Fullscreen HTML5Med Res (??MB)HTML4

A much better result with a b ELECTRIC FLI Our first article in the February 2006 issue showed the potential of the electric Piper Cub model aircraft. This month we continue our investigation by converting the Cub to a brushless motor to obtain a greatly improved result. By BOB YOUNG I n the last article we recounted the trials and tribulations associated with using the original Speed 400 brush motor supplied with the Cub kit. In passing, I mentioned the possibility of fitting a brushless motor to the Cub at a later date. This is the story of how a duckling was transformed into a swan. (For the politically correct, there is of course nothing wrong with being a duckling. Ducklings are nice too.) For those who missed the previous article, the brush Speed 400 motor supplied with the kit caused serious radio interference, especially with FM receivers, resulting in greatly reduced range. This problem was eventually overcome by replacing the FM radio with an AM radio (see box “AM & FM?”). Therefore there was much interest in examining the outcome regarding radio range when the brushless motor was fitted. In addition there were several minor flaws in the aerodynamics of the Cub as originally supplied in the kit. These included too little down-thrust and adverse yaw during aileron turns. Even with these drawbacks, the Cub flew very nicely in its original form and provided several hours of very happy test flying. (To think I get paid for doing this...) However it was obvious that the Cub could stand some improvement. Aerodynamic improvement Altering the down-thrust angle on the motor originally supplied would have been very difficult. Luckily the decision to fit the brushless motor solved that problem as well. The brushless motor chosen was a small LSE outrunner. The outrunner style motors have the windings (stator) at the 78  Silicon Chip centre, fixed to the motor mounting flange. The bell housing holding the magnets (rotor) surrounds the windings and the propeller is fastened to it. Thus the rotor and propeller spin together. Long-term audiophiles will recognise these brushless motors as being very similar in construction to the highquality Papst motors used in reel-to-reel tape recorders. Outrunners deliver more torque than the brushless inrunners but do not rev as fast. It is interesting to note that the outrunner with no gearbox drove the same size prop at approximately the same RPM as the brush motor fitted with a 3:1 gearbox. Mounting the outrunner required the removal of the original front former and the construction of a small box to move the motor/prop forward into the correct position for the re-fitting of the engine cowling (see photo opposite). The outrunner mounting flange was attached to a plywood plate that was mounted on this small box. Thus it was a simple matter to shape the box to tilt the plate at the correct angle for the down-thrust and right thrust required. As it turned out, I had set a fraction too much down-thrust and the angle had to be adjusted after the first test flight. With the correct down-thrust and right thrust angles, the aircraft will fly straight on full power and continue on flying straight when power is suddenly reduced. If the aircraft turns right, there is not enough right thrust and if it dives, there is not enough down-thrust. Conversely, the model needs less right thrust if it turns left and less downthrust if it climbs after reducing power. This is a most important part of setting up an aircraft correctly. A well set up aircraft is much easier to fly and places a lot less strain on a tyro R/C pilot. Learning to fly siliconchip.com.au brushless motor IGHT Part 2 an R/C model is difficult enough without the additional complication of a badly trimmed aircraft. The adverse yaw was cured in a relatively simple manner. A small amount of rudder movement was mixed into the aileron channel via the mixing function in the transmitter. Thus when the ailerons are moved, a smaller but proportional amount of movement was programmed into the rudder. At full aileron deflection approximately 10° of rudder is applied. This is sufficient to overcome the adverse yaw and now the model does well-balanced turns with no sign of adverse yaw. A switch was inserted into the mixer to switch the coupled aileron/rudder (CAR) mixing on or off. It is a good idea to switch the CAR off during take-offs and landings, as the CAR will induce a heading change each time the ailerons are moved to level the wings during the landing approach. This makes accurate landing approaches more difficult and that is not a good thing when learning. The pilot that can land the aircraft well and from even the most difficult of situations or emergencies will be rewarded with models that live a long time. This is why I practise landings and take-offs at every opportunity. Forget the loops and rolls. They are easy. The landing is the manoeuvre that separates the boy pilots from the men. Electronic improvements The difference in the electronic performance of the model with the brushless motor was staggering. I expected an improvement but not as much as was obtained. First of all, the radio interference completely disappeared and we could return to the original FM receivers without the range problems previously encountered. siliconchip.com.au June 2006  79 Secondly, the performance of the model was greatly improved. Where take-offs were marginal even on tar with the original Speed 400, take-offs are now so brisk that short, tufted grass is no longer a problem for the take-off area. The model will also climb more briskly and on a fully-charged battery, loop