Robustness of snakes

[Erfolg]

I have just been messing on the work bench with some Sullivan snakes.

 

The first steel golden wire, just fastened (superglued balsa blocks) at the ends, bent as suggested into an arc. One wire was just held with a cloths peg. I then did some pulling and pushing. The outer did twitch a little, the movement of the free end seemed to be about 1/16" or +/- 1/32. There seemed little difference to tension or the compression which surprised me, as i expected more movement for some reason on the compression.

 

I then super glued two additional blocks at intermediate positions. Without using a DTI (dial test Indicator) I could not see any difference.

 

I also have some of the Golden rods, perhaps not surprisingly there is more end float on these, I would suggest about 3/32", a lot more than the cable. I strongly suspected that gluing the intermediate blocks would make a big difference. Surprisingly it does improve things a bit, or so it seemed, but not enough to show up using a steel rule for measuring.

 

I have done one other little test, which was a bit obvious really, I removed one of the end blocks nearest where I was pulling and pushing. Perhaps not unsurprisingly, the free or unwanted movement increased dramatically, the main problem in my quick test was keeping the end of the cable along the axis I wanted to measure. On a model this would not be a real issue due to the clevis and horn. But it was clear to see that now I had movement, So I removed the middle block, so fastened at two ends again. The result seemed much the same.

 

So this rough and ready test seems to indicate, if the two ends are fixed securely, there seems very little difference with fastening the intermediate points. However leave a long distance of unsupported cable and outer at the ends, there is a lot of lost movement. It seems from buckling.

 

I am a little perplexed as the movement seemed to suggest a change in length, but then again the length does remain the same. So I guess the results speak for themselves

 

It would be interesting if others try the experiment, if their results are the same. Particularly if they take a bit more care than my test. Using a DTI and some set end load. Although I would argue that my set ups were not much different to many models, other than my cables are much longer, therefore should represent the worst case.

 

The original issue with IPA, suggests ESC (environmental stress cracking) not an unknown issue with many polymeric materials. I understood one of the limitations of many plastics, yet we all live quite happily with issue, so perhaps not the show stopper that initial reactions to its discovery that it may suggest.

[PatMc]

In disagreeing with David on whether snakes need supporting or not seems to have led to 2 separate issues. The first is whether supporting them prevents “blow back” the second is whether it’s best to support them in in case one of the end fixings fails.

Taking the second issue first IMO it’s a matter of confidence. Personally I’m happy to rely on well made end fixings using the installation procedure I outlined previously, if others prefer the belt & braces approach it’s their choice but it has nothing to do with the first issue.

There’s also a practical consideration with some fibre glass and plastic fuselages in that it would be very difficult to make intermediate fixings. In which case unless it is really necessary I contend that it isn’t worth the effort.

The first & main issue is whether intermediate supports do anything to prevent blow back.

I repeat that it doesn’t.

If identical snakes are employed using the same material both secured at the same distance apart at each end one with several intermediate supports the other with none then they will perform identically.

 

Posted by Biggles' Elder Brother - Moderator on 19/05/2011 00:32:03:

Pat, I think what Peter and DF are refering to is this;

 

A snake operates two ways, in one direction it pulls - this is no problem, a snake under tension won't bow.

 

But in the other direction it pushes. Now suppose, hypothetically, the control surface couldn't move - but the servo was still pushing it through the snake. Something has to give - but what?

 

If the snake were perfectly rigid - like a very stiff pushrod - then the something that would give is the servo and the outcome would be that the servo will stall.

 

But if the snake can bend relatively easily (because it has a significant unsupported length) then it will be the something that gives and it will bend or bow along the unsupported section. This will happen because the servo end of the snake will move, but the control surface end cannot - the movement has to be taken up somewhere - the somewhere will be the bowing of the unsupported section of the snake.

 

Basically the system will give at its weakest point - either the servo or the snake. If the snake is well supported at regular intervals along its length then it will be effectively very stiff and it will not be the "weakest link".

 

If the exposed inners ends are so stiff that they can’t bend at either the servo or the control surface ends the situation will be exactly the same as with a stiff pushrod.

