Bell 429 by ROBAN/Scaleflying.de
Review:
How it came about
This page is about my sample of the ROBAN Bell 429 and how I fared with it. That may be helpful for some people, just as it's always helpful for me to read other people's reviews. Some general information about this model is given here as well, some information can be found in the Web (see "More"), but most of the following is what I would have liked to read before I assembled this helicopter.
In 2020, I had accumulated six quality flying hours with my HIROBO Schweizer 300. When we had the next Covid-19 lockdown in the following winter, I wanted to own an even bigger model helicopter – probably just craving. The heli should have at least four main rotor blades and again be "scale", so not for aerobatics – sort of appropriate to my age (now even more important).
A model of the Hughes 500 still would have been to my liking, but its body is even bigger than that of the Schweizer 300. That would not only look massive; together with the especially high landing gear it would be "tall" and would hardly go in my car (at least I thought so back then, but I was wrong).
So it had to be something more "slim" – like the Bell 429. I knew the Air Zermatt helicopters from TV and YouTube and liked them a lot. A clubmate who owns the 700-size Tiger model by ROBAN/Scaleflying.de hinted at the model. He strongly recommended the 700-size B 429 instead of the 600-size (and would have recommended an 800-size model if there would have been one).
When the package lay on my carpet I had yet a scare about what I had dared to do. But once the helicopter is ready-built I will have gotten used to its size. Moreover, it will look not that massive out at the flying field, as I know from the 700-size Tiger and an 800-size Airwolf – and bigger flies better, no doubt about that.
The model's size is specified as 700 but the new main rotor blades are 720 mm long – makes 1.60 m main rotor diameter (1.56 m by the instructions). The mechanical assembly is specified as 800-size because it is made for this size (and this is a 4-blader) and a 750-size motor with 450 kv is mentioned in the instructions. A 700-size motor (of course with a 12s LiPo battery) just fits by diameter and is powerful enough according to Scaleflying.de (this is a scale heli), as well as the recommended 1100 rpm main rotor speed is fast enough. Weight is specified as 8 or 9 kg, depending on where you look for specifications. Expectedly, 9 kg turned out reasonable and 8 kg to be absolute lower limit. The dimensions in the instructions are correct, the model is scaled 1:7 – slightly smaller than the HIROBO Schweizer 300 (1:6.32).
At YouTube I found more than 40 videos of the ROBAN Bell 429 (see "More"). There, mostly a bigger motor (750 or even 800) with higher kv is used (if mentioned at all) and higher main rotor speed. To me that seemed excessive and I rather trusted the recommendations by Scaleflying.de, which have been even confirmed in a reply to my e-mail inquiry (thanks!). Just to turn trust into certainty, the model's drive (see below) and rotors have been rendered in a flight simulator – and yes, the recommendations are correct.
I was a bit worried about my lack of experience. I know how things work but have no hands-on experience. The HIROBO Schweizer 300 was good practice for me, but its whole build was perfectly prepared and described. Before buying the B 429 model, I had downloaded the instructions from the Scaleflying.de website. They seemed to be rather short and a bit confusing, and that proved to be true when the kit had been delivered. Three additional sheets added important information not yet included in the instructions (version 2.0 – May 2016) and that was it.
Scaleflying.de had four useful videos at their YouTube channel which showed important aspects of their 800-size mechanical assembly. They pointed out that its main part is pre-assembled but has to be dismantled just to re-assemble the whole thing using threadlocker. Besides, the bevel gears' backlash, or mesh has to be checked and, if needed, adjusted with shims included in the kit, as well as – in this new version – the bearing clearance of the angular ball bearing on the main rotor shaft.
I was annoyed when I heared that I should use heat to unloose bolts if they are stuck by threadlocker. My clubmate thinks that pre-assembling saves packaging the many parts and describing their assembly. That makes sense to me, but why did they use threadlocker then and – as it turned out later – shims to adjust the bevel gear as well as the angular ball bearing, and why did they grease them? To me it rather looks like an exclusion of liability and that hampered me at first.
Assembly
The helicopter is sold for a really reasonable price but still it costs a lot of money. That's why I declared it a Christmas gift and didn't unpack it until Christmas Day. The remaining time was used to purchase the necessary components (altogether costing a quarter more than the kit): motor, ESC (still without capacitors) with USB adapter, safety switch (still without resistor) with receiver adapter, servos, gyro (flybarless) system, receiver, telemetry (GPS, two voltage sensors, data logger), dual receiver-battery switch, cables and connectors – but no batteries yet. At Christmas Day, the kit was unboxed and – first of all – the mechanical parts were inspected.
Mechanical Assembly
From right to left: mechanical assembly, wrapped parts of the motor shaft support bearing, main rotor head, tail rotor assembly with bevel gear, tail boom with tail rotor shaft (wrapped), main and tail rotor blades – wonderful!
I was especially glad to have the latest, improved version: motor shaft support bearing, helically toothed main gear, angular ball bearing on the main rotor shaft, longer servo-swashplate levers, hence longer servo arms, blade pitch linkage without swashplate driver, longer rotor blades, reflexed main rotor blade airfoil.
But should I really dismantle that? Enjoyment gave way to annoyance. All too often I didn't want to concern myself with the helicopter. At times when I was more sober I watched the instructional videos by Scaleflying.de, read about gear assembly, and inspected the mechanical assembly and the rotor heads again and again.
Eventually, there are nine months between this picture and the next one. In the summer, things started to happen really fast. The rotor heads were checked and turned out to be securely bolted (with threadlocker). The mechanical assembly was partially dismantled in order to reverse the bevel gear, then reassembled with threadlocker.
The result is really to my liking, even if I'm still not confident – about the screwlock, the bevel gear backlash, the bearing clearance. By the way, the tail boom is really a bit inclined upwards; that's how it runs in the tail cone. The ROBAN 800 mechanical assembly (called SM2.0 Mechanics) is basically the same in all of their models, but the carbon frame sides are different (as well as some shafts). In the B 429 model, the tail boom is just inclined and quite high in the body. And there is no 45° bevel gear in the tail boom like in the Tiger model.
The two carbon tail-boom struts are bolted and clamped with aluminum fittings. In one of the videos by Scaleflying.de they mention that these struts are affixed differently, depending on the mechanical-assembly frame in the respective model. They have to be bolted to the carbon frame sides where washers are under the bolt heads. That's how it was and now the whole assembly goes through the rear opening in the body, so all seems to be correct.
The motor moves a cogged belt which is visible outside the carbon frame side. Above the big pulley is a steel pinion which drives the big white plastic cogwheel visible here, and thereby the main rotor shaft.
The three 120°-swashplate servos are linked by one 90° bell crank each. In this latest version of the mechanical assembly, the bell cranks are bigger so the servo arms are longer than before, and all three pushrods have the same length now.
The KONTRONIK Pyro 700-45 (kv 450) has been chosen due to the positive experience with the Pyro 600-09 (kv 930) in my HIROBO Schweizer 300. It has about half as big a kv and a 12s LiPo is used instead of a 6s.
A more powerful motor is not needed according to Scaleflying.de (and I think so too now, see below). If I would like to use one, though, its diameter should be not much bigger so it would still fit in the frame's recess.
When mounting the motor, one has to decide how the very stiff connecting wires (not stranded) run forward: slanting to the left or to the right. That depends on how the ESC will be mounted in front of it (see below).
The motor sits on a carriage which is pulled forward (left in the picture) by two bolts to tension the cogged belt.
The two tensioning bolts are visible here as well as the whole cogged belt gear train. When the belt is tensioned correctly, the four bolts clamping the motor are tightened. I used shiny bolts and washers (well visible here) while (quite short) black ones were included in the kit.
The T-shaped part is the motor shaft support bearing. It has three "legs" which are clamped – together with motor and carriage – by three of the four clamping bolts. That's quite a fiddling but doable. I just hope that I didn't cause any tension on the motor shaft when tightening the bolts.
Only in this case the support bearing prevents permanent cantilever load on the motor shaft so it's a very good thing. A precondition is choosing the motor version with the longest shaft journal available – 38.5 mm (6 mm diameter). And the cogged-belt pinion has to be mounted with its set screws towards the motor – not like shown in the standard instructions but like shown in the special instructions for the support bearing and in the video.
By the way, I had hoped that the cogged belt would run centered between the pinion's lateral disks. However, the test run showed that it follows gravity and runs against the lower disk. I had to loosen the pinion again and shift it one millimeter upwards (towards the motor) so during operation the cogged belt runs centered on the big pulley. I deem that important because running on the pulley's edge would destroy the cogged belt rather sooner than later. I needed an Allen wrench with ball head which can work aslant (not in line) to the set screws. And I had to make sure that the set screws ran straight into the threads and not askew.
The cogged-belt gear train's numbers of theeth are 22 and 78 – 3.55:1 reduction ratio. The following gear train with helically toothed steel pinion and plastic cogwheel has 20 and 78 teeth, respectively – 3.9:1 reduction ratio. Hence the total reduction ratio is 13.83:1 from motor to main rotor. Assuming the recommended 1100 rpm main rotor speed, the motor has to spin at 15210 rpm – no problem considering 30000 rpm maximum rotational speed (and 12s LiPo).
Below the big cogged pulley is the first (white plastic) gear of the tail rotor gear train. The second behind it moves the vertical shaft of the bevel gear below it. The latter's gear ratio – as well as the bevel gear's on the tail rotor – is 1:1.
The two plastic gears' numbers of teeth are 40 and 32, what makes for a 1:1.25 transmission ratio. Together with the cogged-belt gear train's 3.55:1 reduction ratio, the total reduction ratio is 2.84:1 from motor to tail rotor. At 1100 rpm main rotor speed, the tail rotor spins at 5363 rpm (1:4.9). With its 290 mm diameter (280 mm by the instructions) and its four blades it should have quite some power (and the simulator confirmed that already).
The bevel gear has been "reversed", that is the bevel wheel on the vertical shaft has been relocated from top to bottom, to reverse the tail rotor's sense of rotation. Now it turns the right way round, like on the original helicopter and like it should be "aerodynamically".
This bevel gear still worries me even though all looks good and works correctly. An instruction sheet was included in the kit, which points out that the bevel gear has to be adjusted, and there the bevel gear is shown like here. Probably the picture shows another helicopter model's bevel gear, though, perhaps with main rotor turning clockwise. Anyway, the shaft's flat portion for the wheel's set screw was not symmetrical and had to be extended (with a grinding wheel on a Dremel). So the tail rotor's wrong sense of rotation must have been intended by the designer. I asked Scaleflying.de by e-mail whether there is a special reason for that or any reason to abstain from reversing the gear, but they just didn't respond. My only clue was an exploded view:
It shows the gear's components (spare parts?) very well and in the position where I have them now. The note below reads: "To ensure proper tail rotor rotation and functionality, the input bevel gear and sleeve might need to be swapped around." But does that really mean what I understand? Anyway…
Reversing the bevel gear was an opportunity to inspect all details. To this end, I removed the whole gear from the carbon frame, what was quite elaborate. I could have done it easier by removing the lower bearing plate and unloosing the wheel set screws. The shaft could be pulled out then as well as both wheels and the brass spacer sleeve. Originally, the bevel wheel was on top and the spacer sleeve below. Now it's just the other way round and all works well like before.
