Robustness of snakes

[Erfolg]

I will have to do my test again, as my results are at odds to yours.

 

Unfortunately I have a tailplane on my board at this moment, so it will probably be next week until I will have some results.

 

I do notice that you have a lot more sticking out unsupported than i did. I found that although my own was deliberately bent, that moving the "S" about had no effect on the inner. It stayed where it was.

 

I will do a straight test as well.

 

I glued my tests on with thick superglue, so that I would not crimp the outer onto the inner

 

But where there is discrepancy, we need to understand why. It could be the difference between a system that works and one that does not.

[Tom Wright 2]

Erflog.

Super glued my outer to the board then added saddles to make things clear on the photo.

[Erfolg]

Tom

 

I have just set up my test rig, whilst weather stopped me working.

 

I have endeavored to limit any change in length due to load, therefore no horns etc.

 

My results are totally at odds with your own.

 

I will post pictures a little later, when I have taken some. Together with the details of what I have done.

 

The indications are that a snake is far better than a long push rod, using balsa, any deformation is also substantially lower. But more importantly end point movement is minimal.

[Ben B]

Interesting experiment Tom. Can you try repeating it with the middle of the outer clamped to see what effect it has? It would be interesting also to use an indelible pen to mark the inner at intervals to see whether the length that is changing is the length inside of the outer or the two unsupported ends. Keep up the good work though, might get out some snakes myself later and have a go!

[Ben B]

Of course in this situation (where one end of the inner cannot move) you could say that a little give is a good thing, otherwise the servo at the other end would burn out in no time . But I still would be interested on knowing where the additional inner length is disappearing to if there isn't any slop in the system!

[Erfolg]

I have finished writing, somewhat hurriedly,as I want to get back to building my Bachem. In some ways I have probably wasted some 3-4 hours building time.

• Quote

[Erfolg]

I will try again, but in two parts

 

I have finished writing, somewhat hurriedly,as I want to get back to building my Bachem. In some ways I have probably wasted some 3-4 hours building time.

Title:- End load test on Sullivan Snake

Objective :- To establish change in length of a snake held at two points only.

Equipment:- Standard Sullivan snake, board for mounting, two restraints, one disc for point loading, two lead ingots, misc of office equipment.

Set Up :- The snake was attached to substantial board using balsa saddles with thick syno. The snake was off set by 2” or 50mm so as to be as typical for many RC installations. The outer case was arranged to have the minimum curvature practical.

An end stop was attached to the end of the board to prevent movement when under a compressive load. The snake impinged directly onto this stop, without any restraint.

Both ends of the snake were cut at 90 degrees to the axis of the snake.

The outer ends of the casing were arranged to offer as much support as was practical to prevent deflection outside of the area which was unsupported by the case.

The axial load at the active end was introduced by a disc, arraged to provide a point load. The disc was free to move being guided only.

The test board was stood as near vertical as possible, ensuring that the arrangement would not fall over and that the weights introducing the end load, would not be affected significantly by friction.

Method

The vertical board was stood near to vertical with the inner snake, free within the outer case. The inner was allowed to contact the end stop. A light transient load was applied to ensure that the seating had been achieved.

The position of the free end of the snake was marked, onto the board.

The 400 g disc was introduced into the end housing and placed onto the snake. The intension was then to mark the new end position of the snake. A check was made to see if any deflection had occurred with the snake casing.

A lead ingot was then carefully balanced on the disc to increase the end load to 1600g. The end of the disc. Free end of snake was marked, a check made for deformation.

The above procedure was repeated with an end load of 2800g.

Measurements

Test 1, no movement of free end, no change in position of snake outer

Test 2, end movement 1/32, minimal movement of case(sideways deflection)

Test 3 end movement 3/32, 1/8 movement of case.

It was noted that the outer case did not initially return to the unloaded position. After a period of some 30 minutes, it became apparent that the original position was regained, without intervention.

[Erfolg]

Discussion

Although the mass’s used were not totally idealised, they are were selected with a view of representing particular category of operation.

The 400 gram mass was thought not to be untypical of the forces (multiply by 9.81 if Newtons are wanted) encountered on small models.

In the case of 1600 gram this was thought typical of small to medium models.

The mass of 2800 gram is an extreme value, where a snake of the type used would not normally be thought appropriate.

The experimental apparatus could be significantly improved to obtain results which could be considered far more definitive in value. Typically a lever and weight arrangement could be used to apply the axial load, making a truly convenient vertical arrangement viable. In this case the objective was to establish typical values of lost movement due to the snake components themselves.

There is also the apparent deficiency where there is restraint in two plane of the three planes, the movement not being totally free in the “x” and “y” plane, the “z” axis purposely restrained.