from level flight. This is a staggering improvement from a motor that is smaller, lighter and runs much cooler than the Speed 400. In addition, the speed controller supplied with the brushless motor has some very natty features although these had to be found the hard way as the original instruction sheet was a piece of paper approximately 75 x 25mm covered in Chinese or Korean writing! The first of these features is a programmable voltage cut-off to allow for the use of two or 3-cell LiPO batteries. As discussed in the first article, LiPO batteries are very fussy to work with. They need to be handled with great care in both charging and discharging. The point about discharging LiPOs is that from about 2.7V per cell downward, irreversible chemical changes begin to occur in the cells and by the time the cell voltage falls to 2.4V or lower, the cell is permanently damaged. It is very difficult to obtain agreement on a definite voltage at which the cell is destroyed, with figures quoted from 2.7V down to 2.4V. However what is almost universally agreed upon is that speed controllers should contain a voltage cut-off circuit that will cut off power to the motor at a minimum of 3V per cell. The LSE electronic speed controller (ESC) used in the Cub is fitted with a lead ending in a 2-pin header plug. If a micro-shunt is fitted to this plug, the cut-off voltage is set for two LiPO cells. If the shunt is removed, the ESC is configured for 3-cell operation. The nominal voltage of a LiPO cell is 3.7V per cell, giving a total of 7.4V and 11.1V for two and 3-cell packs. Be very careful with this because if you run a 3-cell pack on a 2-cell cut-off, terrible things are going to happen to that 3-cell pack. While we are on this point, it is mandatory that LiPO batteries be unplugged from the model at the end of each flying session. There is no ON/ OFF switch in this type of ESC, thus there will be a small current drain on the battery, eventually leading to cell voltages falling below 3V and thereby damaging or destroying the batteries if the battery is left plugged into the model for an extended time. For the same reason, do not leave LiPO cells lying around for extended periods without recharging, as selfdischarge will eventually destroy the battery, Voltage cut-off circuits will not protect against the last two scenarios, so make sure each battery pack is recharged at least once every three months to be on the safe side. The second feature of the ESC is a self-arming/calibration routine. The ESC will not operate until the throttle is moved to low throttle and the transmitter turned on. At this point the ESC is armed and will now operate the motor via the throttle channel. Another interesting point here is that the ESC will set the low throttle point when the transmitter is switched on. It is also possible to program in a dynamic brake during the switchon routine. The dynamic brake will prevent the prop from wind-milling in gliding flight. Electric models can be a dangerous to work around as the motor can start suddenly when least expected and low throttle arming is a great safety feature. Having a model leap off a bench, or worse still, inflict a nasty cut from a prop is no fun. So be very careful at all times when working with electric-powered models, especially with high-powered models! Unplug the motor battery whenever possible, keep the transmitter off as much as possible and if it must be on, put it somewhere where the throttle lever cannot be accidentally bumped to high. The motor will not start accidentally with the transmitter switched off. The scope grabs tell the story . . . This first ’scope grab shows the output of the FM detector on the bottom (magenta) trace, the supply rail is on the middle (cyan) trace while the active lead on the motor is shown on the top (yellow) trace. All ’scope grabs use the same order for the traces. This grab was obtained with the Speed 400 motor stopped. 80  Silicon Chip The motor signal with the Speed 400 motor running. Note the spikes on the supply rail. However, have a look at the receiver detector trace. It has been obliterated and this was with a moderately strong signal from the transmitter. Three brands of FM receivers were tried and all gave much the same result. siliconchip.com.au One of the things that I cannot get used to with electric models is the lack of a definite throttle-stop as I prefer to land with the motor set at an idle of approximately 1000 RPM. In IC motors the throttle barrel is pulled hard against a mechanical stop so that the idle RPM is constant at all times. In an electric model the idle point will vary a little with battery voltage, making landing approaches less predictable. If the throttle trim is pulled too far back, then the motor stops completely then restarts if the throttle is advanced. All this takes some getting used too after years of flying IC engines. I also miss the noise and have been toying with the idea of sticking a bit of cardboard into the prop like we used to do when we were kids on pushbikes. Either that or perhaps I will fit an onboard tape recorder with speakers and recorded motor sounds. The sight of a Piper Cub roaring off the runway with absolutely no sound is a little unnerving for those who cut their R/C teeth on IC motors. Still I do appreciate being able to go to a local cleared area and not disturb the neighbours. Another nice feature with the ESC is a motor cut-off that is instantaneous upon striking an object. The modern brushless ESC uses back-EMF sensing for its timing. These motors are wired as 3-phase motors in either star or delta configuration and therefore need to be timed