Above the bevel wheel were two shims to adjust the backlash. Of course, the shims were put under the wheel when it was reversed. As far as I can see, the gear was adjusted correctly then. I just can't rely on that because Scaleflying.de would not be liable for any damage. That's not a theoretical problem – my clubmate lost his Tiger model twice in crashes because the 45° bevel gear in the tail boom failed. Of course it was his fault, meaning he got no compensation. (Now he thinks about installing a flexible shaft instead, or the new 45° bevel gear by ROBAN which seems to be much better.)
For three reasons I believe (or hope) that my B 429 won't crash due to gear failure: (1) The 90° bevel gears seem to be not prone to fail or in any way critical, after all Helikopter Baumann makes 45° bevel gears with three 90° bevel wheels – here. (2) When inspecting the bevel gear I realized that both bevel wheels have to be adjusted so that their teeth are not offset to each other. (The teeth of the failed 45° bevel gears are stripped, but on one of the wheels only partially, that is not their outer part.) If that's not correctly done, also the backlash can't be correctly adjusted at all. (3) The 90° bevel gears have rigid open chassis so correct adjustment is visible. The 45° bevel gears have somewhat sloppy split closed housings so that the final adjustment is not visible and perhaps a correct adjustment is not even really likely.
ROBAN should really do like HIROBO: By all means, such a gear like the 45° bevel gear has to be prefabricated – assembled and adjusted. Perhaps they could design it in a way that no adjustment is needed, or at least they could use a fixture to easily and reliably adjust the gears – in the factory. They ought to accept responsibility, that is assume liability, but they won't possibly get that right.
However, I still don't know how to adjust the gears correctly and can only guess. Mr. Illig from Scaleflying.de showed in his video how to tension the cogged belt. But he just poked around in the bevel gear with a screwdriver (there is no help for it, there's nothing to be seen) and told that the bevel gear has to be adjusted – but not how. As far as I can see, the bevel gear was adjusted correctly, that is I couldn't do it any better.
That applied to the bevel gear on the tail rotor as well. It's made from the same parts as the front bevel gear; the bearing plates are just closer to each other. It's easy to check that the two bevel wheels are correctly positioned to each other. And the correct backlash can be checked by turning the shaft to and fro. The wheels were greased already.
The tail rotor has been converted to the reversed sense of rotation: the pitch lever links removed from the link balls, the blade holders reversed so the pitch levers "lead", the link balls unscrewed and screwed in from the other side (with threadlock of course), the pitch lever links snapped on again. In this way the right-angled bellcrank lever is again – geometrically advantageous – in normal position parallel to the tail boom and the tail rotor shaft, respectively. As shown here, that means some degrees positive pitch to cancel the main rotor's normal torque and equal throw from here to both sides.
Now the tail rotor turns clockwise (as seen like here in the picture), the tail rotor blades moving backwards at the top, with the main rotor's wake and not against it. That's how it is on the original helicopter and (almost always) aerodynamically correct.
The tail servo (XServo X75 BLHV) is clamped to the tail boom – a simple and neat solution. In the instructions and in pictures I saw at least four different ways to mount the clamps, the plate and the servo and how straight or oblique the whole thing is clamped to the boom. This was one of these ways and it seemed to me the way actually meant by the designer (but I may be wrong).
The tail boom is made from very thin aluminum and its front end was a bit crimped in, probably during transport. With some force, it had to be stuck and turned into the receptacle on the front bevel gear. Fortunately, I didn't crush anything in the process.
Also the swashplate servos (XServo X70 BLHV) are well linked. They are mounted in a special frame so that all three pushrods to the bellcranks have the same length. The length specification (96 mm) is correct, just the middle pushrod has to be a bit shorter because it's less oblique.
At YouTube, some people claim that the middle pushrod has to be offset (bent) a bit to avoid a clash with the middle bellcrank. As Mr. Illig stressed in one of his videos, no clash can happen in practice. I can endorse that there is no need to bend the pushrod.
By the way, the middle link to the swashplate is triangularly shaped so it can hold the swashplate's lower part against turning. This way no special swashplate holder is needed.
The ball links are made from a quite hard plastic and (hence?) have a little bit of slackness. That's strange – so far I had only links made from ductile plastic and had to press them with flat pliers so they are not too tight.
There's an own place for the gyro (flybarless) system so it's close to the servos and exactly parallel to the main rotor disk. The servo leads are protected from rubbing on the sharp edges of the carbon frame by wrapping them in slashed pieces of fuel tube.
The Microbeast Plus HD has been chosen due to the positive experience with the Microbeast Plus in my HIROBO Schweizer 300. This HD version just has an additional MPX socket for high-amperage power supply, advisable for the heavy-duty servos used here, which are in turn needed for the modern swashplate linkage. The four high-voltage servos draw about 5 A on average, much more than those in the HIROBO Schweizer 300. There is even a switch for the Microbeast but I don't use it because I anyway have (rather than a BEC) the dual switch mixer (see below) for two parallel batteries, which I have for safety's sake.
There was some bother about the servo arms: The instructions recommend metal (aluminum) arms. An additional instruction sheet specifies at least 13 mm servo lever arm for the new-version swashplate linkage. Later this proved to be a minimum requirement indeed because the Microbeast wants to know how much servo travel is needed for exactly 6° pitch lest it overreacts. With 13 mm servo lever arm, the maximum adjustable travel resulted in just 5.6° – the Microbeast is fine with this value but the lever arm could be a bit longer, at least 14 mm long.
Aluminum servo arms are hard to get in any case, especially 13 mm lever arm, and I found only one offered in the Web. Unfortunately, it doesn't fit the servo shaft (not properly deburred, swollen when anodized, but also servo shaft too big!) and has a 3 mm threaded hole instead of 2 mm. In a video by Scaleflying.de, Mr. Illig said that metal servo arms are hard to get but using plastic arms is OK. So the supplied plastic arms were put on the servos – first those with 13 mm lever arm (picture above).
The mechanical assembly is an operative unit. Earlier versions even had a "porch" for the ESC with two more mounting lugs, but obviously that has been abandoned. Now four lugs must do and the ESC is screwed to the top of the cabin ceiling in front of the mechanical assembly.
After it's completion, the mechanical assembly has been test-run with the 12s drive batteries. The ESC was set up as a governor and the set rotor speed has been defined. After a short break-in of gears and bearings all ran smoothly and with low friction. Amperage – as indicated by the ESC's telemetry – is a measure of friction moment and suggests about 94% overall (main and tail rotor) gear efficiency.
Corrections
For several reasons, further construction had been put on hold for a long time. Then I got a new transmitter (PowerBox ATOM), the Microbeast got a useful update and all the settings were checked. At the same time, the pitch angles of the four main rotor blades (correctly: blade holders) were adjusted. Because the ball joints can only be adjusted by half a turn, a difference of 0.2° maximum remained.
The ESC got a useful update, too, and slightly changed settings. During the subsequent test run, the set rotational speeds were readjusted. At some speeds, vibrations were seen in the tail rotor drive. This looked like a ball bearing slipped on the shaft, but could also be due to a large amount of play in the drive (in the two couplings on the shaft) and would disappear once the blades were fitted.
In the meantime, a friendly helicopter pilot had informed me by e-mail that the coupling pieces of the tail rotor shaft are screwed onto the shaft with a left-hand thread. (Thanks Roberto! I had hoped to get this information from Scaleflying.de, but they just didn't respond.) Hence the parts can come loose if the shaft runs clockwise, as in my case, because I reversed the front bevel gear. So what should I do?
First, the rear bevel gear was inspected. It would stand to reason to move the driven gear wheel on the tail rotor shaft from the right to the left to reverse the tail rotor's direction of rotation. Then I could un-reverse the front bevel gear and the shaft would run anti-clockwise. Unfortunately, the rear bevel gear is not symmetrical so it cannot be reversed. Nevertheless, it was disassembled, cleaned, oiled and greased, and reassembled. Now it runs with less friction than before, as does the front one. By the way, there were (and are) two shims under the driven gear wheel so obviously the gear was already adjusted.
Then I found a discussion about the ROBAN tail rotor gearboxes in a German forum (August 2020, here). Someone wrote there that he secured the coupling pieces on the shaft by drilling through the threads and installing 2mm screws. (That's what the ROBAN distributor Baumann recommends – here). My thought: Simple dowels could be enough, i.e. drilling 2 mm holes and glueing 8 mm long pieces of steel wire with Loctite glue. (That's what I did on my HIROBO Schweizer 300 – here.) But then the holes became so wide that the wire didn't stick at all. So I used screws with nuts only loosely tightened and secured with Loctite.
By the way, drilling the 2mm holes was not easy. A dab with a center punch has to be made exactly at the tube's crest, and the drill bit has to go in exactly vertical. Clamping in a machine vise seems necessary to achieve this. The thread of the coupling piece is hollow, as I noted when drilling. The inner radius of the hollow is so small that the drill bit's tip can't reach its bottom. Instead, the drill bit's edges cut at the sides. That lets it float around so the hole gets uneven and too big. It would have been better to drill both holes separatly from outside. That means to clamp and punch exactly opposite (turned by 180°) and to drill only halfway into the hollow. However, this matter is obviously not critical.
Pulling the shaft out of the tail boom was surprisingly easy. It turned out that both ball bearings are firmly glued to the shaft, but the front rubber piece had partly slipped off the ball bearing - apparently due to too much friction during installation. So it wasn't correctly guided and that's why there were vibrations. Unfortunately, the front ball bearing is now a bit stiff, perhaps damaged when it was pushed in, or by the lubricant, or when running wrongly loaded by the half-slipped rubber piece. The bearing could be loosened with a bit of heat, but how could I get a new one? The bearing may run in again and for now it just has to work.
Pushing the shaft into the tail boom was easy as well once I knew the problem and found a trick: the rubber pieces have two "lips" which are greased with dish soap. I had done that the first time but then pushed the shaft all the way in one go. The grease hadn't made it to the foremost position hence now I pushed the shaft only until I felt some drag, then pulled it somewhat out until I felt the grease again, and back and forth until it was finally all the way in.
So the coupling pieces were secured and the shaft reinstalled (correctly). The front bevel gear remains reversed, the tail rotor rotates the right way, no more vibrations (test run) – and hopefully no loosening of any threads (not even the pitch sleeve's).
After the mechanical assembly and the tail had been mounted on the body, the swashplate control setup was reexamined. Even at extreme throws of collective and cyclic pitch, no collisions occured. But only now it occured to me that one or two of the swashplate servos had extreme angular throw, more than 60° off neutral. So the 13 mm servo lever arms are rather too short and longer ones would be at least better.
