Robustness of snakes

[Erfolg]

Ben

 

Although I have not attempted to measure the clearance between the inner OD and outer ID.

Hazarding a guess, from sliding back and forwards, waggling etc. I would estimate the clearance of in the region of 0.005" (I still estimate fits in inches, it is an age thing, lengths and distances metric, strange but works for me).

 

This does not result in play in length by that amount, as the change in length, is determined by the radius of the circle formed by the inner bending , as in crippling, until the side wall is contacted in the middle and the two points of the side wall at the ends. The radius of the circle is dependent on the snake outer length, none the less the radius will be very large as the arc length is a small part of the circumference. The change in length being the difference (or delta as scientists like to use) between the chord of the segment and the length of the arc (the length of the outer casing). It is times like this that I wished my scanner worked, as a sketch would show easily the geometry, where words are clumsy.,

 

Bob

 

I have also considered the weight issue. It became apparent to me when loading my rig with weights that 1.8 kg is a lot of weight and that 3kg was a significant weight. My comparison was in my mind, could any of my control surfaces withstand that weight even is distributed over the whole of the area. I was not at all convinced that they could. I had even less confidenced with the horn. So what am I doing, I have fastened a horn to a Small piece of balsa, with a frame to it, so I can stick it to a board, When dry I will be hanging some weights from it to see at what point it fails.

 

I also thought of the servo and concluded that the quoted torque is with the servo stalled and not at all indicative of the torque normally generated to move the surface whilst flying etc.

 

I seem to remember the BEB has done some calcs on this, perhaps a pointer to the thread will illuminate both the issues and the calcs he did.

[Erfolg]

With respect to push rods and hard woods. I am not sure that is the answer initself.

 

Probably the worst push rod system I ever owned was in a Elly Superfly, which I bought second hand. It came with a probably 1/4 dia dowel to the elevator and probably the same to the rudder. I cannot remeber well.

 

In operation the push rods could be seen to deflect all over the place. As I was learning, it was the least of my concerns, as the model flew like a brick. Perhaps not the best model to learn with, or may be that particular one.

 

According to memory, the most important aspect of a column was the " second moment of area", to obtain stiffness. Hence many columns seen, are hollow tubes etc. Woods are/were considered poor from the view of there consistent of properties, hence there was a significant "K" factor, a corrective constant, to take account of the non isotropic (if I remember the term correctly) of wood.

[Bob Cotsford]

The Wot4 XL is the first pushrod I've used in 40 years that I've been happy with - as you say, the traditional balsa pushrod with a bike spoke tied on the end is a total disaster on anything of any size, especially with the recommended z bend that was often specified to ease it's route round the tailplane or fin. This one is an easy 13mm, maybe 15mm or more - I'm not dragging it back out to measure it - and the 3mm rods are short and straight.

In Tom's example I'm quite sure the plastic outer is stretching, it's being subjected to much more force than is reasonable.

Anyway, I'll be sticking with my Sullivan plastic snakes for the moment, they're just so easy to install.

[Tom Wright 2]

Posted by Erfolg on 28/05/2011 15:34:44:

 

In operation the push rods could be seen to deflect all over the place. As I was learning, it was the least of my concerns, as the model flew like a brick. Perhaps not the best model to learn with, or may be that particular one.

 

Any applied force transmitted via a rotating arm will produce a deflection in any system ,not just dowel rods.

[Erfolg]

I just tried the horn with a weight.

 

Bit of a disaster, as the horn broke the mounting plate when i hing a +1kg billet on it. It did not last long, failing catastrophically.

 

I have an old large rudder, lying loose in the workshop. I need to think of a way to use this, to produce a result, that is, specificl, but not necessarily unrepresentative of a typical glider surface.

 

To be representative, I now know I need a spring balance, to load the surface assembly progressively. The use of weights would only work if i had a set of weights, so that I cold load the bracket with small increments of weight. The other big issue is distributing the load uniformally over the complete control surface.

 

I guess that in practice the stress distribution will be very similar to the contour lines on a map, concentric around the horn fixing. On this basis the horn attachment could easily be a weak point, if not carefully designed and made.