to sustain rotation. The timing No, this isn’t the front of the motor or just a part of the motor. It is THE motor – it’s an LSE Outrunner and it gave a staggering improvement over the original electric motor supplied with the Piper Cub kit. Minor modifications were required to fit it, though . . . signal is derived from the motor windings and no longer requires extra wires for the timing signal. If for any reason the shaft stops rotating for even the briefest period, the motor drive current is cut off immediately. This is also a great safety feature. This type of sensing is extremely clever and calls for a simple explanation. By summing the two driven windings, a reference voltage can be derived. The undriven winding, which will have a voltage induced into it because it is being moved through a magnetic field, is compared to the reference. When a zero crossing is detected it is time to rotate the magnetic field to the next position. The above works once the motor is spinning. This shows the same order of traces but with a Silvertone AM receiver fitted and the Speed 400 stopped. Note the inverted detector signal and higher detector output. The next grab (at right) shows the Speed 400 running with the AM receiver and the same signal strength from the transmitter. siliconchip.com.au However before it starts spinning there is no zero crossing to detect so the designer must resort to interesting techniques such as ringing the windings to try to work out the position of the motor so a clean start may be performed. It is much easier to do this if the controller knows the motor characteristics. If they are not known, the controller must learn them. Some It was impossible to sync the ’scope due to noise but the detector signal is clearly visible and at full strength. However the audio filter must have been working its little capacitors to the bone, filtering out the rubbish being passed on from the detector. In spite of the noise the receiver still had full control of the ESC at 80% of its normal range. June 2006  81 controllers are so smart that they can recognise brush or brushless motors and configure themselves accordingly. Incidentally, brushless motors may be reversed by swapping any two of the three motor wires. If the ESC is disarmed after striking an object, the transmitter must be switched off and left off for 10 seconds or so and then turned back on, making sure that the throttle is correctly set at the desired low throttle setting. The most common cause of this type of disarming is clipping the prop in a bouncy landing. Finally, the last feature is the low voltage cut-off that will determine for you when flying stops on that battery pack. Once the pack hits 3V a cell, all temptation to continue flying is removed. That pack is now out of bounds until recharged. Do not be tempted to let the pack rest and pick up a surface charge and so continue flying. Go home and recharge the pack. With two 1800mAh LiPO packs and an 8 x 6 slow-fly prop on the Cub, I find I am gone from home for at least an hour and a half, sometimes more. This includes a five-minute walk each way to the field and the rest is taken up with non-stop take-offs and landings with the occasional loop and roll thrown in for good measure. I did say I bought the Cub because it was a pretty aircraft when taking off and landing! How is this possible? At best, the maths show 30 minutes flying time. This is one of the most complex questions in aerodynamics so the following answer is much simplified. Take a look at Table 1. This shows static (0km/h) current consumption against RPM for the 8 x 6 slow-fly prop. The table shows a non-linear rise in current for each 1000 RPM, with the last 200 RPM requiring as much current as the change from 3000 to 4000 RPM. Why is this so? The ESC appears to deliver a linear current with throttle stick position, so what is happening? The answer is found in the formula for aerodynamic drag: FD = 0.5 CD p A V2 The culprit is the V2 bit of the equation. As the prop speed increases the power required follows a square law and this is what makes selecting the correct prop for any motor/model combination so difficult and so important. The same applies to the model as well. To double the flying speed requires four times the power. Do you want or need a Piper Cub with a scale speed of 400km/h? Why not opt for a Cub with a scale speed of 140km/h and a battery life several orders of magnitude greater? Each branch of electric modelling is different and the prop requirements must be balanced for the task in hand against the design of the model. A slow-flying, high drag aircraft requires a larger diameter, fine pitch prop whereas a slick, fast-flying model requires a smaller diameter, coarse pitch prop. The choice of prop is one of the most difficult parts of aircraft Now have a look at these ’scope grabs. On the left, the brushless motor is stopped while on the grab at right it is running. The gain on the supply trace has been increased tenfold to get a look at the noise on the supply rail. This 82  Silicon Chip Table 1: static current consumption vs RPM for the 8 x 6 slow-fly propellor. design. Most modellers using 2-stroke IC engines tend to over prop their motors (too large a diameter mainly) and I suspect that electric flyers tend to do the same. For example, an increase of one inch in prop diameter can increase current consumption by 25%. However, electric flyers have a method of recognising prop efficiency in that flight times will indicate an efficient or inefficient prop/motor combination. Measuring flight time in the absence of in-flight telemetry is an important method of determining