was some beefy rudder action and something that could be mounted in the tail but not on the surface. This fitted the bill. By mounting it in the tail like this the push rod run is completely unaffected by tail trim and of course the surface is also protected by the CF frame.


I had some encouraging flights right from the offset but they were only that. Initially the crank geometry meant I was loosing a lot of power due to the toggling action (which I countered with a stop) but the stability was nice. It was hard to keep in trim as well, as in the early stages it was throttle only and the tail was held by a length of ali tubing. The crank geometry was changed and there was a good improvement with some better flights, including climbing flight however I was dogged by mechanical failure. This was fixed with the crank driver and gluing of all parts involved in the gearbox. Later it was improved further with a beefed up bumper and the gearbox stiffening rods. The tail was replaced with a CF rod frame and 0.5mm depron covering and the ball joint added.

At this point I got some very encouraging flights. Still a few mechanical problems but mainly crash induced. There was one thing for it and that was control.

Once the ornithopter is basically trimmed by adding tip weights or crank tweaking or both the most important setting is the elevator. The hole tail was cranked up much further than seen in these pics but is was critical for good flights. Too much and the flight it stally and control is poor. Two little and the flight is very short indeed ending on the floor. Just right and the model moves purposely through the air. I'll add trim tabs to achieve this later or perhaps elevator. Anyway I tend to give the model about half throttle on the hand launch and give it a fair heave ho. I then throttle up or down depending on the elevator trim situation. When it was right it was a doddle to fly. Very controllable and capable of figures of 8 easily in the smallish hall it was flown in. It reached the ceiling on more than one occasion and I had the throttle back, not something I ever expected. The hardest part is orientation, it made me feel a little sick after a while. Oh that was 3min and 50seconds by the way.

What next? Well a 6mm pager motor and IR and there are plenty of other things to try as well as general improvements.