Quasi "at the last moment" they were replaced by 17 mm arms (30% longer). Now some of the arms collided so those not needed had to be cut (right and middle servo).
Because the linkages are more aslant now they had to be elongated by half a turn of the ball link threads each to have the swashplate horizontal again.
In the gyro system (Microbeast), the swashplate throws had to be reduced markedly, but that was the whole point of this measure since it makes the throws about linear again. The downside is that the servos draw more amperage now since they have to apply more torque. 15 mm arms would be better in this regard. Anyway, the Microbeast now gets its full 6° pitch at smaller angular servo throw and hence can react appropriately to attitude deviations. The tail rotor servo still has the 13 mm arm because there is no excessive angular throw here.
Body
The struts of the landing gear halves are stuck through holes in the fiberglass body into a plywood structure inside, then fastend by stick-through bolts. Actually, that's a good solution but the right rear strut went not deep enough into the hole so finally some pieces of epoxi chipped off – fortunately only below the strut where it's not visible.
Just because they are well visible here: door hinges and handles are not really prototypical but not too bad either. They are a bit coarse but stand out just as little as the badly glued-on footsteps.
My problem is visible here: The footstep's midlle attachment is glued-on flush but the other two at the front and the back stick out a bit. The footstep is so rigid that rubber bands, adhesive tape, and later the glue slackened. Now some Canopy Glue will somewhat fill the gaps.
This is a bottom view of the gaps; later they will stand out very little, see previous picture.
The doors are flush-mounted and look pretty good. The interior equipment, mostly seats, has to be glued to the beige-colored plywood cabin floor – that's all. Looks pretty good as well.
There are tabs on the bottom of the seat pedestals, which are stuck into slots in the floor (hardly visible in the picture). So it's perfectly clear where the seats go in the cabin; they can be glued on only there.
A whole row of seats should be glued in one go, that is the two pedestals per seat into the slots and the seats onto their respective pedestals. Glueing the pedestals to the seats first won't work because all is too inexact for that. Canopy Glue allows enough time to align the row of seats.
The box in the middle has a fuction: It's a handle on a slider which locks and unlocks the hatch in the cabin floor in front of it. The box is grabbed to move the slider back and forth. Works great!
Three seats are on the cabin floor hatch, and they are used as a handle to remove the hatch and get at the battery compartment.
In the middle is room for two 6s 5000mAh LiPo batteries and on the sides for two receiver batteries. Behind them (left in the picture) is room for other things to be placed here (master switch for the drive batteries, dual switch mixer for the receiver batteries, telemetry sensors) as well as cables and connectors.
The cables to the ESC on top of the cabin ceiling can be threaded through the vertical plywood duct visible in the background (behind the corner of the hatch opening). In front of the battery compartment (right in the picture) are the pilots' seats, which are seen from above in the next picture.
This picture is about sticks, though, that is the cyclic and collective control sticks.
In the full-size helicopter, the pilot is in the right seat (left in the picture) and his collective stick is on his left, not on his right like here in the model helicopter. The model's seat would have to be more to the right side and the collective stick between seat and the center console (but it's not really easy to glue them that way due to the tabs and slots). There is even another, smaller console on the center console's right side (below the panel, not visible here) so a pilot manikin's left leg or knee, respectively, would have not enough room.
Seems the model's designer missed something, but then again he went to great effort: The switch box on top of a cyclic stick is asymmetric and here both sticks are inversely asymmetric because the pilot uses his left hand and the copilot his right hand in this arrangement. Actually, both cyclic sticks should be equal because both pilots have them in their right hands (and the collective sticks in their left).
Anyway, I should have changed the right pilot seat's arrangement. It would have been possible to drill new holes for the two sticks. The tabs on the seat's pedestals could have been cut off and the seat just glued flat on the cabin floor – in the right place. And I could have lowered the seating surface so that a pilot manikin's hands reach to the sticks (to their grips) and its helmet doesn't push against the cabin ceiling (see third picture below).
The tail has to be bolted to the body with six 2mm bolts, a foam piece on the tail boom carrying the tail's weight. At least I don't think the small bolts are capable to do that on that long lever arm. The holes can or have to be drilled – with tentatively fixated tail – without support. The drive-in nuts on square plywood pieces were included in the kit. Because they are hardly accessible later they have been glued in already.
The front window has been screwed on. Two wiper dummies are included in the kit and the Bell 429 generally has wipers, but especially this model's original doesn't have some so they are dropped here.
The two dummy antennas on the top hood are glued on. Glueing another five dummy antennas is saved until last so they are not broken off while I'm further working on the helicopter.
The decals have been applied – following the instructional video by Scaleflying.de. My UAS operator ID (e-ID) as a QR code is hidden between the stars in front of the black exhaust nozzle.
Especially without direct sunlight the model's red color seems too dark to me – compared to the original as well as to other models. Perhaps they used a paint that looks different in different lighting, but I think this shade of red doesn't quite match the original (but it does match the carpet).
Addition
A pilot manikin scaled 1:7 has been put into the right front seat where it belongs in the B 429. Seat and sticks have not been relocated to the correct places because they are bonded too tightly to the cabin floor. The manikin's left leg just goes through the gap between cyclic stick and console.
The manikin sits firmly on the seat and pushes its helmet against the ceiling. The arms do not reach to the sticks but the legs are just long enough to reach to the pedals. Hence rather the seat, particularly the seating surface seems to be too high.
As long as you don't look too closely all this is hardly noticeable. So no seat belts are needed and no fixture at all. But in flight the helicopter looks even more prototypical now.
Tail
On to the tail: bright red color here in direct sunlight. In the tail cone runs the actual tail boom with the tail rotor shaft. That's why the horizontal stabilizer can't be sticked through in one piece, and two halves have to be glued into recesses in the cone instead.
At the stab tips are fins (screwed but glued as well) with very nice, small position lights (glued on). I was scatty again and fastened them the wrong way up, their longer side upwards, but that's not a problem and just one more non-prototypical trifle.
By the way, the tail is here upside down because the stab halves were aligned while the glue was setting. It's Canopy Glue again because it bonds so well on smooth plastic surfaces and because it stays a bit elastic. And when it's cured it hardly stands out.
This big fin with tail skid and two lights has to be screwed to the tailcone's rear. Four indentations for screw holes are visible around the cable exit. The cables belong to the lights which are glued to the fin's top. The skid is simply screwed into a wooden filler.
The nice red anti-collision light was included in the kit, but no white rear position light – another mysterious deviation from "scale". I took a white LED from a Multiplex light set and glued it on where the original's rear position light is. Drilling a hole for the cable was no problem since the whole fin is hollow plastic.
The big central fin is screwed on – one black screw head is visible in the opening. In the tail cone as well as in the fin were four indentations each for screw holes. In the tail cone, the fiberglass plastic has to be drilled out for the 3mm screws. In the fin – plastic with wooden filler – 1mm or 1.5mm holes suffice to let the screws go in without breaking. (All screws and bolts included in the kit seem to be brittle and are sheared off easily.)
The holes in tail cone and fin matched, so the fin was screwed on. Now it seemed to be badly bent forward, and compared to the original helicopter it was. So it was taken off again and three of the holes in the tail cone were made slotted holes (more forward and/or upward) using the drill bit's stem (side) like a milling cutter. (The upward slotted hole is visible in the picture.) When the fin was again screwed on it was fairly upright. Seems to be a flaw in the molds.
It would have been better to drill the holes in the tail cone, then tentatively fix the fin in the correct position, and then drill the holes in the fin through the previously drilled holes in the tail cone. Vertical tail cone like in the picture and vertical upper edge of the fin might be good for alignment, or parallel leading edges of central and side fins.
The fins attached to the stab the wrong way up are visible in this picture. Now their upper and lower parts' sweep doesn't match the central fin's but that is virtually invisible. The main rotor is high enough to avoid any fin strike. (And at Scaleflying.de I saw a picture of another B 429 model with the side fins just as wrongly attached as here.)
All four lights are plugged and tested – looking good. The light controller included in the kit is not too bad. (Newer kits include a different controller.) It has three free connectors for rear position lights of which one is used here. It can be switched (with a transmitter switch) to different blinking patterns of which one is prototypical. (I had to find out by trial and error, no instructions.) Only the red anti-collision light is blinking then, in fact so smoothly that I don't want a different controller. Because it works only at voltages up to 6V I even had to slot a voltage regulator in ahead, but that was worth it for me.
The decals are applied (well visible) and the six holes for the fastening bolts close to the forward edge are drilled (hardly visible).
Correction
After a while, there was a certain disillusionment: The horizontal tailplanes came off. The Canopy Glue – so valued by me – had become unstuck while it actually bonds very well especially to plastic. Maybe it was too old, or on the fiberglass parts was still release wax even though I had rubbed them with alcohol. Anyway, I could simply rub away the whole glue and clean the surfaces with alcohol.
The same way I also could have detached the fins and re-glued them correctly. However, this was not possible because the position lights would have been destroyed if I had tried to detach them. So the fins have to stay the wrong way around.
To improve the glue's bonding, the surfaces were now sanded and roughened, first (especially in the corners) with a small wire brush in a Dremel machine and then with sanding paper. Finally they were again rubbed with alcohol what may also activate the surfaces.
So it had to be epoxy after all. Actually I don't like it (and get allergic sniffles), but the glueing went unexpectedly well. A good amount of 30-minute epoxy was mixed on a piece of plastic film and with a toothpick smeared especially into the corners of the tail cone's recess. The quantity fit perfectly, because as the tailplane was pressed from above into the recess, the epoxy flowed up and formed a fillet. Hence the tailplane was fixated in vertical position with adhesive film to the tail cone and the whole thing let alone for curing. The other side then followed the next day. A bit tempering with a hair dryer has accelerated the curing, maybe even improved.
Now the tail looks like in the picture above, except the glue is clear-transparent and not milky. It seems to bond well this time. But the tail cone as a fiberglass part is quite elastic and the tailplanes are bobbing when vibrations occur. So at this occasion I could or should have installed some stiffening elements like shown here (by the pictures at the page's bottom half) but now it's too late.
Electronics
This picture has been shown above for the 17mm servo arms. Here it illustrates positions and wiring of the servos, the flybarless system on the platform behind them, and the receiver beside them on the fuselage.
The four servos are plugged to the flybarless system, the serial-servo-signal/power lead from/to the receiver, and a lead to connect it via USB adapter to a PC for parameter setting. This lead runs in an arc into the heli's bottom. From the flybarless system's bottom runs a thick black lead down to the "receiver batteries".
That's a good place for the receiver (Multiplex RX-12-DR compact M-Link) because the leads from the ESC are just long enough and the two antennas just reach to the dummy exhaust nozzles. The lead to the left is the ESC's, to the right go the serial servo output (sum signal) for the flybarless system, the telemetry bus (Y-cable), the lights, the drive battery safety switch, the telemetry data logger activation, and two black antennas.