Why on earth should the pressure on the inner cause the snake to bend between the fixed ends and how would supports make any difference?

But if it did bend why should that have any effect ?

Inner & outer are identical lengths between end fixings. For the inner to shorten at one exposed end only (blow back) would require the outer to expand in length or break.

[PatMc]

Peter Beeney,

No need to apologise, I hadn't noticed the misspelling but that's because it's one I use elsewhere.

BTW my snakes have been tested in gliders using "ping" launches from full pelt hand tows & electric winching in sport & competition flights so I'm quite confident in my methods.

 

[PatMc]

Posted by Stephen Grigg on 19/05/2011 21:49:04:

Ive built a Red Zephyr and it has a long fuselage.I thought I had supported the snakes ok but Im wrong.The Rudder is badly effected in the air it hardly lets the model turn right,.Its perfect on the ground and left in the air,but the snake is pushing when you turn right and the wind flow is obviously pushing the rudder back and flexing the snake.I may improve things if I can support the snake where it comes out of the fuselage.Its over 12 inches to the rudder and I believe this is where the problem lies.

Stephen I suggest that the snake outer needs to be fixed as close as possible to it end with as little as possible of the inner exposed, even on a slow flyer I'd limit it to a couple of inches.

Have you considered a mini servo mounted near the tail with a short wire push rod ? The Zephyr rudder won't require a lot of force, a small servo will probably weigh less than the snake even with the necessary extension cable so there should be little if any change in cg.

I use a couple of tiny Hobbyking servos mounted in the tailplane of my Veron Deacon. I can't remember what they weigh but it was less than snakes or pushrods. Pull-pull fishing trace wire would have been lighter but less convenient.

[David Ashby - Moderator]

Good interesting debate though guys, thanks and I'm always happy

to be 'put straight' Pat You never stop learning with this hobby.

 

 

 

 

 

Edited By David Ashby - RCME Administrator on 20/05/2011 08:17:38

[Alan Cantwell]

that red zephyr rudder doesnt need a snake, it needs closed loop,

[kc]

This has been an exemplary discussion, but I still think David was right originally ..... snakes do need support in practice. It's what works.

[Peter Beeney]

Pat,

I certainly accept what you say, and it undoubtedly works for you. I would not argue with that. However I’m sure that’s fine all the time the servo lines up with the exit point, line of sight, so that you can keep the snake taut, but I thought one of the advantages of using snakes was that they would go round corners, such as a right angle turn and running up the fin on a t-tail, a nice example was the Radio Sailplanes Scirocco. There’s a blast of nostalgia! I flew one of those for a long time. Plus, as you say, you can make sure you get the very best quality items. I’m afraid the beginners are not always so good at it, and invariably not nearly as knowledgeable as well. Some of the early varieties of plastic control rods were a bit flimsy, too, as I remember. Although I think things have improved greatly in that respect since ARTF as apposed to building the kits of yesteryear.

I can certainly identify with the wire insert to support the ends, when I examined the ends of my first snakes I decided they needed a little reinforcing and I remember scraping and sanding some cocktail sticks down fine enough to fit! I also carefully adjusted the outer tube for length, so that I had full movement with the adaptors just about touching the ends! Minimum amount of inner showing! I must have been really keen! Cocktail sticks, and the longer type, I think they may be kebab skewers, have almost infinitely different uses within modelling!
Later I seem to remember some tubes having the requisite wire in the package.

I’m a bit too long in the tooth to change ideas much now anyway, so I shall go on checking the soundness or otherwise of the control linkage on any unknown quantity. Although I must admit, I haven’t had any problems in this respect for a very long time. The last little cause for a ripple of interest was when someone recently bought a DeAgostoini Spitfire via eBay and I was asked to punt it around for a test flight. Apparently it’s first ever. On the first return pass over the strip the tail plane suffered severe flutter, enough to instantly take the back end right off! I had remarked on the check out that it looked a bit flimsy, it was obviously worse than I thought!
This may or may not be tied up with the control linkage, and I can’t remember what fashion it took, it might have been a snake, but I reckon if there was any slight play in the system it quickly became a full blown oscillation!