It became apparent that a photograph of the end positions would be useful to those interested in the performance of a snake assembly in compression.

It is not unreasonable to assume that a similar amount of lost movement occurs when a tensile load is applied.

Conclusions

With a load of 400g there was no movement of the inner, or noticeable movement of the case certainly at An end load of 1600, although some change in the end of the snake was apparent, at the level practical model usage the lost movement would acceptable to most if not all modellers.

However when a load or mass of 2800 grams is applied, there is a measurable movement of the snake end. Although purely subjective, a impression was gained from handling the weights that a mass of 2800 is quite a high force. Probably outside the intended duty of a snake assembly. Forces at this level are almost certainly going to reveal the limitations of other parts of a typical control system. Components such as control horns, clevis, servo output arm, would be severely tested.

With respect to an overall impression of the Sullivan Snake assembly, if installed carefully, there should be very little, if any loss of movement due to the characteristics of the assembly within its intended operating regime.

It is apparent that that acceptable results are obtained by firmly anchoring the extreme ends of the snake outer casing, additional intermediate fastenings if undertaken with care, should do no harm. There could be benefit in reducing both the stresses and forces seen by the end connection, thereby increasing reliability.

The results presented and writings have purposely been limited by both time devoted to the process and the lack of presentational media.

Recommendations

There is clear benefit to modellers in quantifying the capabilities and limitations of the individual components that comprise the control; surface control systems.

It is apparent that in addition to snakes, that long push rods are frequently utilised. On that basis it suggested that a similar test of a pushrod assembly, comprising the two bicycle spoke and balsa rod, is subjected to a similar test.

Pictures

Click onto the images to enlarge.

This is my test rig. Unfortunately you need to rotate your monitor clockwise by 90 degrees to view

The following are the weights used

Quote

[Erfolg]

The following are the weights used

Marking unloaded snake

Test number 1

Test 2

Test 3

Careful examination of the picture you should just see the additional mark

 

Image of the two marks, showing the change in length.

As these two images are so poor, I will see if I can get a sharper picture tommow.

 

[Peter Beeney]

From my experience at least, it seems that if any of the types of control systems are installed incorrectly they can be a source of trouble; and probably when when the load is greatest which means that the model is most likely to be in an unusual and invariably higher than normally stressed situation. Even closed loop wire can break down, I’ve seen a c/loop crimp fail on the elevator of a largish model, allowing the wire to unravel. Fortunately it only resulted in a heavier than normal landing. The crimp had not been sufficiently squeezed together, thus allowing the returned end to pull free. When we checked the other crimps they were similar, so we concluded that the constructor did not have the correct tools; and perhaps also not enough experience or knowledge to realise that it has to be done with at least some accuracy, otherwise sooner or later it will make it’s presence felt.

‘Installed incorrectly’ is probably only in the eye of the beholder, i.e. - me! I’m sure most people would be quite happy they’d got it about right; and it’s quite possibly it’s me that’s wrong, but I have seen the rare model go in due to lack of control which I think can be attributed to loss of control movement.

Erfolg, Whilst I would applaud your tenacity in undertaking those tests on the snakes and it certainly proves something, and I certainly don’t wish to take anything away from that but by using the Sullivan brand are you not guaranteeing that there will not be any problems? You also appear to be using very short exposed ends of the inner tube, too. Sullivan tubes may well be the best of the best. Modern snakes seem to have a very stiff walled outer plastic tube. You need to do the same tricks on some of the much older, more light weight varieties to perhaps get an overall picture, how about a Bowden wire type, with it’s very thin flexible outer, and with about four inches of wire exposed unsupported at each end? That’s how some installations are. Or, at least, have been in the past.
I’m quite happy to do a standard check on the model. Such a check would simply be operating the control surface by hand, and that has to at least turn the servo, switched off of course, without any sign of sponginess. If there is any ‘give’ in the system at all I would want to try and eliminate that because I would consider that it can translate into blow back.

Having said all the above, I’m sure snakes can be very reliable, I’d personally use them anywhere, and have done for years. I secure them along their length, but that’s just an individual preference. These days, though, the servo mounted close to to the control surface and a short pushrod seems as good as any. So maybe the issue about the quality of snakes is gradually going to diminish anyway.

PB

[Tom Wright 2]

Ben.

Yes will do the test with the middle clamped tomorrow.

 

Erflog

I think if the tests are going to mean much the load applied needs to be quantified , have you got a strain gauge ?or suitable measuring device.

