Take Off Dolly

[Gordon Tarling]

Erfolg - from my own limited experience and observations that I've made in the past at my flying field, one of the biggest dangers of dolly launching is of the model jumping out of the dolly before it's ready to fly, due to bumps etc. Whilst I hear what you've said about your particular field, I've never considered steering on the dolly to be necessary. The model's rudder seems to give enough steering effect once it has a little airspeed.

[Erfolg]

Gordon

Most of my models without UC have no rudder.

The bumpy field and lack of rudder, is the reason as to why i have provided steering.

I have now connected up the Nicad and Rx and will attach a photo tomorrow.

I have a servo reverser for the models that do have a rudder, which acts in reverse to my dolly std, set up.

It is the issue which you have raised is the final conundrum for me. The fact is that no amount of pondering, compares with the experience gained from actual operation and observation of the real world, to form some real understanding of what really matters.

[Frank Skilbeck]

Erfolg - don't forget to post the video when you test it

[MABX]

Erflog - i have been thinking of building a dolly for the same reasons that you have identified and will be a must for a couple of models i plan to build over the winter. I can see the benefits of the rear steerable wheel for the dolly but did not like the idea of connecting this to the model and trying to find a fool proof way of not interfering with the model on take off.

Our club field has one or two bumps and there is a slight camber to the strip being on top of a hill which might cause a problem for the straight running of the dolly which got me thinking. I am going to wire a cheap heading hold giro to two servo testers, one to set the centre point and one set the gain. The whole lot should cost under £20 pounds from HK and will be completely standalone from the model. This system would have the advantage of holding a straight course and see any bumps and correct long before i could correct manually. Like putting a gyro on the rudder of a normal model was my thought.

I could probably refine this with an Ardunio set-up to "unlock" the model to allow release after a given distant and handle the gyro function as well. Anyway small steps first, going to try the servo testers and gyro first.

[Simon Chaddock]

From a purely technical point of view rear wheel steering becomes dynamically unstable as the speed rises.

Tail draggers rely on the fin becoming the predominant force before the 'speed of instability' is reached.

Fast heavy planes usually had to have locking tail wheels to avoid the problem.

Of course a servo is non reversible so in effect the tail wheel behaves as if it is locked but any 'steering' at speed will make the dolly unstable unless it is damped by the fin.

It will be interesting to hear (& see) how you get on.

[Stuart C]

I have never used steering on a dolly, though did once consider it. Realising that steering inputs were more likely to cause, rather than rectify stability problems at speed, I opted for small rudders. After constructing this particular dolly I then subjected it to a number of dry runs over a long smooth surface, to discover that the friction from the wheel axels was not only high but also varied enough to give adverse yawing. These were proprietry plain bearing castors, which I then had to machine out to accept twin ball races. Hard work but well worth it. The rudders became totally redundant - but looked good! I might add that I have only used dollies off tarmac or concrete. Off grass? Good luck.

[MABX]

A heading hold gyro will dampen the swing at lower speeds much the same as it would if fitted on a cub at speeds where the fin/rudder is not yet fully effective. I have seen a very good gyro setup on a 30cc cub also emulated on a much smaller model cub with very similar predictable takeoff characteristics and virtually eliminated tail swing.

As the speed and airflow increases over the fin of the model attached to the dolly the fin should provide more stability as speed increases with less correction force needed from the gyro. If this keeps the dolly +- 20% from centre then this will fine for my purpose, i just want a repeatable track. Most videos I have watched on youtube with a dolly seem to suffer from being knocked off their track relatively early in the takeoff and end in a sort ground loop. The gain on the gyro will be key to the dampening of the swing and could differ depending on model , anyway will give it a try as i think this will be worth exploring and is a work in progress.

[sticky fingers]

Erflog after many months tested my launch trolley as romeo whiskey or vice versa what a cracking bit of ki, works a treat ,I used a foamy acrowot as a test plane next will be a cambrian fun fighter I,ll post the results

Dave

[sticky fingers]

PS you really dont need the steerable tail wheel REALLY

[Erfolg]

Most unfortunately the indications are I do need a steerable dolly. The reason being our field (club No.2) is an old rubbish tip, that has settled unevenly. The consequence of this is that most take offs require a rudder input, sometimes of a continuos nature.

I have been thinking about the shimmying that is possible of castering wheels. Particularly Supermarket trolleys. Noting that although they all seem pretty much identical at a location, most do not shimmy, the odd one does. I think that potentially that the variation is dependant of the freedom of the spindle to revolve, a inherent damping coefficient, and what ever the mechanism is that controls the oscillation. I know from my youth, that there will be a Sine value, probably gyro effect, dependant on the wheel characteristics, the speed of rotation involved as well as the coefficient, beyond these rather basic thoughts, i know nothing.

