Andy Ellison has advice on getting that model set-up just right!

So you've flown your latest ARTF or scratch-built pride and joy for the very first time. How did it go? A few beeps of up trim and a couple of left? A click of the needle valve, maybe? Job's a good 'un! Are you happy to leave it at that, or would you like to try and get the best from the model with a little tweak here and there?

## Step 1:

It's true that many of us are very happy to leave well alone after the first flight of our latest toy, never quite getting the time to explore its true flight characteristics with a view to optimising its aerodynamic trim. Having lectured on this subject at various club nights over the years I also know it to be true that even the term 'Aerodynamic Optimisation' is well over the heads of many club fliers - increasingly so in this ARTF age, as we lose the basic skills that 'old school' aeromodelling used to provide.

You may think that such optimisation only really applies to aerobatic aircraft where ultimate precision is being sought, but this isn't the case. All R/C model aircraft can benefit from an optimised aerodynamic set-up. The objective is to make the flying experience even more enjoyable at the Tx end by reducing the pilot's battle against the airframe, and with computer radios now being the norm there aren't really any excuses for not doing at least a little bit of work in this area. Whilst it would be impossible to elaborate for every model type in this increasingly diverse hobby of ours, you can apply or adapt the following for your own requirements.

## Step 2: The Radio Re-visit

The first and most basic step following completion of a model's initial flights is to re-visit the radio installation. This is primarily to make mechanical adjustments, re-centring the output arms and optimising the torque ratings. Let me explain. When connected to a control surface, a servo generally translates its rotary action into the linear movement of a control linkage.

Consider a servo arm that's set square across the servo. Given that the movement of the arm is equal in both directions, the linear distance the linkage moves will also be the same. Now if you add some trim in flight, therefore moving the servo position from central to offset, there will be unequal linear movement to the linkage and hence the control surface (Fig. 1).

This 'differential' throw is sometimes very useful on an aileron, and you might choose to purposefully offset your servo arms in order to get it, but it's much less desirable on, say, a rudder. True, you can mix it out using the Adjustable Travel Volumes (ATVs) on your Tx, but the rate at which the control surface reaches a given point will then be different for both sides of the deflection, and the control surface response will 'feel' different in the air. It might take you a few flights to get the linkage lengths correct or the servo arms to an acceptable position but stick at it, as getting the basic steps right is very important to the rest of the set-up process.

## Step 3:

The second reason for re-visiting the installation is to optimise the useful torque of the servos. Servo torque is usually rated in 'kilograms per centimetre' (kg/cm), e.g. a 3kg/cm rated servo will provide 3kg of force to the control surface from a point 1 centimetre out along the servo arm. Attach your pushrod at 2cm from the centre of the hub and the servo can only apply 1.5kg of torque. Put an overly long, 3D-type arm on your servo at 4cm from the hub and you'll only get 0.75kg of torque - not much for driving a massive control surface (Fig. 2).

So, servo selection can be a key factor when building a model. For the purposes of this article I'll assume you already have the right servos for the job, so let's set about optimising their torque.

If you found the elevator to be a little twitchy for your liking during flight you probably either rated it down or reduced the throw using ATV. In order to maximise the torque available to deflect the control surface, the correct course of post-flight action would be to move the position of the mechanical linkage: either inwards on the servo or, if this isn't possible, outwards at the control surface horn. You would, of course, then have to reprogram your Tx to increase the rate or ATV (maybe to their maximum levels if the linkages don't bind up) in order to facilitate the same deflection at the control surface for 'best torque' from the servo.

Another thing to assess in this area is control centring and the 'feel' of the model. For example, does the elevator always come back to neutral after a 180º turn or reversal? Do the ailerons 'hunt' a little after a rolling sequence? If you're unsure then fly the model again, getting your mate to make notes if you think it will help. Poor centring of a control surface is due to problems either with the mechanics (binding servos, stiff linkages, etc.) or the electronics (bad servo resolution or 'dead band' in the radio system). Whatever the problems are, don't continue the set-up until they've been corrected.

## Step 4: Motor and Prop

The last step to look at before we get on to the 'heavy' set-up areas is to reassess the motor / prop combination. Selection of the correct propeller will fundamentally affect your model's performance. Many pilots want some level of 3D or 'Fun Fly' performance from their models, coupled with good airbraking and fast acceleration, which generally demands a prop of large diameter but shallow pitch. However, if you're after high top-end speed then you need a smaller diameter prop that's relatively high in pitch (for example, my British speed record-holding model used a carbon 8 x 10" prop turning at 20,000+rpm to achieve its 234mph).