The black cable running helically into the dummy exhaust nozzle is the longer one of two receiver antennas. The shorter one runs into the other dummy exhaust nozzle. Both antennas are fixed inside the nozzles simply with clear adhesive tape what makes (roughly) for a 90° angle between them.
The GPS is velcroed inside and connected to the telemetry bus via a Y-cable. This is a place where it has good "visibility" of satellites and little electronic noise.
The GPS position data are logged for later analysis in case of a mishap, but I also have the transmitter check for too much speed, altitude, or distance during flight. The GPS could be even useful in case of a big mishap: if the helicopter gets lost in an invisible area (corn field). Then it could be found provided telemetry is still working.
The KONTRONIK Pyro 700-45 motor (700 size, kv 450) looks small compared to the big helicopter but these modern premium motors are just very powerful and their power comes to a great extent from high rotational speed (15,000 rpm), not only from torque. It has a built-in cooling fan.
Then again, the YGE Opto 135 ESC (no BEC, 135 A continuous) is slightly oversized because an ESC for 12s LiPo batteries is needed here and I wanted an Opto, so that was the 135. (I wanted an ESC by YGE for Multiplex telemetry compatibility.) Still it's good that it has a big heat sink with a fan because it has to work under the hood with just some venting slots in front of it.
The ESC has been placed this way to have the leads run straight into the vertical "cable duct" (right in the picture) down to the batteries in the heli's bottom. Still these leads got longer than 12 inches so I added a set of five capacitors next to the ESC to protect it from voltage spikes – just to be safe. But then I could as well have mounted the ESC right in front of the motor so its weight would be not so far fore. I just didn't think about the heli's balance at this moment.
The two three-wire leads (signal, telemetry) run straight to the receiver (right side in the background) and their length just fits. I should have let them run down into the heli's bottom and with an extension lead each from there up to the receiver again. Then I could unplug them in the heli's bottom and connect them with a USB adapter to a PC for setup. Now I have to remove the main rotor and the hood if a parameter change is needed. I was just afraid that the leads would be too long.
The distance between the components seems to be almost optimal. There's maximum distance between the ESC in the front and receiver and gyro system in the rear – the leads' length. The motor is in between but is brushless and hence without brush sparking. So if the ESC would be mounted right in front of the motor, there would be probably no EMC troubles, either.
By the way, two of the four aluminum lugs and black Allen screws with washers holding the mechanical assembly on the body are visible here, left and right of the motor. One of the two unused holes with drive-in nuts is visible behind the capacitors, between the two leads.
Into the middle compartment in the heli's bottom go the two 6s 5000 mAh 40C/80C LiPo drive batteries with XT90 connectors. They are plugged in series to the safety switch (EMCOTEC SPS SafetyPowerSwitch 70V 60/120A) at the bottom, which is actuated by a switch on the transmitter via an adpter (SPS Remote Switch Actuator) connected to the receiver. The green LED visible here means the switch is on, the drive is active.
When I later added capacitors to the ESC (see previous picture) I didn't mind that the SafetyPowerSwitch needs a shunt then. Indeed, the switch was damaged and I had to buy a new one – and a shunt (208-8 a 470R 10%). Now the green LED is always on as long as the batteries are plugged, even while the switch is off. So the LED is useless and I'm no longer sorry about putting the adapter invisibly into the compartment. Relevant is that I can activate the drive only at the runway and that there is an anti-flash effect when plugging the batteries.
Into each side compartment goes (not visible here) one 2s 2200 mAh 20C/40C LiPo "receiver battery" with XT60 connector. They are both plugged to the dual switch mixer (Jeti DSM10) hidden further back in the bottom. It is activated by a magnetic switch (see next picture) what makes the conventional switch coming with the Microbeast HD flybarless system redundant. I chose this switch mixer because it is not a voltage regulator so I can see the actual battery voltage (as receiver voltage) via telemetry.
There are two voltage sensors plugged to the drive batteries' balancer connectors. You may see the white connectors, the red Velcro tape on the sensors, and the three-wire leads to the sensor bus. The sensors have been removed later because two of them on two serially connected batteries can't work (other than I had mistakenly thought, see below).
Admittedly, this is a mess, but it just fits in the compartment between the batteries and the former behind if the batteries are slid forward. All is covered by the hatch then, which is part of the cabin floor and has three seats on it (see above). No fasteners are needed, the batteries just can't get out, they can't even move in the compartment, so all is safe in flight – extremely simple. The point is that the heli's center-of-gravity is too far forward and the batteries should be further back – it's just not possible.
The dual switch mixer for the receiver batteries is activated by a magnetic switch. I made a plastic bracket and glued it beside the left pilot's seat. This way I can open the door and hold the magnet to the blue circle to switch on or off, respectively. The green LED indicates that the receiver batteries are connected. When the door is closed the switch is not visible.
The round hole in the cabin floor under the seat is there for the screw which holds the landing gear's front left strut in the body.
This is the FlightRecorder, the telemetry data logger, attached to the back of the left pilot's seat. This way it's invisible when the hatch with the three seats is set in, but I have access to the MicroSD card when the hatch is off.
The green light indicates that the logger is active. Here I connected it to a receiver servo port to switch logging on or off. It's actuated by the same switch on the transmitter that actuates the drive safety switch.
By the way, what I call "cable duct" is quite clearly seen here: on the left and the right side, from above the cabin's ceiling to below the cabin's floor, into the left and right side compartments.
Completion
Final assembly was not really easy. While the mechanical assembly is precisely built, the fiberglass body with its plywood internal structure and the fiberglass tail have greater tolerances, not to say they are unprecisely made. That's the nature of things, though, and just means we have to prepare for it. All assembly steps have to be tentatively done while being reversible.
The cables and leads have to be pulled through the holes in the fuselage formers beforehand, also to see if their connectors go through and their length fits. The mechanical assembly with the mounting lugs has to be aligned in the body so that the tail as well as the hood fit. Holes for the tail's mounting screws might be drilled only after aligning mechanical assembly and tail together. While I did most of that well there were some minor blunders.
This is the state immediately after the final mounting steps and actually all is well. The receiver is still not installed and some leads hang loosely out of the body. It is clearly visible that the rotor shaft is vertical while the cabin's upper edge is inclined. That's just how the helicopter sits on the ground, the tail boom horizontal.
Not as clear is that the main fin on the tail is still not slanted enough. Its top edge seems not quite horizontal, and it should be horizontal when the heli is flying forward and hence is horizontal itself. Another clue is that the central fin's leading edge and that of the small fin on the horizontal stab are not parallel. That's a minor blunder (as well as the small fins being the wrong way round).
The mechanical assembly had been put into the body with all four mounting lugs and eight screws in place. Then the tail cone had been put on the tail boom, and finally the tail rotor assembly had been put on the tail boom. Now the tail cone had been fixated on the body by the six small screws prepared for that. The mechanical assembly was aligned so the tail boom was in the middle of the tail cone and the carbon frame in the middle between the plywood sides beside it. It had to be shifted forward as far as possible (to a former) to have the swashplate just in the hood's cutout. Now the eight screws in the lugs were fastened finger-tight and the lugs tack-glued to the frame with Canopy Glue.
After curing all was pulled out again. The four screws going through the lugs into the carbon frame were now finally screwed in with thread lock. The components were set in again as before as well as aligned as before. Finally also the four screws going through the lugs into the drive-in nuts in the plywood structure were screwed in with thread lock. There's a washer under each of the eight screws so when they are tightened the lugs are not turned or shifted.
The mechanical assembly, that is their lower part with the tail rotor gear, has to abut on a plywood former in the body. It has to be a bit aslant horizontally to go into the tail cone. There's a small gap between frame and former on the left side, hardly seen in the first electronics picture above.
Still the swashplate is farther back in the hood's cutout than this gap is wide, so that's a minor "built-in" blunder. The rear swashplate linkage is just clear of the hood and the juts touching the other two linkages have been filed off later.
Again the strikingly different red color, depending on lighting.
This is the finished helicopter in all its glory, just the main rotor blades missing (mounted at the flying field, and the missing five dummy antennas glued on after the maiden flight). No obvious inaccuracies or blunders.
This view shows another minor blunder which I hadn't noticed timely: The tail is inclined to the right side (seen in flight direction). I noticed that during the tentative final assembly (when the mechanical assembly went in not quite straight) but then it was already too late to correct it.
Early on, I had pressed the tail cone on the corresponding "neck" on the body's rear. It fit well without any gap so I drilled holes for the six mounting screws and glued the drive-in nuts. That was easy. I could have looked for alignment and put up with a small gap on the right side. It would have been hard to tack the tail cone in proper alignment, though, and the tail cone wouldn't fit to the body as snugly as now. I'm not unhappy with this state of things.
Just another thing that is – if at all – visible here: The tail rotor is a bit turned right around the heli's longitudinal axis (seen from behind, turned left as seen here). It had struck me when I had completed the mechanical assembly and I thought it might be my fault. But it must be intended this way because the tail boom tube has a cutout at front and a hole at rear which determine the tail rotor's position. In flight the helicopter is slanted to the left side so the tail rotor is vertical – close enough I just think.
The right-angled bellcrank lever for pitch adjustment is in normal position here, parallel to the tail boom and the tail rotor shaft, respectively. That means some degrees positive pitch to cancel the main rotor's normal torque.
That is hard to see because the blade tips are not square. The tail rotor diameter is 290 mm (11.42 in) measured at the blades' leading edges and 300 mm (11.81 in) at the trailing edges. At least the tail rotor's sense of rotation is clearly visible.
The big central fin looks good but it seems to be still not slanted (backwards) enough. A tad more would have been better but that was not really visible while the tail was not mounted to the body.
By the way, the tail rotor is another detail that is not quite "scale". The original helicopter has a quite particular tail rotor, actually two two-bladed ones stacked on the tail rotor shaft and offset by 60 degrees, not four blades in one plane and offset by 90 degrees like here. But I even like this version better.
All rotor blades had been balanced beforehand. Since main and tail rotor have four blades each, the two lightest and the two heaviest of each rotor have been identified (with the blade balance). The pairs have been labeled with colored adhesive tape (red-green and yellow-blue) and balanced with the white adhesive tape included in the kit under the respective lighter blade's tip. Quite a few people claim that ROBAN blades vary a lot in weight, but not much tape was needed here – perhaps because balance can be done in pairs. The blade holders have been labeled with correspondingly colored adhesive tape so that the blades are facing one another by pairs. That way each blade will always go to the same holder and any individual blade angle adjustment persists. (The linkages to the swashplate stay bolted to the blade holders so no color labels are needed on them.)
Most of the color labels inside the blade holders are visible here. Visible as well are the angular-shaped swashplate linkages, which dispense with the need for a swashplate driver.