But it did strike me early on, that a locked arrangement, may be subject to the forces, but the result, may or may not be some noticeable vibration. As long as the forces are something less than the 2.4 kg cm force that the servo can generate, the wheel will be locked, as far as any oscillation is concerned.

I do have a gyro available, and have though a week back that it could be used to lessen the work load.

Time will tell on all of these aspects.

As far as the WR design is concerned, my own dolly is essentially the same, it may look a little different, although the characteristics are much the same, other than the steerable tail wheel. It is this tail wheel that has lead to the AoA being something different. I will do the calc to try and get some handle on the WR AoA, for purposes of comparison.

Perhaps what needs restating, this is not intended to be a ground breaking new design or concept, the objective is to identify what actually is important and works, what does not and to try to make sure my dolly or the modified WR dolly incorporates them.

Edited By Erfolg on 17/09/2015 11:03:49

[Simon Chaddock]

Erfolg

It is not wheel shimmy that will be the major problem (as long as you keep the free play to a minimum) but the actual dynamics of rear wheel only steering.

No problem at low speed (fork lift truck?) but no high speed ground vehicle uses it for this very reason.

It will all come down to the point at which you stop inputting steering corrections and effectively 'lock' the rear wheel.

[Erfolg]

Simon

I am not getting the picture. i need a bit more descritive text.

My descriptive model has a tailwheel, has in essence the same geometry RW and Electricflyer. As far as I can see as long as the servo is up to dealing with the forces induced from the tailwheel, it is the same?

[leccyflyer]

Gents I think there is a danger that you might be over thinking this a bit - try a simple dolly, without any steering, and see how you get on - if the model has sufficient power she won;t be on the dolly for long enough to make any difference and if your take off is 10 degrees off, will it really matter? The point about the rear wheel steering is probably correct, - you would be trying to chase a shimmy instead of concentrating on getting the model off the ground.

[Romeo Whisky]

Hi again Erflog,

I do have to agree with the above comments. I really cannot see a steerable rear wheel working, but .... hope it works for you. It would be a pity to mess up a nice model if it cartwheels.

For what it's worth there's a YouTube Video of a dolly based on my design being used to launch several models. He's used two rear wheels, but smaller than the single one I used, but I'd guess the weight would be similar to my big wide one. I've never had a failed take-off with my final design, but I do use some up-elevator to keep the nose up and right rudder to keep a straight line in the take-off run, although since it is mainly a torque-roll issue I suspect right aileron stick might do just as well.

The link is here

[Erfolg]

I have a number of models which have no UC that also have no rudders. At the time my decision that the additional servos would be dead weight.

The need for a hand launch is an issue, particularly at one of my clubs, in that there is no longer ant one who is capable of the task.

Recently I went to a fly in at the Bickershaw club. Their field is incredibly flat, the grass short and totally manicured. I had visions of press ganged club members, on their knees, with nail scissors, achieving the perfect cuit.

Then ther is our field at club no. 2, we fly of an old rubbish tip. It is not flat at all, the surface is generally irregular. In one particular direction there is a depression of about 3-4 feet, with probably 30 feet diameter. The grass cannot be cut any where near as close as many clubs enjoy.

In general take offs initially require some rudder input, sometimes many multiple inputs, occasionally none.

On this basis either a steerable dolly is a must. Or the model needs a rudder, when using the dolly, where I would try just not switching the steerable dolly on. Even a fixed rudder will be effective as a damping device as the model speed increases. On normal models, as the speed picks up, far less rudder input is required. This should be no different in principal to a model with a steerable tailwheel.

If at all possible i would much have preferred a non steering trolley.

For me I need as much input from the dollys that work as possible to try and reduce the risks involved.

I am now looking to provide a low AoA, with a pair of arms which allow the model to both find a flying position, and freedom to raise clear of the trolly.

[leccyflyer]

Concurtotally with your views on handlaunching - some folks really have the knack and can hand launch anything - but lots of folks don't and I've seen (and experienced) lots of models wrecked by poor hand launches.

In my experience, off grass, with the simple lightweight dolly pictured above, I've successfully launched models both with and without a rudder with only the occasional failure, where the model has flopped off the dolly without getting up sufficient airspeed. In general they aren't on there for long enough for steering to be an issue, but the tailwheel can be set (by experimentation) to get a straight run.

Before the dolly we tried catapults - now there was a futile, but hilarious, sojurn into making launching a model that could easily be hand launched as complicated and unusccessful as possible.

Cost of importing a pair of catapult kits into the UK -=about £65 each - cost of assembly,=another few quid for some ally extrusion from B&Q - time expended putting it all together,=not too bad, coefficient of friction of the ply wood model carrier along the aluminium bar -=gigantic, -similarity to the opening sequence of Fireball XL5=- high, at least as far as the slow acceleration and rocking from side to side went - length of launch achieved = just slightly less than the fuselage length of my Bearcat - entertainment value for those watching = priceless.