I guess it’s true what David said, we continue to learn all the time.

PB

[Erfolg]

I was interested in some of the comments about just grabbing the rudder and trying to move it. Rather than the laborious test I did, which indicated that the intermediate supports do not do much.

 

Surprisingly I found an old trainer, has a lot of slop, what interested me is that the push rod is almost 3/8 balsa.

 

I will try some more practical tests tomorrow on some of my models, many with vatious guises of snakes.

 

 

[Erfolg]

I was interested in some of the comments about just grabbing the rudder and trying to move it. Rather than the laborious test I did, which indicated that the intermediate supports do not do much.

 

Surprisingly I found an old trainer, has a lot of slop, what interested me is that the push rod is almost 3/8 balsa.

 

I will try some more practical tests tomorrow on some of my models, many with vatious guises of snakes.

 

 

[PatMc]

Posted by kc on 20/05/2011 16:38:18:

This has been an exemplary discussion, but I still think David was right originally ..... snakes do need support in practice. It's what works.

 

kc, the explanations I've given were not just reached from theorizing. In practice supports are not needed for the reason David originally suggested.

 

Posted by Peter Beeney on 20/05/2011 22:43:35:

Pat,

I certainly accept what you say, and it undoubtedly works for you. I would not argue with that. However I’m sure that’s fine all the time the servo lines up with the exit point, line of sight, so that you can keep the snake taut, but I thought one of the advantages of using snakes was that they would go round corners, such as a right angle turn and running up the fin on a t-tail,

Peter, if snakes are kept straight they will have less wasted end play under normal pressure than if they are curved. Remember that the amount of play is the dependant on how tight the inner fits in the outer. It is the result of the sideways movement that the inner can make within the outer when it's easier for it to move that way than against it's load.

With a curve it's easier for the inner to bend within the outer & the friction is greater than with a straight run. A straight run of the same length cable will potentially have the same amount of play but it will usually only with excessive pressure. It doesn't matter whether it's straight or curved any intermediate support will make no difference.

 

Personally I'd never use snakes with a significantly curved run for any control needing minimum slop.

 

A friend of mine used to have a large thermal soarer (IIRC 140" span Medicine Man) that had the AMT mounted half way up the fin driven by a snake through 90 degrees. The model was very well built but up elevator control always had be given a touch of down to return to neutral & vice versa for down but sometimes it needed a bit of juggling to get back to the same trim. I think that he eventually incorporated a spring to bias the movement to up elevator.

 

Recently I saw a pair of Hurricanes built from the old Mick Reeves kits which had a single central servo operating the ailerons via snakes buried in channels in the foam wings. The amount of slop was frightening.

[Grasshopper]

Hot glue is useful for tacking loose servo leads etc. inside the fuselage but do not use it to tack servo snakes to support them as it can melt the outer and cause binding.

I have been caught out several times as the snake appears OK but under the blob of hot glue the outer may have deformed and nipped the inner.

 

You have been warned!

 

Peter

[Bob Cotsford]

I remeber a couple of models (still got the cotswold Models Skylark)where the ailerons were driven by Bowden cable snakes curving through 90 degrees from a central servo. Aileron control was at best a suggestion to the model. Just like Pat, I'd never use a snake where a significant change of direction is needed, just too much friction and lost motion even with the snake embedded in a foam wing. even belcranks and pushrods are better.

 

As for pushrods, as Erfolg has stated, they will allow the surface to deflect one way quite easily as it's almost impossible to stop them bowing under compression - hence my preference for snakes or closed loop.

 

Good comment re hot glue Peter, I always use masking tape + epoxy to hold snakes and the only failure I've had was when the epoxy came away from the glass fuselage after a hard slopeside arrival. The epoxy was still firmly attached to the snake though!

[PatMc]

Bob,

I usually use cyano & micro GF to secure snake ends but if I use epoxy it's with masking tape. Also if I'm film covering a veneered foam wings with a GF bandage centre joint I cover the bandage with a layer of masking tape first. Sticking film direct to GF bandage is never satisfactory but it's fine on masking tape.