They are definitely a nice thing and a nifty design, but still they look a bit flimsy compared to those on 3D helis. They seem to be a bit elastic, not really rigid.
Anyway, I had a hard time adjusting the blade tracking. In several measuring "rounds" with a pitch meter, the pitch angle measurements were not quite reproducable. Only after some trying and adjusting, three more measuring rounds showed an acceptable match of blade pitch.
Lastly it's visible how the cutout in the hood has been widened (straightened) left and right with a file. Now the swashplate linkages can move up and down without touching the hood.
Flying
The following picture shows the helicopter after its maiden flight on September, 29 2024. The tail rotor blades are pitched to the wrong side here because the gyro system reacted when I turned the heli while carrying it to the desk. The shadow on the tail is cast by our birch tree and shows once again that the red color looks much brighter than usual in direct sunlight. You can also see that inside the cabin all is completely "scale". In the 700 and 800 class scale models by ROBAN the mechanical assembly is not visible like in the smaller 600 class models. All the more at least a pilot manikin is needed (and has been added later). I like the heli's sound, I don't need a sound system.
One or two of the main rotor blades didn't track well but that was fixed before the second flight. Otherwise there was nothing faulty, except – when running up, the horizontal tailplanes bobbed in resonance with some vibrations. How to amend that now is hard to see (maybe thin bracings).
The helicopter flies very calm and stable but also very agile. The flybarless system (Microbeast) is set up entirely as per the manufacturer's recommendations. Obviously, this setup is perfect. It included specifying the highest possible swashplate throws. (That's why the servo arms have been replaced by longer ones.) These throws are excessive in normal operation. My transmitter is set up for three combined rate (100%, 80%, and 60%) and expo (50%, 40%, and 30%) values on a switch. So far I'm using the low rate only and intend to try even lower rates (and higher expo). All that holds for the tail rotor as well.
But you should observe something before flying: Immediately after activating, the Microbeast flybarless system needs dead calm, that is not the slightest vibration, and the heli has to stand level. Not before the swashplate is moved up and down after a while, the system is calibrated and its virtual horizon is correct. And after the heli has been carried to the take-off point and put down there, I even adjust the tail rotor blades to normal pitch (about 6°). Then the take-off goes smoothly without wiggle and waggle. (I assume that all flybarless systems are the same in this respect but I own only two Microbeasts.)
The heli's center-of-gravity is quite a bit in front of the rotor shaft. Obviously, the flybarless system compensates that "silently", that is it automatically holds the heli horizontal. (Luckily, the main rotor has four blades, not only two.) So I think I won't add ballast in the tail, not even if pilot manikins are put in the front seats.
The helicopter weighs 8.8 kg (19.4 lb) with batteries and 7 kg (15.4 lb) without. Its flight performance is definitely sufficient – the 700 class motor with 12s 5000 mAh LiPo battery, the 1100 rpm main rotor speeed for 1.60 m (63 in) diameter and four blades with cambered and reflexed airfoil. The tail rotor's performance is definitely sufficient to let the heli hover over a spot even in crosswind and to fly sideways. (And performance is still well sufficient at only 1040 rpm main rotor speeed.) The following video shows the first two minutes of the second flight on October, 6 2024 and rather my awkward flying than the heli's flight behavior. I have still to get used to this heli's climb performance and agility (as well as to my new transmitter).
The logged telemetry data show that amperage is – on average – a bit higher than 25 A in hover and a bit higher than 20 A in cruise flight. "On average" is emphasized here because amperage fluctuates heavily between 10 A and 40 A. Only in abrupt maneuvers it goes over 40 A and only once (when I panicked and over-pitched) there was a 87 A spike. (That can be prevented by reducing collective pitch throw in the transmitter.) The ESC (by YGE) can stand 135 A so it's oversized for normal flight. However, due to or despite this low partial load (15% to 20% of nominal limit) its temperature does not go above 50°C (120°F), probably due to "active freewheeling".
Average draw is about 360 mAh/min (equivalent to 21.6 A). After 7 minutes flying time (including runup) the two 6s 5000 mAh batteries still have 3.8 V cell voltage (storage voltage). The two 2s 2200 mAh "receiver batteries" lose up to 0.1 V cell voltage per flight so they are good for three or four such flights. Still safely for all batteries, one could do 8-minutes flights, four of them in a row with the "receiver batteries", but would have to re-charge to storage voltage in the end.
Average draw is only 325 mAh/min (equivalent to 19.6 A) at only 1040 rpm main rotor speeed – without any noticeable loss of performance and agility. Even lower main rotor speed (and amperage) is possible. One could do 8-minutes flights, now only three of them in a row with the "receiver batteries", and would still end up at about storage voltage.
By the way, we have an estimate for average servo amperage draw now. Four 7-minute flights mean about half an hour, 3.8 V idle cell voltage means – very roughly – half charge of the two parallel 2200 mAh "receiver batteries". So they are loaded with 1C amperage, that is 4.4 A. Liberally rounding up we assume 5 A as a ballpark estimate (even 5.4 A at closer look, and still nearly 5 A with 15 mm swashplate servo arms). Servo amperage draw during hover is distinctly lower than in cruise flight (see lower voltage drop at the end of the example diagram below).
Flying this helicopter is enjoyable but there are some caveats: I'm still learning and getting used to, and in one case I even panicked when the heli dropped suddenly in hover flight. Full collective pitch let it jump up again but I felt reminded of an advice in the instructions – to keep forward motion except close to the ground (in ground effect). There might be a tendency to vortex ring state.
And I really like the heli's shape, paint scheme, and color, but just these can cause bad visibility. The paint scheme has no clear straight lines at all and not even the landing skids are parallel. When the sky is overcast, the fuselage looks dark and contrast is low. Then, the rounded shape has a silhouette which sometimes makes it hard to see the heli's flight attitude and keep orientation. At least I'm more relaxed when flying this heli in sunlight than when it's overcast. This heli is very nice but not for rookies. You can't have it all.
At last two comments: One at YouTube calls the helicopter a beauty – agreed. A clubmate's grandson calls it a Christmas helicopter – not agreed but understood (and unknowingly he has a point, see above).
The Bottom Line
… just as a few notes:
- Nice, prototypical model.
- Minor details are not scale.
- Substantially built, good quality, attractive price.
- Well-designed mechanical assembly.
- The tail rotor's sense of rotation can be reversed to CW.
- The tail rotor does spin during autorotation.
- Flies calmly but is still agile.
- Not for rookies – experience helps.
- But not really challenging, either.
For – even some challenging – details see previous sections.
The most important information is condensed in the following notes, for important details see above.
Notes
A 700 size kv 450 motor is just about perfect for this heli. A bigger motor (750 or 800 size) would just have an excessive power reserve, one with higher kv (520 or 560) would draw more amperage and even burden the ESC. Only if cheap components (motor, ESC, battery) with high internal resistance are used, a 750 size motor might be advisable to have a good power reserve.
The motor should have not much more than 50 mm diameter (probably about 60 mm maximum) to fit in the recess. In any case, it has to be a version with a long shaft journal: 38.5 mm or just a bit longer (and 6 mm diameter).
An ESC rated for at least 100 A amperage is at the safe side. In any case, it should be part-load resistant, though ("active freewheeling").
Two 6s 5000 mAh LiPo batteries are good for 8 minutes flight time then (as long as they are new), hence smaller batteries are possible. Average amperage is 20 A to 22 A; peaks go to 40 A, only in extreme cases up to 90 A.
Because the leads from the ESC to the batteries are long, additional capacitors are advisable. A safety power switch may need a shunt then.
Good standard-size heli servos (for at least 600 size helis) are required. As high voltage types they will together draw about 5 A on average from a 2s LiPo battery. A flybarless system with high-amperage power connector(s) or a separate servo power supply is advisable.
13 mm servo lever arm is a minimum requirement, but actually it should be 14 mm. 17 mm lever arm works really well but 15 mm might draw 10% less servo amperage. Aluminum arms are good, plastic arms will do in a pinch.
The swashplate pushrod length specification is basically correct, and there is no need to bend the middle pushrod (no clash).
To be more "scale" and aerodynamically better (more effective, less
noisy), the tail rotor's sense of rotation can be reversed:
(1) The front bevel gear has to be reversed.
(2) Both coupling pieces of the "torque tube" have to be secured with
screws.
(3) The tail rotor's blade holders have to be reversed as well.
The right pilot seat as well as cyclic stick should be mounted more to the right side than provided for by slots and holes in the cabin floor. The collective stick should be on the seat's left. The seating surface should be lower.
The horizontal stabilizers need some kind of stiffeners which have to be glued in stab and tail cone.
The central fin has to be inclined backward, more than the indentations (for drilling holes) in fin and tail cone suggest.
The central fin needs a white LED as rear position light (not included in the kit). The light controller yet has a port for it.
My sample weighs 8.8 kg (19.4 lb) with batteries; 8.0 kg (17.7 lb) seems to be bare minimum. Pilot manikin and accessories come in addition but weight is still nowhere near an issue.
The model's center of gravity is quite in the fore; that should be minded when building the model.
1100 rpm main rotor speed is good for all practical purposes. The heli is agile then but also economical. It is still agile but even more economical at even lower main rotor speed, for instance 1040 rpm.
After 7 minutes flying time (including runup) at 1100 rpm main rotor speed, the (still quite new) 12 5000 mAh LiPo cells still have 3.8 V at idle (storage voltage) – after 8 minutes at 1040 rpm. A BEC for servo power would cost about ¾ of a minute flying time.
The model is scaled 1:7 so that has to be the scale of a pilot manikin or other "accessories" as well.
Thoughts
Some things just turned out all right, a few other things can be seen – in hindsight – as flaws, especially when seen in context. I might consider different solutions if I would build the helicopter once again.
The motor recommendation by Scaleflying.de was very good. A bigger motor (750 or even 800) would just have an excessive power reserve, but weigh (80 g / 2.8 oz or even 125 g / 4.4 oz) more and cost more. A higher kv (520) motor would draw (3.2 A) more amperage, what diminishes flight time (by 1 minute). The instructions anyway specify only 5 minutes flight time (even 3 minutes less), for whatever reason.
Not a problem but not really good either is the center-of-gravity too far fore. Two drive and two "receiver" batteries contribute a lot to this, but they can't be moved backwards. The ESC contributes less but it could be moved backwards.
Considering the more than sufficient capacity of the drive batteries, smaller and lighter ones could be used, for instance 4000 mAh (instead of 5000 mAh) with a 20C/40C (instead of 40C/80C) amperage rating. That would be still enough for this application (1½ minutes less flight time) and would save quite some weight (415 g / 14.6 oz far fore) and cost.
The ESC could be one size smaller as well, only 105 A (instead of 135 A). That would save weight (50 g / 1.75 oz) and cost.
The ESC could be mounted more rearwards, close to the motor. The leads to the batteries would be even longer then, but that is yet compensated for by additional capacitors.
