Back to Part 3

Each Way Bet - The Full Story
Part 4
by Trevor Hewson

 

Eventually, the weather calmed down a bit and there was nothing for it but to commit the unlikely contraption to air. The first test took place at our outdoor site, and was an attempted ROG in conventional flight mode. Here is the subsequent e-mail report:

 

The first flight lasted all of two seconds, the model rearing up off the ground uncontrollably, managing half a stall turn before sliding in on its wingtip. It did at least confirm that there was enough power.

 

After patching up with zap and masking tape, a second attempt was made with the cg forward a touch and full down trim. This time it got airborne, but still needed the stick held well forward. The model steered well but, since I had no more downtrim available, it was hard to control the airspeed. This was doubly awkward because the vee tails fluttered badly at anything other than minimum flying speed.

 

Eventually though, we got up to a good height so I closed the throttle and flipped the mode switch. This resulted in the most perfect harrier descent you have ever seen. I did try to persuade it into canard flight but can't now be sure whether I opened the throttle or just pushed forwards on the elevator. Anyway the slow, level descent continued, seemingly regardless of control input - I guess all flying surfaces were completely stalled. At about 20ft, I flipped the mode switch back and, more by luck than judgement, the model flared for a perfect landing.

 

The plan now is to make a proper job of the repairs, re-work the hinging of the vee tails, and adjust the elevator trim. I suspect that there is still not enough cg shift so I may try to make the wing move a fraction further, but there really is not much more that can be done here without major re-work.

So, the concept is not yet proven. However, the model was good value today, if only for the reactions of people when I sat it on the ground and flipped the mode switch!

 

So far other people's reactions to this model, whilst amusing, were begining to make me think of re-naming it from "Project X" to "Project Why?". However, I had a good list of possible names to choose from - Flip-flop, Shuttle, Recoil, Boomerang, 2-way stretch, Bipole, Return Flight, Loose Change and many others came to mind. It would be some weeks yet before I settled on Each Way Bet, partly because it seemed such a gamble every time the model took to the air.

 

If the results of the first flight test were less than totally successful, worse was to come. On the second outing, I opted to try a flight in Canard mode:

As Clive witnessed, an attempt at flight in canard mode today was very short lived. It all seemed to happen in a blink of an eye but, as far as I could see, the model set off from the hand launch straight and level, albeit only for a fraction of a second, suggesting that the cg was somewhere near right. However it then seemed to flick round horizontally (i.e. around a vertical axis). The model was now trying to go forward (in the conventional sense), but with the wrong geometry (i.e. cg on the trailing edge) and with full reverse thrust! Not surprisingly, the result was a flick to inverted followed by a crunching return to earth.

 

The model looks very sorry for itself but, being Depron, is actually quite repairable. The question though is what to do about the yaw instability. I had always feared that the vertical component of the V- tail would destabilise the model in canard fight but I must admit that the shear violence of the turnaround did come as a surprise - it really was reminiscent of launching a dart flight-first. I could try flattening the V, in the hope that there is an amount of vertical area which the canard can cope with but which will still provide a working fin/rudder function in conventional flight. Frankly, I am not convinced that such a compromise position actually exists.

 

The only alternative which I can think of at the moment is to rotate the whole V tail assembly when changing modes so that it becomes a swept back canard rather than a V canard. Apart from the formidable mechanical challenge of actually achieving this, it does also pose the question of how to steer in canard mode (the model has no ailerons, remember). However, With independent control over the two canard halves, they might work okay as tailerons - or should that be canarderons?

 

The damage to the model was a bit of a setback, but the real problem was the manner of the crash. I was fairly confident that the problem wasn't simply a cg issue but was more to do with yaw stability. More e-mail correspondence followed:

The problem is that, whilst the moveable wings are a large proportion of the horizontal surface area, the only portion of the vertical area which moves with them is the interplane struts, so the 'vertical aerodynamic centre' doesn't shift much. Hence the challenge of achieving positive yaw stability for both flight modes.

 

It seems to me that the possible remedies are:

i) Increase the moveable vertical area. e.g. replace the inboard carbon rod struts with sheet depron.
ii) Reduce the V angle when changing to canard mode.
iii) Take the Eurofighter route and resort to electronic aids, e.g. Gyros, to stabilise an inherently unstable aircraft.
iv) (I have just thought of this one!) Render part of the vertical surface ineffective when changing modes. e.g. replace the V tail by a conventional tail and let the all-moving rudder swing freely in canard mode, perhaps by slackening a closed loop control system?!

 

So much scope for mental exercise. I think I might make up some cardboard shapes representing the side view, and see just where they balance. Maybe more chuckies too, but the difficulty there is that the fuselage accounts for a significant part of the vertical area, so a chuckie with a stick fus really doesn't help much.

 

At this point the project was in real danger of being put on the back burner once again. However, as the new Winter flying season loomed, I forced myself to confront the problem and chose to go for remedies i) and ii) from the above list. Repairing and re-working the model was a tedious business but eventually it all came together again.

 

In these pictures, you can see that the all moving tail / canard surfaces are now attached to a small root section which is in turn hinged to the fuselage. A bellcrank underneath, operated from the moveable nose/tailwheel causes the 'V' to be flattened when switching to canard mode.

 

I duly took the model along to the first meeting at Calshot and Nigel Potter kindly agreed to try to capture any airborne moments on video. As Before, I ran the model back and forth on the ground to show off the mode change, before taking a deep breath and opening the throttle. You can see the results by clicking on the 'Video Clips' link below. To avoid too much frustration for readers with dial-up connections, here is my subsequent postmortem, again in an email to Mike Roach:

The model did get into the air, albeit briefly, in both flight modes. The really incredible thing though is that I brought it home in one piece! The conventional mode flight was okay for 3/4 of a circuit then it seemed to do a quick dutch-roll type of wallow, tip stalled and levelled out just an inch or so below ground level. Damage was limited to a loosened wingtip panel.

 

The canard flight was somewhat briefer. A clean rise off ground into a steep climb, levelled off, then an indescribable manouver which resulted in a vertical landing from what Nigel Potter described as a "prop push" - the canard equivalent of a prop hang. Tentative diagnosis was that the pivot/torque rods (2mm carbon rod) driving the canards were too flexible allowing the control surfaces to deflect uncontrollably as soon as the model built up speed. I have had a go at improving this, so we shall see if it is any different next time.

 

At the time of writing (Late November 2007) the model is ready to go again. Watch this space!