motor/prop efficiency. Propellers have many important characteristics including blade shape, blade section, pitch and diameter. The latter two are the most easily explained. Diameter is the length of the prop from tip to tip and will determine the mass flow through the prop. The pitch is how far the prop will pull the model in a single revolu- is processor switching noise and is barely visible. The FM receiver detector output is clean and stable with only processor noise visible and the audio filter easily filters this out. siliconchip.com.au tion. Broadly speaking, the prop is the aircraft transmission and a fine pitch gives great pulling power at low speed whereas a coarse pitch prop will give less pulling power. However there is a complication in that the faster a model flies the smaller the effective pitch angle becomes due to the angle that the air meets the prop being reduced as a result of forward motion. So once the aircraft gets moving the coarser prop becomes finer in effective pitch; take-off being the big problem for fixed-pitch propellers. Full-size aircraft usually use variable pitch props or constant speed props. Thus during take-off the pitch is set to fine (low gear) and at high speed the pitch is gradually increased to keep the engine RPM constant. Now this has important ramifications for electric flight models in that current consumption drops as the model gets up to flying speed. For example a 9 x 7 Masters e-prop that draws 11A at 5400 RPM at 0km/h will draw only 7.5A and the revs will increase to 6700 RPM at 60km/h. At 40km/h, the motor draws 8.5A and at 20km/h 9.5A. Thus referring back to Table 1, at anything less than 3,000 RPM in flight the motor in the Cub would be drawing fractions of an amp. And this is precisely how the model is flown when doing circuits and bumps. Full power is only applied during take-off and climb to altitude (about 100 ft). From then on the model is throttled back and cruises on approximately one-third throttle or less for the rest of the circuit. In conclusion then, where I began flying the outrunner with the recommended 9 and 10-inch props (props are still measured in inches) on the Cub, I finally settled on an 8 x 6 inch slow-fly prop, thereby reducing the static current from 11A to 7A and thus extending the flight time accordingly. The aircraft also flies at a more realistic speed – another nice touch. REAL VALUE AT $12.95 PLUS P & P Radio performance improvement A major, staggering, improvement was obtained upon fitting the brushless motor. It was impossible to obtain a reliable range with an FM receiver with the Speed 400 installed and we had to resort to installing a Silvertone AM receiver to obtain the desired range. The series of scope screen grabs tell the story. Field-testing told the same story. In hours of test flying there has not been a single glitch from the FM receiver, even with the model at extended range. So there you have part two of the electric flight story. I could say with warts and all again but there were no warts. Was the effort and expense of the change of motors worthwhile? Too right! SC AM and FM? For the technically-minded, scratching their heads and trying to make sense out of the statement that the AM receiver out-performed the FM receiver in a high noise environment, fear not! The laws of physics have not been rewritten. What the model trade refers to as FM is, in fact, Narrow Band Frequency Shift Keying (NBFSK) and R/C-type AM is ON-OFF Keying (OOK). NBFSK as applied to R/C equipment employs a ±1.5kHz frequency shift as against the 70kHz frequency shift in a true FM broadcast system. The signal-to-noise ratio of an NBFSK system when compared to an FM system is very poor. OOK on the other hand is not AM as the data transmitted is carried in the position of the OFF spikes (Pulse Position Modulation, PPM) and not the amplitude of the carrier. This means that the carrier is either full ON (100%) or OFF (0%) for very brief periods – about 350ms. The AGC time constant holds the AGC on during the 350ms OFF spikes. Thus the receiver AGC clamps the receiver in the least sensitive state until the last 20% of the receiver range, at which point the AGC is almost useless and the receiver is wide open to noise, as it is now at full sensitivity. This gives the OOK system a much improved signal-to-noise ratio compared to the true AM system. In practice, the signal-to-noise ratios of the two systems (OOK and NBFSK) are about equal. However, in many cases the AGC provides superior protection against electric motor and spark ignition noise than the limiters in the NBFSK receivers. This does not always hold true but experience has shown that swapping an OOK receiver for an NBFSK receiver will often give better results. siliconchip.com.au Silicon Chip Binders These binders will protect your copies of S ILICON CHIP. They feature heavy-board covers & are made from a dis­ tinctive 2-tone green vinyl. They hold 12 issues & will look great on your bookshelf. H 80mm internal width H SILICON CHIP logo printed in gold-coloured lettering on spine & cover H Buy five and get them postage free! Price: $A12.95 plus $A7 p&p per order. Available only in Aust. Silicon Chip Publications PO Box 139 Collaroy Beach 2097 Or fax (02) 9979 6503; or ring (02) 9979 5644 & quote your credit card number. Use this handy form Enclosed is my cheque/money order for $________ or please debit my  Bankcard   Visa    Mastercard Card No: _________________________________ Card Expiry Date ____/____ Signature ________________________ Name ____________________________ Address__________________________ __________________ P/code_______ June 2006  83