 

All

A moderate load perhaps roughly equivalent to full rudder deflection on a largish model at speed produced a bend in push mode which equated to a significant reduction in rudder movement ,the out of tube 3mm rods did not deform in any way ,the bending in between the clamps does reduce the distance between the two exit rods by up to 1/8"hence the control surface movement is reduced by a factor that is a function of the movement and the length of the control horn.This rather obvious result is similar to the results expected from a thin push rod under push load conditions.

[Tom Wright 2]

Photo shows the the amount of bend that occurs when a firm push load is applied.

 

CLICK TWICE TO ENLARGE.

 

Edited By tom wright 2 on 27/05/2011 15:06:35

[Erfolg]

Peter

 

My objective was not to suggest that one system/method is better than the others,

 

The objective in this case was to establish what is fact and what is myth with respect to snakes.

 

As you have suggested it is imperative that as little as possible of the inner is left unsupported by the outer case and the outer case is firmly attached as near their extreme ends as possible.

 

Before reading some contributions to the thread, I had been of the opinion that you should support the outer at intermediate points to minimize excess slop in the system.

 

The argument presented with respect the outer is a fixed length, as is the inner, therefore they cannot get longer or shorter, other than through strain or changes of temperature and both would be expected to be relatively low, during operation.

 

I thought about this and the argument seemed compelling.

 

I later realised that I have a number large gliders where the ends of the snakes only is restrained. I then thought, none of these machines have caused me any problems, particularly on the elevator.

 

It was these thoughts which caused me to do a quick initial check and the later check when Tom could not replicate my results.

 

My 120" Open glider used the Bowden cable type. The outer is significantly stiffer than the Sullivan type. This system does have issues, that is of high friction if not maintained. Again I pinched an idea from Don Edberg/Lee Renauld from their Sagittas, in that it is possible to remove the cable for periodic maintenance. Again it is not slop that is the Achilles Heel, but friction.

 

Am I an advocate of Snakes, not really. I much prefer to have the servo close to the surface to be controlled, with a short push rod.

 

Although once a user of "long push-rods", I tend to avoid them now. The reason being that getting a push-rod anywhere as slop free as a snake is difficult. The last long push-rod I made and used, was constructed from a "Aluminum Arrow shaft" with the shortest bicycle spoke connectors that was possible. This was stiff, but not perfect by any means. Yet as far as I know, no one has tried to quantify "long Push-rod " performance.

 

For a period I used closed loop systems on my big gliders. They were very light, did not suffer from inertia, which can be a real issue with a competition glider, that is forced down into the landing circle. As the saying goes "Landing points, make trophies". Although I did use them increasingly until stopping modelling, I became aware of the limitations. You need the wires tight, if flutter of the control surface is to be avoided. Typically encountered when forcing the model down from height as the slot time comes to an end. The tension required seems to put the bellcrank/horns under great stress, commercial components were not up to it. The areas where the wires are bent sharply are also prone to wear/failure.

 

I do not think that there is a perfect solution, each system has its strengths and weakness.

 

I am convinced that most systems with excessive slop, has more to do with the builder than the system, or lack of maintenance.

 

Finally do you have to support a snake at intermediate points, I would say no. But intermediate attachments should do no harm in themselves, yet are no cure for unrestrained ends with excessive unsupported inner.

 

These tests have also left me with one overriding thought, that 3kg or approx, 30Nis a lot of force, and I suspect far more force than any normal model system usually sees. I suspect that it will not be the snake that fails first but some other component making up the system.

[Erfolg]

Tom

 

Something has to be moving/slipping.

 

I cannot believe that amount of strain can occur without permanent deformation. The mers will have been permanently realigned.

 

Are you pulling my leg

[Tom Wright 2]

Erflog.

My conclusions on the subject are ........The average model control surface load is unlikely to cause significant deflection of the snake when clamped at both ends.

Higher push loads say above 5 kg ,produce progressive bending that results in reduced control surface movement .

Clamping /securing the snake at additional positions as well as either end significantly reduces the loss of movement in high loading conditions.

The finer points of snake quality ,exit rods ,clevis type, horn type,servo gear play ,and control surface horn fixing methods all have a bearing on the final performance.

[Tom Wright 2]

Erflog.

Would not dream of pulling your leg ,unless there was good reason. the snake inner is put into compression during a push load ,eg the servo arm movement via the control surface load ,if the loads are sufficient ,as with any "beam" bending occurs.

 

Edited By tom wright 2 on 27/05/2011 17:53:05

[Erfolg]

I do not think I understand what is happening.

 

If you were take the Euler relationship for slender columns, the assumption is that the column does have a bending radius, and the ends move closer together.

 

Yet this is not a column of that type as it is a compound column, with an outer which is restrained at its ends. These ends are restrained to ensure that they cannot move towards each other or away from each other.