The ESC could be one with a BEC (instead of the Opto one), saving the two "receiver batteries" and the dual switch mixer altogether (270 g / 9.5 oz) but weighing (40 g / 1.4 oz) more. About 250 mAh per flight would be additionally drawn from the drive batteries (equivalent to ¾ of a minute less flight time) but that would be still no problem.
ESCs with BEC seem to have a vogue, anyway. At least the Opto ESC used here is no longer made for this reason. And the BECs seem to be quite reliable now, other than a few years ago. They can really deliver 12 A amperage at really constant voltage, and they even have a mixer for a backup battery, which is usually quite small and lightweight.
This battery would be important for my feeling of safety. The BEC could be set to 8.4 V, the battery's maximum voltage. In case the BEC fails and the backup battery stands in, its voltage will rapidly drop below 8.0 V and that can be proclaimed by telemetry. But since such a case shouldn't actually happen, the backup battery has to be charged before flying and discharged (to storage voltage) after – more inconvenient than dual "receiver batteries", which are flown to storage voltage.
With a BEC, the switch for the Microbeast HD flybarless system would be useful. But even the special lead from the BEC (in the ESC) is quite thin and has only a small "servo connector" on its ESC end, not an MPX connector – not adequate to the amperage peaks to be expected.
These are the two or three reasons why I would still stick to the solution with two "receiver batteries" and a not-voltage-regulating switch mixer. Rather I might even use only one (but bigger) "receiver battery" and the Microbeast switch and I would still feel safer than with a BEC. If I had decided for a flybarless system without high-amperage connector, I would even use a separate servo power supply with telemetry.
By the way, I'm suspicious of the BEC and the drive battery as well. Both sometimes failed in the past so it was twofold risk. Both has happened to me but no receiver battery failure yet – that could still happen but is far from likely. That's how we are informed…
More
Wikipedia article about the Bell 429
Air Zermatt Bell 429 pictures at airliners.net
Web page by Bell about their 429
ROBAN MODEL (in China) web pages (English)
Scaleflying.de web pages (English)
Web page by Scaleflying.de about the ROBAN Bell 429 Air Zermatt (English)
Web page by Scaleflying.de about the ROBAN Bell 429 Air Zermatt in the Web Archive
YouTube channel of Scaleflying.de (German)
Web page by KONTRONIK about the Pyro 700-45 motor (English)
old Web page by KONTRONIK about the Pyro 700-45 motor in the Web Archive (English)
Web page by YGE about their high-voltage ESCs, especially Opto 135 (English)
BEASTX Microbeast Plus Wiki page
ROBAN Bell 429 build thread at rc-heli.de (German)
another ROBAN Bell 429 build thread at rc-heli.de (German)
ROBAN Bell 429 build thread at vstabi.info (Mikado forum, German)
ROBAN AS350 review by John Salt (English)
Telemetry
This model helicopter, like my HIROBO Schweizer 300, has a modern electric drive with a governor. Hence it just has to be equipped with modern telemetry as well to ensure flight safety.
Receiver and ESC provide even half of the displayed values; just two extra sensors are used to monitor the voltages of both drive batteries. For later interpretation of these data, a GPS adds 3D location data. The FlightRecorder logs all telemetry data so they can be analyzed later – routinely or possibly after an accident.
Most of the values are not just logged and transmitted but also monitored for exceeding or falling below adjustable limits. Values and any "alarms" are shown on the transmitter display and announced by voice-output. That needs a sensible setup and therefore I devise a coherent plan in the form of a spreadsheet.
Sensor Setup
My Cockpit SX 12 transmitter (bought in 2020) has an operating system similar to that of a smartphone and a touch screen. In the telemetry display pages, the font size and hence the number of displayed values is adjustable. For me, six rows (values) per page is the best compromise between overview and readability. The display has been adjusted to that by dragging the touch pen. In this plan, two pages with six rows each are marked off by red lines so a third page with four rows remains:
These are the basic telemetry device settings. Up to 16 values may be transmitted on the MSB; the display sequence (leftmost column) follows from the chosen bus addresses (fourth column).
Alarm means optional upper and lower alarm thresholds while Parameters are mandatory and have to be set in the respective devices. (Options – that is maximum, minimum, or mean values – are not needed here because all telemetry data are logged.)
All of the possible 16 values are used here but no special allocation to pages has been tried. The values of the five devices on the sensor bus have been just put in consecutive rows. Yet the resulting partitioning is quite useful: The first page with values interesting during flight is followed by a second page with values meant for later analysis and a third (shorter) page with all drive battery voltages.
The two values delivered by the receiver have been left in their default places 0 and 1 so the transmitter's warning LED for too low receiver voltage will work. For a 2s LiPo receiver battery, 7.4 V under load is a safe warning (or alarm) threshold because it implies there is enough charge left for landing, even if one of the two batteries is inoperative. "Priority off" means none of the values (or addresses) is transmitted more frequently than the others – that would be actually needed just for a high-resolution variometer tone.
The ESC provides seven values – all of them interesting. The controller's "actual opening" and "set opening" are new to me and especially interesting. The order of values has been changed only in that "drained charge" has been put in place 4 so the transmitter's warning LED for too low battery charge will work. In exchange, the "main rotor rpm" has been put in place 6. Number of motor poles and gear ratio have to be specified because rotor rpm is derived from the motor's field frequency by calculation.
"ESC input voltage" is actually redundant since it will be only slightly lower than the voltage measured directly by the voltage sensors on the batteries (due to the resistances in between). Still the low-voltage warning is activated as an alternative solution. The trigger level is set to 42.0 V total battery voltage, corresponding to and as replacement for 3.5 V per cell – the level actually wanted but not possible with the voltage sensors.
For "amperage", the high-amperage warning trigger level is set to 90 A – far more than the expected 40 A peak, far less than the ESC's 135 A rating and the (40C) batteries' 200 A rating, but about what the motor can stand for a short while.
Both batteries – 6s 5000mAh 40/80C LiPo each – are charged to only 4.18 V (instead of 4.20 V) per cell, which is why capacity is only 4850 mAh (97% of nominal capacity). 20% is deducted to anticipate degradation by usage and age, so 3880 mAh are left. At 30% remaining charge, that is 1200 mAh, a low-charge warning should be issued. The ESC simply counts "drained charge" so the warning trigger level has to be set to 2680 mAh, or rounded 2700 mAh.
The GPS 3D location data are logged by the FlightRecorder in any case. They are used to render a flight path in Google Earth. In diagrams, "spatial speed 3D" and "altitude AGL" as well as perhaps also "spatial distance 3D" are useful data to be plotted over time. These have to be displayed (sent on the MSB) during flight so that the FlightRecorder can log them. The three upper warning levels have been activated. The GPS is set to "slow aircraft", meaning horizontal speeds up to 79 km/h and vertical speeds up to 54 km/h. The "spatial speed 3D" warning trigger level is simply set to 79 km/h. An "altitude AGL" warning is triggered at 50 m, our least limit when the control zone is active and my limit to keep orientation. For the same reason, my personal limit of 200 m is set as "spatial distance 3D" warning level.
Immediately after the special voltage sensors have been plugged to their respective drive battery, they check their cell voltages and report their state of charge (in %) in the place of cell voltage. If one of the cells is below the voltage characteristic for "LiPo" cells charged "80%", a warning is issued – a safety feature. Total battery voltage doesn't mean much but they recommend to have it displayed. Then, if the 3.4 V cell voltage threshold (set here) is underrun, the respective warning is displayed in the place of total voltage. Only the so-called absolute low-voltage alarm (threshold 3.1 V preset for LiPo) is shown in the place of cell voltage. The 3.4 V warning threshold is the default value and is tried and trusted by me – it's not too high. It should be even raised to 3.5 V for an earlier warning (at 25% remaining charge instead of 15%), but the sensors don't allow that. Still they are good for another warning stage, detecting a too low single cell voltage. If the sensors know the amperage value's address ("3"), they refrain from issuing low-voltage warnings in case of short amperage peaks.
Safety Plan
The warning trigger levels have been chosen so that the helicopter is as safe as possible and the batteries are taken care of:
In normal circumstances, the ESC's low-charge warning comes first at 30% (nominally) remaining charge. Then, at about 25% (really) remaining charge, the 42.0 V battery voltage level – equivalent to 3.5 V per cell – is underrun. If not after the first warning already, the helicopter should be landed after this second warning.
For once the voltage sensors warn at 3.4 V cell voltage (or 40.8 V battery voltage), remaining charge is only about 15%. That is already detrimental to the batteries but still enough for a safe landing before the ESC reduces power so much that cell voltage is kept up at 3.3 V in order to avoid an immediate battery damage (or, without power reduction by the ESC, before the battery is discharged too much).
So the low-charge warning is the actual advice to land soon – like a white reminder light in a cockpit. 42.0 V battery voltage (equivalent to 3.5 V cell voltage) is an urgent request to land very soon (ASAP) – like a yellow warning light. 3.4 V cell voltage (equivalent to 40.8 V battery voltage) is the last warning to land immediately – like a red alarm light. After that, you would at least partially lose control because eventually power is reduced too much for flying.
Circumstances are not normal if the batteries are cold or old. In these cases, they have not (any longer) their nominal capacity and the low-voltage warnings are triggered soon after the low-charge warning, in extreme cases even before. We can rely on that because a helicopter – unlike an airplane – permanently draws a lot of amperage.
In case of a bad cell, the voltage sensors can warn just in time, possibly without any other warning issued before. So you must always pay attention to a warning's type – the red light may light up unexpectedly.
(This safety plan's rationale is explained here.)
Voice Output
My Cockpit SX 12 transmitter (bought in 2020) has built-in voice output, which is configurable but simpler compared to the Souffleur used with the old ROYALpro9 transmitter (bought in 2008). With a bit effort, it's possible to generate special sensor value announcements and load them into the transmitter. However, these announcements are bound to fixed addresses then. So for all models the same respective addresses (for instance for rotor speed) would have to be used. That's virtually impossible so the standard announcements (for instance “sensor 3”) are even more meaningful. They are useless, though, which is why I replaced all of them by a short beep to shorten announcement time. It's still possible to distinguish all announcements by their units, or at least by their values, and short announcements are good in any case.
Several values can be included in only one group which is announced at one go. But not only a switch can trigger the annoncement but also tilting or moving the transmitter. That's how I trigger the announcement of a few critical values: amperage, drained charge, ESC temperature, and lowest cell voltages. Separately, remaining flight time is announced by tilting the transmitter.
By the way, voice output is also assigned to switch positions. That's very convenient for me since I keep forgetting which switch and position is assigned to which function. Now I can simply try and I'm prompted. Of course, I do that only before the helicopter is activated; then I can again keep it in mind for the next few flights.