 

Does anyone know how to locate BEB's estimated calc on control surface forces. I suspect that 1 kg could easily exceed the load actually seen. Could the stalled servo be more of a hypothetical risk than a fact, on small models at least, or even on F3b types, as the bash about at +100mph, whilst pulling tight turns, I suspect most with no better than 2.1kg rated servos.

I am convinced more than ever that correctly installed, there is no issue with snakes. I now suspect that long push rods are inherently poor in performance, than any other method of moving a control surface.

[Shaunie]

Just to add my tuppence worth.

First off I am not a lover of snakes, preferring to use a closed loop setup if possible. Local servos with short pushrods must by definition give the best fidelity of control surface movement for servo operation but I do think them unsightly. Also sticking three servos in a tailplane is poor for CG and longitudinal moment of inertia reasons IMHO.

However, back to snakes. By way of visualisation:- Bowden cables used for throttles,clutches etc.work only under tension, however similar systems used for steering on boat outboards and gear mechanisms in many vehicles are similar to snakes, working in push-pull. In many instances there is no rigid link (the fuselage in our application) between the two ends of the outer, thus the load on the outer is the direct inverse of that on the inner as result the inner compresses and the outer must stretch, to combat this the core and outer are usually metallic. Transfer this image to a model:-

When a snake is loaded with the inner in compression the outer therefore will try to stretch, as its ends cannot move it will achieve this by bowing in the middle. Imagine a "perfect" snake with no sideplay and zero coefficient of elasticity it would be unnecessary to support it midway and there would be no loss of movement. In the real world however there is sideplay, this results in buckling of the inner under compression, a sideforce is generated on the outer which then gives the initial sideforce to start the bowing. This bowing then is caused by a net difference in movement between the input signal and the output one. It seems to me that the more constrained the outer is to remain in a straight line the smaller this effect will be, effectively the outer is only strengthened by the fuselage if it is constrained so as not to be able to move at all.

The big question is whether in the real world this is big enough to be a problem. It seems to me, keep your snakes straight and support them as much as possible.

One downside of plastic snakes I found on one model was a huge variation in trim due to expansion and contraction brought about by temperature changes. Before every flight I had to eyeball the rudder and elevator neutrals because I could never be sure where the trim would be.

Shaun

[Erfolg]

Shaunie

 

It is some time ago when I did my test, so i do not remember accurately what I did and what i found.

 

As I remember the experiment, the forces as end loads was far higher than my models could reasonably expect to see. If they were to see the forces loaded, I would have expected the servos to have stalled. In reality the weights appliedI were so high, I was fearful, if they were to fall of onto my foot, it would have hurt considerably. I suspect that the control surfaces would not have been capable of withstanding the forces.

 

It is worth recognizsng that I was considering models up to about 50". Not giant scale.

 

With respect to the snakes used, which were Sullivan Golden Rod, the strain was believed to be low, the calc would be easy.

 

strain = f/a, stress = delta (l)/l and E = stress/strain

 

combining

 

delta l = E*l*f/a

 

Rather than bothering with all the theory, measurements are more important. Taking into account all the slop (which is very little with Sullivan snakes) and strain, bending of fixed stop etc, I seem to remember that the change in length was about 0.005".

 

To my standards nothing, my servos probably move that much, against their screws.

 

I was testing a hypothesis put forward by I think Steve?, if the outer is firmly attached at both ends, overhang of the inner is minimised (to reduce/prevent) crippling, very little motion will be lost and fastening at intermediate points is not necessary.

 

I think in principle he was correct.

 

That it is with the provisos. The ends must be very firmly attached at both ends. The mention of the overhangs is also important, these need to be minimised, as these are a real source of lost motion.

 

Although I accept it is not strictly necessary to fix at intermediate points, I often still do. It helps keep things in order, they might reduce the forces that the ends see.

 

I consider Steves contention proved, that ends must be firmly attached. That no amount of fastening around the centre area will compensate for excessive unfastened outers at the ends