[Erfolg]

One of the contributors wisely suggested what matters is practice. I agree, recognising that the theory suggests what is happening, the testing and research, recording results and critical peer examination of results until the theory becomes fact or dismissed as a false preposition.

 

So I have done a few tests previously, with some snakes on the bench. I then decided to examine all the models I have, to try and measure, or at least a empirical assessment.

 

I did raise an eye brow, to those who suggested that they would never use a snake on a full blooded towline launch, although i may be misquoting.

 

So my results are a little surprising.

 

1) Long "push rod", even with an intermediate support/guide are my poorest linkages. Used on an electric Tomboy. The rod flexes and the end wires/bicycle spoke, is less than absolutely rigid. The same is true of my Ridge Rover, although no intermediate supports/guides and short bicycle spokes.

 

2) A 144" open glass bodied glider using "closed loop" on both all flying tail and rudder, are very good. To get the best out of the system I have two straight bell cranks, which the servos drive via a push rod, to avoid wear of the servo output bearing.

 

3) Now addressing what was originally asked, " Snakes".

 

The two models above both use snakes. Both have no more slop than the best of the rest.

 

The small glider, has the snake bound on as a fishing rod whipping is dead straight, there is no slop, in the linkage.

 

The large body, was a 120" open glider, since motorised. This model only has the extreme ends fastened, being inside a fishing rod. It has endured more lung busting, full blooded, tow line, power winch launches than my granddaughters have had hot dinners. I could not estimate how many bungee launches, a great deal.

 

PatMac is absolutely correct, there is no slop at all on this snake system. It is a s good if not better than all the other examples I have.

 

I would suggest that anyone making a poor system, probably has more to do with what they have done than the system.

 

There is one proviso, I always in the past arranged that a pull would give up elevator. As I was concerned that a unsupported, wire could buckle , as a slender column. I have noted that the slender column effect seems worst on push rods, particularly where the metal bicycle spoke is above an inch or so (some of my own are as on the Tomboy, no problem with Ridge Rover).

 

My current favorite system is to mount the servo as near the surface as possible, with a short push rod. As the amount of slop is very low, even if ham fisted. Even then a piece of CF over the spoke improves the crippling (slender Column) issue.

 

With respect to my test, and practice seems very close theory. In that the outer cannot change length, in any material manner at the stress levels we can develop. The same is true for the inner, though the ends could cripple, if care is not taken to support with the outer to as near the end connectors as is practically possible

Edited By Erfolg on 23/05/2011 14:43:31

[Bob Cotsford]

Thanks for that Erfolg, it confirms what I thought from past experience. I second your comment that 'Up' should always be a pull action, so I was surprised on assembling a Wot 4 XL ARTF to find the elevator pushrod is below the tailplane and pushes for 'Up'. Admittedly it does use 3mm rods for the twin elevator connections, but I'm still not entirely happy with it.

[Erfolg]

My results caused me a little concern. The concern centered on pushrods and how frequently they are used, how passionately they were pushed in the past as the ideal means of connecting a servo to a control surface when separated by a longish distance. Often used as the "gold standard" against which other systems were compared.

 

What concerned me was how much they flexed.

 

To be completely honest, i had expected and hoped that BEB would have waded in with an explanation why they are less than ideal.

 

A little thought did convince me how with a slenderness ratio of between intermediate and long as "Euler", they were susceptible to buckling at low loads. This is made far worse by the axis of loading, not being along the axis of the rod assembly, as idealised by friend "Euler". Made even worse by the fact of the non, isotropic (free of irregularities, such as grain etc) nature of wood, plus the binding arrangements of the wire appendages etc.

 

So perhaps it is not surprising that long push rods do not perform well when compared to a snake, in empirical and probably theoretical examinations.

 

I guess what has happened, that results with Bowden cable could be erratic in the past, due to high sticktion/frictional forces with some combinations, often requiring careful lubrication to obtain acceptable performance. With the event of Sullivan type snakes, sticktion (sticking and then moving) is not an issue, friction is low, and the fit between the inner and outer is so good that little slop occurs within the snake.

[Tom Wright 2]

Some of the comments made on this interesting thread seem to be inadvertently in defiance of the laws of physics in that if a strait line is then bent the distance between the two ends will reduce,or if the ends are immovable the material must expand,either way a change in length will occur .