New Transmitter
My latest transmitter, PowerBox ATOM M-Link (acquired in July 2024), is especially good at telemetry and voice output. Alarm trigger levels have no longer to be set in the sensors. Instead, there are up to four trigger levels (two upper and two lower limits) per displayed value in the transmitter. Hence they can be modified easily any time. The kind of alarm is highly configurable, from beeps over vibrations to voice output. The latter needs no prepared sound files since it's a full text-to-speech system, hence very easy to set up. Of course, voice output is also assignable to switch positions. The whole telemetry concept was implemented in this transmitter, especially the safety plan.
The transmitter automatically records all telemetry data so the FlightRecorder in the model is actually redundant. But transferring the data via an USB stick to a PC is just as inconvenient as taking the MicroSD card from the model and reading it on a PC. Then, I don't like the style of diagrams in the PowerBox Terminal app so I'll stick to what I have – FlightRecorder and LogView.
Though not telemetry-related, another good feature should be mentioned here: some kind of pre-flight check list. That means the transmitter starts only if predefined conditions are complied like master switch off, collective pitch at a certain position, and so on.
Problem
Like I "forgot" the shunt for the safety switch, I realized only as an afterthought that the two voltage sensors can't work as I thought. (There's a thread about that in the Multiplex forum for years, yet not one word about it in the instructions.)
The sensors would need a ground connection between sensor bus (receiver) and drive battery (due to ESC without BEC, with opto-isolator). They can't work potential-free (what I had mistakenly thought). This connection is possible to the first 6s battery but of course not to the second because two different potentials are not possible. And if I would connect the second sensor on the second battery to the same ground, it would have the full 12s LiPo voltage (up to 50.4 V) while it can stand only 36 V.
So both sensors have been removed from the helicopter and the related display widgets in the transmitter have been deleted. The safety plan is incompletely implemented now. In the new transmitter, the "red alarm light" is represented by a second battery-voltage warning – 40.8 V instead of 3.4 V cell voltage. There's no cell-voltage monitoring, so I have to listen even more for warnings during flight. All 12 cells have to be checked on the charger before flight and especially after, to see if their state of charge is still equal and if any cell lags in charging.
Example
The following diagram shows a telemetry log excerpt in that it plots the values of receiver battery voltage (blue) as well as drive battery voltage (green), amperage (red), and discharge (pink) over flight time. This was an average flight with several horizontal figure eights flown at cruise speed or slow and with intermediate hover. In cruise flight, amperage is remarkably low as are the fluctuations, which reflect straight parts and turns and go up to about 30 A. In slow flight and hover, amperage is substantially higher and fluctuations go up to about 40 A.

"Receiver battery" voltage (blue) is actually what the receiver measures as voltage coming via the serial signal lead from the flybarless system. Originally, it comes over a MPX connector and a thick cable from the dual battery switch, which taps two parallel 2s 2200 mAh 20C/40C LiPo batteries with XT60 connectors.
Sensor resolution is 0.1 V what makes for the jagged blue line. Before and after flight, the electronic devices (including servos) and the lights draw little rest current. During flight, discharge is still slow on average but voltage fluctuates with 0.3 V amplitude, sometimes even 0.5 V. Obviously, the flybarless system lets the servos work fast and hard to make for the calm and steady but agile flight so appreciated by us. Interestingly, voltage drops less in slow flight and hover (around 2:30 on the time axis). Obviously, the servos are less busy then.
The receiver batteries had started that day at 8.2 V and this is the second flight, starting at 7.9 V. After the flight, rest voltage is by almost 0.2 V lower than before. Of course, discharge is not linear but degressive first and then progressive. The third flight started at 7.8 V and ended at 7.5 V rest voltage or 7.6 V idle voltage, respectively (3.8 V per cell, storage voltage).
Drive battery voltage (green) is what the ESC measures as voltage coming via the thick leads from the heli's bottom, where the master switch is plugged with bullet connectors and the two 6s 5000 mAh 40C/80C LiPo batteries in series with XT90 connectors. The green line shows a regular discharge curve seen over the whole time of measurement, but there is a lot of more or less deep voltage drops which correspond exactly to the peaks in the amperage line (red). Idle voltage – before and after flight compared – dropped by about 4.6 V. That is 0.38 V per cell, from 4.18 V end-of-charge voltage to 3.80 V storage voltage – bingo!
Drive amperage (red) is drawn by the ESC from the batteries. I didn't find the right setup option for slow runup so I do it myself with a rotary knob on the transmitter. There are drops and overshoots, but finally amperage settles at 16 A when rpm settles at 1100. Take-off and acceleration need a lot of amps, especially spot turns as well. Slow flight with hover (around 2:30 on the time axis) produces huge fluctuations. Cruise flight with turns (3:00 to 4:20) results in low draw and small fluctuations (peaks during turns). Approach for landing (from 6:40 on) produces high peaks, but the closer to the ground the less amperage is drawn. After touchdown, amperage settles at 11.5 A with lower collective pitch than before take-off.
Drive battery discharge (pink) is actually amperage draw accumulated over time by the ESC. It starts slowly like amperage and then reflects kind of an average amperage because short-waved fluctuations are smoothed out by accumulation. High amperage for half a minute (around 2:30) makes the pink line's slope steeper, cruise flight (3:00 to 4:20) less steep. All in all the line is still not far from being a straight diagonal (except start and end) so it lends itself to calculating average amperage: 20.8 A in this case.
ESC
Probably the instructions for the YGE Opto 135 ESC are comprehensible for insiders only. They don't make clear how a slow runup for helis has to be set up. The setup app shows more details but doesn't explain anything, either. At least I managed to set up the governor-store mode, just by considering how the ESC could work. In case someone has the same problems as I had, maybe the following two diagrams are of some help.
The light green line up and the red line down show again battery voltage and amperage during the same flight as in the previous section. But in between are now shown (dark red) main rotor speed as well as (green) throttle preselection and (blue) throttle opening. The two latter terms are explained by the diagram.

It shows that main rotor speed (dark red) is maintained constant very well. (The resolution of the rotor speed value is obviously 10 rpm.) Just that is prescribed by the so-called throttle preselection (green). It's kind of a set value, in this case 76% of the highest possible value. In governor mode this may be seen as the highest possible rotor speed, albeit only "theoretically" possible with the same battery voltage as when setting up and without any load.
Then, the so-called throttle opening (blue) is the corresponding actual value. In simplified terms, it may be seen as percentage of the highest possible voltage available to the drive. This gets lower and lower while the battery is discharged so the percentage has to be raised more and more. The blue line is quite jagged but tends to go up over time (from left to right in the diagram). It's an exact mirror-image of the (light green) voltage line above since – to keep rotor speed constant – the governor has to open up the more the lower voltage falls.
The jags in the (red) amperage line below correspond exactly to those in the (blue) throttle-opening line since throttle opening is what releases the amperage needed to keep rotor speed constant, for instance when pitch is increased. Accordingly, there are a few drops of rotor speed immediately followed by higher throttle opening, increasing amperage and bringing rotor speed back to set vaule – sometimes so fast that the set value is overshot. The converse holds for less load by decreased pitch.
Probably, the values of throttle-preselection and throttle-opening are equal if and when voltage is the same as that applied during setup. The whole preselection setup consisted in just running up to a certain rotor speed after the governor-store mode had been set. That should be done at more than 70% throttle preselection as per the instructions what just ensued at 1100 rpm rotor speed, probably due to proper layout of motor and gear. So this rotor speed can really be seen as "standard" (term in the YGE instructions, in contrast to hover and 3D) for this helicopter since the 76% throttle-preselection required for it is just in the middle of the recommended 70% to 80% range.

By comparison this diagram shows the values of a later flight with the now-preferred lower rotor speed 1040 rpm. 72% throttle-preselection is required for this speed so it's still in the "standard" range, and the heli's flight behavior still feels like "standard" as well. For low rotor speeds (hover) the YGE instructions recommend the 60% to 70% throttle-opening range. Indeed that could be tried with this heli for it feels nowhere near "soft" in flight so far. In any case this heli is at the safe side since throttle-preselection and throttle-opening must not be too low.
For heavy scale helis the instructions recommend high throttle-opening to prevent high heat losses, whatever that means. However, heat can't be a problem here because the ESCs temperature did not exceed 33°C (92°F) during this flight and 36°C (97°F) during the previous flight. Of course this heli could haul a heavy load but then it would be flown at high rotor speed, that is high throttle-preselection and hence throttle-opening as well. Rotor speed could be markedly above 1100 rpm, 1200 rpm or even more, for the drive has the power reserve needed for that (see next chapter).
As soon as all that has been tried, the telemetry values of throttle-preselection and throttle-opening are actually no longer needed. Still I won't "deactivate" them in the ESC because they are logged, and just these logged data could be important or at least helpful in case of a drive problem or an incident.
"Throttle-opening" plays a role also in the next chapter (about the drive). There it's called "throttle ratio" as the percentage of full voltage which results in the amperage necessary for a certain rotational speed. While full voltage continuously decreases here, constant nominal LiPo cell voltage (3.70 V) is assumed there in the first place – a value so low that it's not reached here (only 3.80 V). Hence the "throttle-openings" here are lower than the "throttle ratios" there but rise more and more.
Drive
To render the model's performance in the REFLEX XTR² flight simulator, the drive characteristics had to be ascertained. "Drive" means electric motor, main and tail rotor gears and shafts, and both rotor heads. Some parameters were specified by the motor and model manufacturers, respectively, some had to be measured, and the rest was calculated. After the helicopter had been flown for the first time, telemetry data logged in-flight allowed for calculating realistic operational figures. The calculation spreadsheets are available for download.
In the following sections, nominal LiPo cell voltage (3.70 V) multiplied by cell count (12) is assumed as battery voltage. To be easy on the batteries, discharge should be limited to 3.75 V per cell, starting from 4.20 V or less. Therefore, voltage is practically always (up to 13.5%) higher than "theoretically" assumed here.
Parameters
The motor's manufacturer specified 450 rpm/V specific rotational speed (kv) and 32 mΩ (milliOhm) internal resistance (impedance, Ri). The third characteristic required to calculate the motor's performance, idle (no-load) amperage (current, I0), had to be measured. To this end, the motor was just clamped to the workbench and run with the designated ESC/governor and a 10s LiPo battery. The ESC's telemetry showed both rotational speed and amperage. I0 turned out to be 2.3 A at 15800 rpm (equivalent to 1143 rpm main rotor speed).
Similarly, amperage was measured after the mechanical assembly had been made ready (see above where also the gear ratios are specified). The ESC/governor was set to 1100 rpm main rotor speed and after some run-in of the whole drive the no-load (no blades) amperage I0 turned out to be 3.3 A now. So main and tail rotor "gear" made for an additional 1.0 A amperage draw. A simple calculation resulted in 94.5% "gear" efficiency (which, as the percentage of torque transmission, by definition includes any additional friction losses under load).
That, the specified gear ratios, and the measured blade lenghts and widths are enough to render the drive in the flight simulator. I didn't even check the specified values (except the gear ratios) because that would be hard and because the calculations are simplifications, anyway.