Spend some time deliberating over the way in which your model performed on the initial flights, and if it suits you, then great! Don't overlook the ever-present noise issues, though. On most conventional models the main noise producer is actually the propeller, so if you fly from a noise-sensitive site you'll need to drop the tip speed somewhat. It's important to settle on the prop early, as any changes here will affect the thrust line steps that we'll look at later. When you're happy with your prop choice make sure it's well balanced and stick with it, buying spares of the same make and size. Before we start with the trimming proper, something to bear in mind: trimming is a bit like squeezing a balloon. Change a bit here and something somewhere else will be affected!

## Step 5: C of G

It's best to start with the C of G (Centre of Gravity), and indeed you might have had to do a little work here to get your new acquisition to fly half decently in the first place. Most kit instructions give a good indication as to where the balance point should be, but this position isn't sacred; it only represents the point where the prototype handled in the way the designer thought it should.

Testing the C of G is fairly simple, its effectiveness being demonstrated best on models with fully symmetrical wing sections set at low incidence angles. With the model at height, observe a flight path where you're viewing from the side and flying into wind. Slow the engine to idle (or motor off if you're flying electric) and push the model into a vertical dive. Release the elevator stick and observe the descent of the model. If it begins to pull out of the dive as if up elevator was applied then this indicates a forward C of G. Conversely, a tuck-under as if down elevator was applied indicates a rearward C of G. Many F3A aerobatic pilots set the C of G slightly rearwards to help the model fly 'hands off' down line. 3D models also usually exploit a very rearward C of G, and it's not unusual to have to apply slight up elevator on these to hold the dive straight (Fig. 3).

When applying this test to a glider, observe the model in a long, steep descent of say 45 - 60º rather than a vertical dive. A good proving test for a glider is to fly inverted. Hands-off inverted is often desirable for a racing set-up and is often referred to as a 'pitch neutral' C of G position. Some pilots prefer to prove the C of G position in a vertical climb rather than a dive, observing if the model falls backwards or tucks under as it begins to slow. Personally I find this method unreliable as the motor's running flat out here, and any thrust line inaccuracies will create a similar effect. We'll come to those later. In the meantime let's look at the other balance issue.

## Step 6: Lateral Balance

This is an aspect of trimming that's often overlooked. I've read of many 'old school' modellers adding weight to wingtips after a session of balancing the wing on the workbench or hanging the model from strings attached to the roof. Whilst the latter method may get you somewhere near, the former is totally wrong; the model must be balanced as a whole assembly, and the best way to prove it, is in flight.

To determine if your model has a heavy side, fly it into wind straight and level and directly away from you. At a reasonable throttle setting pull tight consecutive inside loops without using any other control except elevator, and leave the throttle alone. Watch carefully to see whether the model begins to track off to one side, as if it had a heavier wing (it will track towards the heavier wing). Note the direction carefully, then bring the model around the circuit to the same position, throttle setting and heading as before. This time roll to inverted and push down elevator to fly tight, consecutive outside loops. Again note the direction of any drift. If the model has a heavy side it will now track in the opposite direction to before. A drift in the same direction is indicative of a warped wing rather than a heavy one (Fig. 4).

Obviously the lateral balance of models with side-mounted engines is going to be off from the start, so you might expect this side to be heavy. Lateral balance can be corrected by adding small weights to the lighter wing, as near the tip as possible, until the model tracks straight both upright and inverted. On moulded glider wings the weight can be added to the outboard aileron servo bay or under a gap seal near the tip. If you're unlucky enough to have a warped wing and decide to stick with it, you might consider lessening the effect by adding a trim tab on the underside made from trailing edge stock, or experimenting with the mixing on your Tx. Mind you, this is really a compromise and you'll probably never get it perfect.

## Step 7: Thrust Line

It's quite common these days (especially on high-end ARTFs) to find the front bulkhead factory-fixed with an offset for motor side-thrust and down-thrust. Whilst these might seem like they're set with intent, any variations in motor or prop size relegates them to the position of 'best starting point'.

The spiralling slipstream from a propeller hits the model's fin at an angle, and the yaw force this generates is dependant on the motor rpm and forward speed of the aircraft. To a small extent this effect can be mixed out with rudder trim but this will constantly vary as the rpm of the motor changes.