A long snake in push mode,just clamped at either end pushing against say a heavily air loaded rudder can bend thus changing the length.,and reducing the rudder movement available.

The classic case of rudder control were the fus interior is not accessible is IMO best dealt with by the closed wire system.,and long snakes surly must be better supported along the length at regular intervals.

Sorry to come in so late its a good thread with some interesting views the above is just my inital take on the subject.

 

Tom.

[Bob Cotsford]

Tom, in practise if the ends of the outer are held securely then there is a minute amount of movement as the inner takes up the few thou of clearance in the outer, but the forces required to get a significant loss of movement through deformation of the material of the outer are more than flight loads are likely to be. More common will be the situation where the length of exposed inner is too great and that section will buckle. Easily cured by bunging a length of wire down any exposed inner, or just keep the distance from the end of the outer to the horn to a minimum.

With most ARTFs I would be more worried about flex in the plastic horns supplied.

[Ben B]

Tom- I think you're thinking of these

 

because that's the only way the length of the snake outer between two fixed (clamped) points can change length!! Forget physics, try logic An outer snake clamped at both ends cannot suddenly change the length held between the two clamped points without stretching. Snakes are generally made of a non-stretchy material.Therefore the length between the clamps cannot suddenly change.

 

Most outers are inserted in such a way that they are slightly curved. Whilst you can get movement or writhing of the curved snake when the inner is moving the outer length remains constant. Unless of course the outer snake is made from a stretchy material or points A+B are moving in relation to each other. And if the fixing points are moving in relation to each other then they're not very good fixing points.... and you probably have more significant things to worry about (like the fundamental lack of structural intergrity being shown by the plane).

 

If you have slop within the snake due to the I/D of the outer being greater than the O/D of the inner then you could problems creeping in due to the length of the inner being held within the outer changing due to flexing. However most snakes do not have much slop therefore negligible improvement with intermediate clamping. An exception would be certain snakes of the "thin piece of wire in considerably larger plastic tube" type which can demonstrate this. With a slop-free tight fitting snake this isn't a problem. If there is slop in the system having the outer's clamped (therein stopping movement) will be of benefit.

 

A big source of slop in systems is long unsupported inners at either end of the outer. But this won't be cured by having intermediate clamps.

 

The only advantage of intermediate clamps in a slop-free system is that in the case of a clamp coming undone or breaking you still have two fixed points so at least the length of outer between those points can't change.

 

 

[Tom Wright 2]

Thanks BOB and Ben for taking the trouble to re - explain your findings , i will now go away and give things some further thought,should that result in any startling revelations i will let you know.Oh i forgot to thank Ben for the graphic illustration.can i get those from GC ?   .

Thanks again.

Tom.

Edited By tom wright 2 on 24/05/2011 16:14:08

Edited By tom wright 2 on 24/05/2011 16:15:13

[Erfolg]

Ben B

 

You would be a real hero with my two granddaughters, a bag of gummy snakes, and not even opened.

 

I must keep an eye open for them in the shops, obviously not Haribo, but who makes them, I cannot make them out?

 

I guess not much good for models though.

 

To be honest this exercise has opened my eyes, the order of preference for me, now is

 

1) short stiff pushrod, where servo and surface are closely located.

2) Snakes, not quite as good, but not bad and flexible (get the .......).

3) A distant 3rd, close loop, as good installation takes some effort, that is both plastic horn and driving horns must be equal centres, adjustment is needed to get the fishing trace wire slop free, ideally the run should be straight, ideally the trace should not be crossed over, as wear of the plastic outercould causes metal to metal contact. Ideally should not be driven indirectly, unless servo has a good bearing to resist wear. Ohhhh!!! so much effort.

4) Long push rods are not good at all, must be light if inertia does not become a problem, particularly on an arrival, where the gear train could "wipe out". Needs to be stiff, straight etc. Some of the requirements do not sit well together. I suspect that these systems now belong in the past, or a desperate last option.

[Tom Wright 2]

Erfolg .