 

In the case we are considering, if the inner is put into compression, any attempt to cripple is restrained by the outer, which goes into tension, resisting the crippling.

 

I have not attempted to calculate the strain involved to get the outer to extend any significant amount, but guess the force would have to be large. Your picture seems to indicate an amount of extension that is so large, that there certainly would be permanent extension/deformation, which I would expect to result in a reduced inner and outer diameter of the casing, this would result in at the very least binding if not seizure. With most polymers that I am aware of ,realignment of the mers/fibrils in polymers is pernament or plastic deformation on this scale is none recoverable.

 

I really do suspect that you are winding me up, hoping I will reach for my calculator, get my text books out, lose sleep, or even build my own tensile test machine for plastics . Hmm you are very naughty

[Tom Wright 2]

Erflog .

This discussion is far to simple to say, a man of your modeling and engineering experience ,does not under stand.

Are we at some yet unrecognized cross purpose ?

 

By the way an additional clamp in the middle results in just what would be expected .....an"s" bend.

 

 

Edited By tom wright 2 on 27/05/2011 19:15:21

[Erfolg]

Oh how you flatter as you mock

Which type and size of model do you think could put 5kg force down a snake or any other device. I an see you are tempting me to check by calculation how just the dowel part of a push rod may behave & then as an of se,t loaded column., Nah, seems like work, did that, got the tee shirt, not going back

[Tom Wright 2]

I don't know if its of any help or significance my demo snake has a 2mm piano wire inner and a plastic outer ID 2mm OD 3.2mm.

[Tom Wright 2]

Erflog.

We must be at cross purposes somewhere,this is not a wind up,why should i,?

Is it the fact that the behavior of the snake under high compression loads is an extreme case not likely to be encountered with the average model.?

After all the difference between a snake and a push rod is that a push rod is less convenient to support along its length,other wise the behavior under compression load is similar.And as mentioned before the inertia of a push rod can be destructive.

 

 

[Bob Cotsford]

Two points regarding Tom's results - 5Kg is more than most servos are capable of providing unless using a minute horn on the servo, and I don't count wire-in-sleeve as a 'real' snake - don't like 'em, don't trust 'em and only use them in ARTFs where they are already moulded in place and won't come under much load (Cularis).

If you are expecting 5Kg servo loads you really would be better not using snakes.

 

Back in the real world decent snakes (eg Sullivan) just don't behave like that, even with 5Kg + servos such as I'm using in my Joker (.90 2 st with twin elevator snakes coupled at the servo end). If I try hard I might get a mm or two of flex, no more without being silly. Remember that a 5Kg rated servo will only generate about 2-3Kg of thrust when using a standard sized output arm - 5Kg.cm torque at ,say, 2cm radius = 2.5Kg going into the pushrod, snake or whatever. Most average servos such as 148s or 300x are maybe 3.5Kg torque?

 

As I think I said earlier, at the sort of loads needed to deform a snake, I'd be more concerned about the strength of the control surface horns supplied with most models and their location in the balsa surface, and I wouldn't recommend standard 2mm connectors for such loadings, so that's snakes out anyway. The Wot 4 XL comes with something around a 13mm dowel elevator pushrod with 3mm end fittings. I'm using 996 (10Kg or so?) servos and I can't get the pushrod to flex for fear of breaking the elevator in the attempt.

[Colin Naylar]

TW2 - I think you are wrong. Put simply, if a force is applied to the control surface eg rudder, then the snake outer would tend to bend and therefore not transfer the movement back to servo (or visa versa) - ok granted. Now if the ends of the snake were fixed, as they should be, then the snake outer cannot bend because the length is fixed - the control surface cannot move!

[Ben B]

There must be slop between the wire and the outer. I guess what this shows is if you've got a slop free system you don't need middle clamping, if you've got slop then you do!

[Tom Wright 2]

Bobs observations highlight some good points particularly the reference to hard wood dowel push rods .the higher cross sectional area does make such a component more resistant to bending under compression loads ,so the lower cross sectional areas used in my test would be more prone,unless supported at several points , but with supports this "low profile snake" has proved capable of transmitting very high loads ,and has the advantage of low inertia .

The discussion has revealed the degree of awareness on the subject ,and the many different approaches that can be used to achieve reliable load transmission.

My test set up was not done to advertise the benefits of any system ,but to merely confirm the behavior of long rods in compression.

The last six or seven models i have built from scratch all use servos mounted in the surfaces opposite the control surfaces,and some have carbon rods to keep the inertia effect and weight down to a minimum, this approach for my purposes at least makes other methods redundant.