## Step 8: Sidethrust

This is perhaps the easier of the two to determine, but be aware that anomalies with lateral balance and any fin / rudder misalignment can create similar effects. Fly the model at full throttle, straight and level into wind and then pull vertical at the end of the circuit, as if for a stall turn; it's important to establish the up line quickly without the use of rudder inputs. Be wary of pulling too sharply as any lateral balance inaccuracies will immediately drop the heavy wing and throw the model off the vertical. The model should hold steady for a short while before yawing off to one side. If it yaws to the right then decrease the right thrust, and if it yaws to the left, increase the right thrust. Aim to have the model hold a vertical up line for at least four seconds before a slight yaw to the left begins (Fig. 5). It may take you a few flights to get this and the lateral balance nearer to the mark, but patience will pay dividends so keep at it.

## Step 9: Down-thrust

Again, fly the model at full throttle, straight and level into wind. Abruptly shut the throttle to idle and observe the flight path. If she climbs then the motor is pulling the nose down when under power and you'll need to reduce down-thrust. The correct thrust line should see the model continuing straight and level before slowly starting to sink as speed decays. A more abrupt dive is a result of not enough down-thrust and you should alter the motor angle accordingly (Fig. 6). High lift wings (as used on electric-powered gliders) result in a motor installation that has large amounts of down-thrust, whilst side-thrust may not be such an issue. Models with the motor mounted on a pod above the fuselage will need large amounts of up-thrust to work effectively.

If you've a cowling to cut then make sure your thrust lines have been established before getting the knife out, note also that most F3A aerobatic engine mounts are now fully adjustable with side- and down-thrust trimming in mind.

## Step 10:

This is the last area where big gains can be made towards perfecting a model's performance. It's really aimed at obtaining purity of yaw axis and rudder control, i.e. when applying rudder the yaw is optimised and any associated pitching or rolling is minimal. Whilst this really only applies to aerobatic models, an insight into what's going on can help us understand some aspects of model behaviour. In order to obtain pure rudder input a low-wing model will generally need dihedral (wingtips elevated above the root), a high-wing model will need anhedral (wingtips below the root) and a mid-wing model might need none at all. Let's look at this in practice:

An F3A pattern ship is designed to track straight, roll pure, and knife-edge forever. Its low-wing design incorporates slight dihedral to achieve purity of the rudder control whilst retaining some stability.
A very aerobatic high-wing design such as the Wot 4 usually has a flat wing that's mounted way above the datum line. If rudder is applied in flight then yaw is induced, but there'll also be a large pitch downwards and a rolling motion in the direction of the rudder. To eliminate this mechanically the Wot 4 would need an anhedral wing for purity of roll control, but then the model would lose the stability that's made it a classic design.
Scale aerobatic models like the Extra 300 generally have a mid-wing configuration with little or no dihedral, as they're designed with control purity and some level of instability from the off.

## Step 11:

The test for correct dihedral is made with the model in knife-edge flight. With the aeroplane flying at a decent speed, roll to knife-edge and apply top rudder. If you rolled 'left wing low' with right top rudder and the model slowly starts to roll in that direction (i.e. with the rudder input) then the dihedral angle of the wing is too great. Likewise if the model rolls against the rudder then the dihedral angle is insufficient (Fig. 7).

You'll need to try this on both sides to be really sure, and the remedy is to cut the wings in half and re-join! Once upon a time, maybe, but not so these days. Most modern transmitters with freely programmable mixers will have a dedicated knife-edge mix that applies a little aileron as a slave channel to the rudder input. Experimentation with rudder rate switches for straight and level knife-edge at full throw may result in slightly different figures in the mix for both sides. Models like the Wot 4 will need a great deal of aileron compensation for a given rudder input, whereas a low-winged aerobat might need only 2 - 3% aileron offset in the mix. An additional free mixer will deal with any pitching induced by the rudder.

Most models will pull towards the undercarriage in knife-edge due to the increased drag it produces about the datum, and of course you'll need to set this up for both sides; but the result could easily be effortless knife-edge flight from horizon to horizon with a model that you'd otherwise struggle with.

## Step 12: Fit and Ready

So there you have the basic areas where the biggest improvements can be made. Other aspects can be explored if you really start getting into this kind of model set-up, e.g. equalising elevator throws on split surfaces or raising / lowering ailerons to provide correct tracking - these are areas where smaller gains can be achieved. The level to which you choose to take this is entirely dependant on both your model and your ability, plus of course the way in which you fly. Don't forget to revisit some of the earlier steps as you go through the process to see how the latter changes affect the earlier ones. Go ahead, experiment. It's a lot easier than you might think!