That sums things up nicely IMO, and certainly that nasty Mr inertia has a lot to answer for when it comes to push rods and firm arrivals.

Tom.

[Tom Wright 2]

As Erfolg took the trouble to carry out some pratical tests i thought it would be a good idia to try some of my own .

Inner and outer perfect fit no side play at all.

Outer clamped at both ends only .

Inner clamped at one end to reresent high air load on a control surface .

Push load applied to other end of inner.

RESULT.

The inner and outer bowed in the middle reducing the length of the coupling between the servo arm and control horn by 1/8"and so restricting the conlrol suface movement by a significant factor.

This conclusion is pretty obvious without going to all this the bother but it does confirm my origionl post thoughts,and the fact that snakes with supports just at both ends can bend under load and cause descepancies in control surface movement.

 

 

You need to click on the photo twice to see the set up detail.

 

Edited By tom wright 2 on 24/05/2011 18:49:31

[Peter Beeney]

It would appear that it’s a bit of a personal choice thing when dealing with these mysterious plastic tubes. If you make and fly only your own models then you know it will be ok, it’s when you fly someone else’s model, without checking it, as I did in those early days, does it all go a bit egg shaped and you do the Humpty Dumpty exercise and fall off the wall, never to be repaired.

After reading Erfolg’s interesting and comprehensive narrative, I remembered that I’d been given a glider, about 100 inch span, with a view that it might be of interest to someone. Unfortunately both halves of the all moving tail-plane are missing so I’d never even really looked at it previously.
I dug it out and found the rudder is a closed loop but the elevator is a snake. The fuz is about 51 inches long, including the rudder, and the snake about 34. It’s a Bowden wire running in a black plastic outer. It all looks slightly elderly and has been flown. The threaded adapters are soldered on, quite nicely, and there are metal snap links both ends. Very solid! So it was a bit disappointing to find that when I held the AMT bell-crank locked I could easily turn the servo arm in the pushing direction; that’s for up elevator, needless to say! The back end is completely enclosed, so I cut the covering away to find that the end of the plastic outer stops about 3.5 inches from the bell-crank, with about 2 inches of exposed wire. It’s this wire that’s simply flexing upwards when the servo turns. Plugging into a model with a battery and the servo easily turns again, flexing the wire, effortlessly. Also the effort to hold the bell-crank, with a piece of wire though the operating arm, is not that great, but of course when it’s free it has full travel. It’s ironic to think that the wire is strong enough to almost hold up a house and yet has this very weak area when used to control a model aeroplane. Pulling for down elevator would stall the strongest servo before any thing broke. I’d have thought that if the outer had been taken as close to the operating arm as possible and some solder run into the wire to stiffen it at the ends, that would make it rigid enough. There is about an inch of wire at the servo end. Inverting this action and holding the servo arm, the operating arm can easily be lifted, exactly as you’d expect.
I think there is every chance that if the gilder reached a critical point, and was perhaps over-speeding for whatever reason, the load could increase to the point where this elevator might not operate correctly. On the other hand, it might never reach that point, so it might never become apparent there was a built in requirement for the ubiquitous black bag! The closed loop rudder is solid enough.

For me, this only really confirms what I’ve thought many times in the past. It’s very easy to think I’ve got everything covered, but frequently I’ve missed something; and back in the early days knowledge was the all important thing, too. Sometimes there was a reluctance to pass it on as well, that may still occur even today; and when modellers used to build models from kits occasionally there were some weird and wonderful concoctions. The ‘Z’ bends used to exit the fuselage for instance, one I remember particularly. It really did have a strange effect. Regarding the long balsa pushrod effect, at one time it was de rigueur to use large cross sectional area lightweight balsa sticks, which would flex nicely, undoubtedly at the most inconvenient moment. We soon learnt the trick of fixing ‘gates’ in the fuz to keep them in line.

I think we did all the theory and the testing, way back, but as always, we just give everything a real workout before lobbing off, if it shows any signs of flagging in any way it gets fixed first. That’s always worked for me, at least mechanical failures are very few and far between.

Now I just have to eliminate the all the other kind of failures and then we’ll be on the road to success! Maybe!

PB