[Erfolg]

Leccyflyer, much of my flying has been with flat field Glider Guiding. All launches have involved either, towlines, bungees and power winches.

What ever the claims of some, nothing beats power/tension for successful high launching, with gliders.

Perversely i have also launched some smaller models, which were non glider type. As you found, success is far more difficult to achieve. I came to the conclusion that the issues were a combination of the lower wing span, and higher wing loading. I also fond that some were also unstable in pitch, snapping from nose high to nose low, and visa versa. On that basis other than getting a model a few feet up with a little airspeed, that bungees seem to offer little. I have seen a few DF launched with low power, short length bungees from a frame. Again this seems to work to some extent with pretty low wing loaded models, where the real issue is getting the fan unit working more effectively.

With respect to hand launching, the members of one of my club are much older, than they were and older than my other club. It does seem that age to rob us all of our athleticism. Things that once were done easily, become more of a challenge. I do remember one incident where the model was launched nose high, on its side, with no forward speed. The model saved itself, due to a fine pitch prop, a degree of power and low wing loading. Just a few years earlier this guy was 100% reliable. It seems that is how things go.

[Erfolg]

I thought I would post a picture, which comprise a collection of my odds and sods of dies and taps, with which i intended to alter the thread on a spoke.

The other item of interest is a spoke. I could do with a pile more of these things. They are 0, 095" (2.5mm) dia no longer made for down hill mountain bikes, Apparently quite useless for a normal wheel, as they produce a heavy, unresponsive wheel. Yet for modelling another thing. Unfortunately I do not have any nipples that fit, so will die down the spoke to 8BA. What is good for this application is that they have a smooth black coating. On that basis I do not anticipate that there is any need to cover with heat shrink.

The final item of interest is my late fathers Micrometer. Most, if not all engineers will know that the traditional Mic has a thread of 40 tpi, whic means that each revolution of the barrel is 0.025". Then the thimble is divided into 25 increments. This Mic is different, in that the threads must be a 20 tpi, where each revs is 50 thou. Never seen one before, that is other than my Metric/English Mic made by Shadlow, which is also some strange arrangement. In both cases i have to look carefully to decide how to use them.

[leccyflyer]

The catapult of which I speak wasn't a bungee, it was a genuine catapult, completely self contained within the 8 foot length of the frame. Most of the bungees that I've seen in use have been much simpler and very effective. Not so for the catapult.

[Erfolg]

I have been more than a little concerned as to Simons comments on rear wheel steering. On that basis I have been doing a few tests/experiments.

Over the last two days I have used my PZ Albatross, with a fixed tail skid and my HK Arcus, with a steerable tail wheel,

I fast taxied both models, a number of times on each day. I over steered deliberately each time, slowly increasing the speed until lift off. Then returned.

I am now rather reassured at the speed our models reach for take off, there is no sign of any instability at all. That is on these two models, on essentially still air days. The response in most respects seemed pretty much the same, at the beginning as at the end.

I am hoping that this rather small number of checks indicates that any problems, occur at a higher speed.

[Simon Chaddock]

Erfolg

Whilst your planes might be acceptably stable with their combination of main wheel position relative to the CofG and with the fin providing damping, the dynamics are likely to be a bit different with the dolly in place.

The combined weight of the dolly and plane will be significantly greater and its effective centre of mass rather further back than the planes CofG. The greater mass will also mean the fin provides less effective damping.

ll these factors are likely to make the dolly/plane combo less stable.

The preferred option is to keep the dolly as light as possible and to place the main wheels no further forward than they would be if it had fixed tail dragger undercarriage.

Of course the only way to find out if stability is a problem is to try it!

[Erfolg]

Hmm, it seems that much in life and the real world is not as clear cut as you tend to expect and often hope for.

I am not sure that the fin argument is very compelling. In that with or without steering, the model still has a fin.

It seems that at high speed cars such as F1 cars and their ilk have designed in passive rear wheel steering, to increase stability at speed. Perhaps more surprisingly to me is tat a lot of large commercial vehicles use either passive or active rear wheel steering.

I have only found one reference which suggests that rear wheel steering causes natural oversteer, where front wheel steering causes natural understeer.

This a definite area where the more you read more are the issues, wether front or rear steering, where stability seems to have many aspects. Front wheel steering has significantly more research than you can shake a stick at. Not only the true radius that the wheels roll around, but also the angle of the wheels relative to the forces I am now considering reducing the possible angle that the rear wheel is capable of moving.

What has influenced me now is my observations, that there appears to be non problems at model speeds and this:

The driver, any driver, always has the steering wheel in front of him or her - right? Such astounding good sense leads to the logical conclusion that the steering wheel will therefore be connected to movable front wheels for the purposes of keeping the vehicle on the straight and narrow. Conventional wisdom therefore has it that the steered wheels on any vehicle other than dump-trucks or similar slow moving industrial devices need to be those at the front in order to excercise proper control.

Designers of land speed record-breakers are not usually constrained by what has gone before, so it should come as no suprise to find that a car as radical as Thrust SSC should stand conventional wisdom on it's head and opt for rear wheel steering, a first on any car targeted at the Land Speed Record. The world's fastest dump-truck? - hardly.

Edited By Erfolg on 21/09/2015 16:47:24

[Erfolg]

Like most other elements of it's radical design, Thrust SSC's rear steer layout emanates from a couple of basic principles that have guided the overall design philosophy from it's earliest days. Firstly, the car has to be inherently safe, which at the sort of speeds envisaged can be summarised in one word - stability. The adoption of the stable dart principle dictates that the heaviest parts of the vehicle, the engines, need to be installed towards the front of the chassis. Each needs it's own specially designed intake which at full power will be forcing several tons of air per minute through the engines.

While air intake design is crucial to the correct functioning of the engines, they also compromise the second major principle of the design, that of minimising cross-sectional area in order to reduce drag. A logical place to install the front wheels is at the outer edges of the engine pods thus providing stability by giving the widest possible track and leaving the intakes free from interruption. However, if you make those wheels steerable it increase the cross-sectional area, and therefore the drag, by an unacceptable level. Answer? - fix the wheels and think of some other method of steering.

Fixed wheels and a movable rudder mounted under the nose of his jet-powered Sprit of America, the first jet car to take the LSR, was an idea first tried by American Craig Breedlove back in the early sixties, but it proved to be unworkable at 300mph so there was no reason to suppose that it would work at 750mph. Recently, it has been suggested that computer controlled vectored thrust, essentially a movable tailpipe to direct the thrust from the engine could be used, but that idea is still a long way from reality.

The long narrow chassis section of Thrust SSC that runs between the engines suggested either a tandem rear wheel or offset rear wheel installation in order to minimise drag and to stay clear of the extremely hostile environment generated by the engines' exhaust. Tandem rear wheels, albeit fixed not steered, were another idea tried out in the sixties, this time by Bill and Bob Summers in a streamliner they built to attack the wheeldriven LSR. Although not fast enough to take the record it proved a point and the lessons learned were applied to their next car, the Goldenrod, which did succeeed in taking the wheeldriven LSR. It was a record that stood for many years until taken by another American, Al Teague, who used offset steered front wheels in order to reduce the width of his streamliner.

In a flash of lateral thinking one day, Glynn Bowsher asked why Thrust SSC shouldn't use steerable offset rear wheels. Why not indeed? After more discussion and detailed calculations it became clear that this could be the answer to the problem, but theory is one thing and practical application another. In order to remove any lingering doubts, including those in the mind of the man who would have to rely on this novel method of direction, it was decided that a test car would be built to prove the theory. And so was born a very strange looking Mini.

While the engine and most of the drive train remain standard, the front wheels have been locked up and an extended tubular chassis grafted on at the rear to carry two offset steered wheels in the format to be used on the record-breaker. Although it may look like the automotive equivalent of the Frankenstein experiments using a shopping car and a dragster as the unwilling donors, it does actually work. Lengthy runs at the Motor Industry Research Association (MIRA) test track not only proved the point to those remaining sceptical, but also gave Richard Noble valuable experience driving a rear wheel steer car. In fact, after a while he pronounced it more stable to drive that a standard Mini.

Back in the sixties, Craig Breedlove's failed attempts to use fixed wheels and a rudder were solved by making the single front wheel steerable through only one degree. With a proven rear wheel steering concept and a chassis installation that will allow six degrees of steering movement, Thrust SSC should prove every bit as controllable as it's illustrious predecessors. I wouldn't mind a go in that Mini myself!

As you suggest, that the real test is in operation

[Erfolg]

This link may interest some. Being a little serious and making a link, thiese devices possibly go faster than our dolly ever will.

As many things, the more you read, the less certain you become. Particularly when some of the Internet examples, wither deliberately misrepresent steering forces with driving forces, or are not quite as competent and authoritive as they present themselves.

Edited By Erfolg on 21/09/2015 19:11:08

[Erfolg]

These links are also interesting

Link 1 and Link 2.

A good part of the answer appears to be the amount of trail. The issues with bicycles is a little more complex than a trolley. It can be done, now the issues are identifying what matters and making it common knowledge.

I will post the completed trolly tomorrow. May need some mods before trialling.