Maximum Power
In the simulator, the motor's mechanical-power characteristic is approximated. For that, maximum mechanical power has to be specified as well as the rotational speed at which it occurs. Both can be calculated from the specified drive parameters. However, the resistance of batteries, ESC, and leads (including connectors) has to be estimated. From the charger's display, 30 mΩ have been taken as one battery's resistance, and 20 mΩ have been assumed for ESC, leads, and connectors.
By way of calculation, maximum mechanical power is 4110 W at 718 rpm main rotor speed, corresponding to 9930 rpm motor speed. These values are needed for the simulator only.
In practice, the motor can't deliver this power because amperage would be far too high. Even at the designated 1100 rpm set rotor speed, amperage would be 95 A and still too high. However, if the ESC would draw only short bursts of full power at a constant 1100 rpm main rotor speed, that would be harmless for the motor (for the ESC in any case). On average, amperage is not even quarter of that.
Hence it's not a problem that too low ("conservative") values are calculated here because just nominal LiPo cell voltage (3.70 V) is assumed and no practical working voltage (between 3.75 V and 4.20 V). In practice the values would be up to 13.5% higher (4,20 V), most of the time during flight between 8% (4.00 V) and 3% (3.80 V).
Cruising
In flight, the drive is working at partial power, which is permanently varied by the governor, though, to keep the recommended and favored 1100 rpm main rotor speed. Here we just assume the average amperage of a typical practice flight: 21.6 A. Then, the "throttle ratio" is the percentage of full voltage which results in this amperage at that speed and which is determined simply by trying: 82% for 21.6 A and 1100 rpm.
The "throttle ratio" corresponds to the "throttle opening" described above. But the former refers to the assumed nominal (3.70 V per cell) battery voltage while the latter refers to the actual, (from 4.20 V to 3.75 V per cell) permanently decreasing battery voltage. This is always higher than nominal voltage, which is why the throttle opening is always lower than the throttle ratio and permanently increased. However, just due to this permanent readjusting, rotor speed and all other values are permanently kept constant – at least on average. That's why we can use the constant throttle ratio as a convenient substitute for the variable throttle opening. Both are equivalent power settings, the former being kind of a nominal value (not to be confused with the throttle preselection described above).
So this is the diagram for 82% nominal power setting (at 3.70 V battery cell voltage) and 21.6 A amperage, which is needed for cruising at 1100 rpm and automatically done by the ESC/governor during flight.
Advancing to 100% at still 1100 rpm, to make for a massive climb with maximum blade pitch, could draw up to 95 A at 3.70 V battery cell voltage. This big (and in practice even bigger) power reserve (4.4 times average amperage, 4.8 times average mechanical power) is shown in the previous diagram.
The diagram shows that the drive is operated at fairly low power and thus relatively high motor speed – relative to maximum speed, which is proportional to the respective required voltage. This is where the efficiency curve eta (light blue) is at its maximum, and it's 79% efficiency overall (taking all electric resistances and all mechanical friction into account). So the motor is an excellent choice for this helicopter, both for its fitting kv value and high efficiency. It makes for minimum amperage draw and maximum flight time at any practical rotor speed. For even longer flight times, even lower rotor speeds could be chosen, which require even lower amperage (see below).
All curves in the diagram are "shrunk" horizontally by the "throttle ratio". So maximum rotor speed is here 82% of that in the maximum power case above. At still 1100 rpm the drive is now working a tiny bit past peak-efficiency rpm, but efficiency is not noticeably lower than peak value: still 79%.
Hovering
When hovering without ground effect, the average amperage is noticeably (4.9 A) higher than in cruise flight: 26.5 A. Still the "throttle ratio" determined simply by trying is only marginally (1%) higher: 83% for 26.5 A and 1100 rpm.
How well the "throttle ratio" corresponds to the ESC's "throttle opening" has been checked with the telemetry data log shown in the diagram above. There's a short period (around 6:50) of hovering without ground effect and an average amperage of 26.5 A as assumed here. Battery voltage is 44.5 V but that goes up to 45.5 V at idle after landing. Voltage assumed here is 44.4 V, divided by 45.5 V there makes 0.976. 81.4% throttle opening there divided by 82.8% throttle ratio here makes 0.983. That's a pretty good (0.7%) match considering I assumed 60 mΩ internal resistance for the batteries here while it seems to be only 38 mΩ there (1.0 V voltage drop divided by 26.5 A).
This is the diagram for 83% power setting and 26.5 A amperage, which is needed for hovering at 1100 rpm and automatically done by the ESC/governor.
There is still a good power reserve (3.6 times average amperage, 3.8 times average mechanical power), possibly useful to escape a vortex ring state.
Here the curves are shrunk horizontally to 83% so they are 1% more to the right than in the previous, the cruising case. Only 1% more voltage is needed to increase amperage by 4.9 A because speed is not increased. At still 1100 rpm the drive is now working pretty much at peak-efficiency rpm: peak efficiency is 79%.
Cruising 2
Here as well we just assume the average amperage of a typical practice flight, but now at 1040 rpm main rotor speed (9% less). Amperage is noticeably (2.0 A, 9%) lower than at 1100 rpm: 19.6 A. Now also the "throttle ratio" is noticeably (5%) lower: 77% for 19.6 A and 1040 rpm.
This is the diagram for 77% power setting and 19.6 A amperage, which is needed for cruising at 1040 rpm and automatically done by the ESC/governor.
Advancing to 100% at still 1040 rpm could draw up to 111 A. This massive power reserve (5.7 times average amperage, 6.3 times average mechanical power) is not usable in practice because blade stall could occur at this amperage but should be prevented by limiting the blades' pitch. The amperage is not too high for the ESC and even the motor could stand at least short bursts of it.
Here the curves are shrunk horizontally to only 77% because main rotor speed is only 1040 rpm now and because 2.0 A less amperage is needed for that. Operating at 1040 rpm is even a tad further past peak-efficiency rpm than at 1100 rpm, but efficiency is only a tad (1%) lower: 78%.
Cruising 3
Here as well we just assume the average amperage of a typical practice flight, but now at 1160 rpm main rotor speed (5.5% more). Amperage is a bit (0.9 A, 4%) higher than at 1100 rpm: 22.5 A. The "throttle ratio" is a bit (5%) higher: 86% for 22.5 A and 1160 rpm.
This is the diagram for 86% power setting and 22.5 A amperage, which is needed for cruising at 1160 rpm and automatically done by the ESC/governor.
Advancing to 100% at still 1160 rpm could draw up to 78 A. This power reserve (3.5 times average amperage, 3.8 times average mechanical power) is usable in practice because blade stall would not occur at this amperage. It is not too high for the ESC and even the motor could stand it.
Here the curves are shrunk horizontally to 86% because main rotor speed is even 1160 rpm now and because 0.9 A more amperage is needed for that. Operating at 1160 rpm is just as far past peak-efficiency rpm as at 1100 rpm and efficiency is the same: 79%.
Cruising 4
Here as well we just assume the average amperage of a typical practice flight, again at 1100 rpm main rotor speed, but now with a 700-52 kv 520 motor (15.5% more). Amperage is noticeably (3.4 A, 15.7%) higher than with kv 450: 25.0 A. But the "throttle ratio" is noticeably (10%) lower: 72% for 25.0 A and 1100 rpm.
The KONTRONIK Pyro 700-52 (kv 520) motor is assumed here. It has the same dimensions as the Pyro 700-45 (kv 450) but less stator windings and thicker winding wire. Since the manufacturer specifies only kv now, internal resistance has been taken from the old web page in the Web Archive and idle current has been simply estimated 15% higher for the -52.
In the calculation spreadsheet, throttle ratio and amperage have been set by just trying so that – at still 1100 rpm rotor speed – mechanical power is the same as in the -45 case. After all, that's what the helicopter needs for cruise flight. The electrical power needed for that is slightly (1.7%) higher here because the internal resistance losses increase with amperage squared. Efficiency is thus 1.6% lower. But the calculation is not that accurate because the idle-current value is a rough estimate.
This is the diagram for 72% power setting and 25.0 A amperage, which is needed for cruising at 1100 rpm with a kv 520 motor and automatically done by the ESC/governor.
Advancing to 100% at still 1100 rpm could draw up to 145 A. This power reserve (5.8 times average amperage, 6.4 times average mechanical power) is quite similar to the previous case of 1040 rpm rotor speed with the kv 450 motor.
In both cases, the relatively low rotor speed requires more curtailing of voltage. Consequently, full voltage results in more increase of amperage and power.
Here the curves are shrunk horizontally to only 72% (instead of 82%) because kv is now 520 (instead of 450). That means the same voltage would result in (15.5%) higher main rotor speed, hence the smaller throttle ratio is needed for the same rotor speed (1100 rpm). Conversely, 3.4 A (15.7%) more amperage is needed. While operating at 1100 rpm, efficiency is still slightly past its peak and is a tad lower: 78% (instead of 79%).
Higher amperage is not a problem for either motor or ESC. The latter could have a problem with throttle ratio, though. This is still above 70% but – due to higher battery voltage during operation – throttle opening and throttle-preselection are yet below 70% in cruise flight ("standard" for the ESC). According to YGE, that can cause "unneeded heat losses" (in the ESC and maybe also the motor) and to avoid them they recommend a higher-reduction-ratio gear ("lower pinion") – or lower kv. (Both are equivalent, as is lower battery voltage.) So we already have too high kv here – for the ESC, which has to curtail the voltage even more and has to stand higher amperage.
Review
All telemetry measurements and the calculations based on them suggest that a 700-size 450 kv motor is just about perfect for this helicopter. Bigger size and higher kv seem to be useless, if not detrimental. In the end I went to the motor manufacturer's (Kontronik) website to check with the heliCalc tool offered for free there. It's definitely good for retrieving internal-resistance and idle-current values, which are no longer available elsewhere at the website, and even internal-resistance values of ESCs and batteries. But first and foremost it's meant as a help in choosing a proper motor by trying different ones and comparing the calculated characteristics. So I tried the drive used here.

However, the general free heliCalc version has been used (instead of the Kontronik version, which is free as well) because it has data for my YGE ESC and for a better fitting 4500 mAh LiPo battery. Both versions are restricted in that the rotor fudge factor (PConst/TConst) is fixed (not modifiable). And there is no way to specify blade width.
Differences to my own calculations are the motor's idle current (2.2 A instead of 2.3 A) and the battery's internal resistance (33 mΩ instead of my firstly too high 60 mΩ, but then 38 mΩ as per telemetry data as well as charger).
With the benefit of hindsight, it was easy for me to just retrospectively calculate drive parameters so that they match set rotor speeds and measured amperages. These calculations turned out to be correct (see above). On the other hand, heliCalc tries to forecast amperage and the rest of the parameters by estimating the rotor's power requirements. That seems to be harder than thought.
At least the tool calculates more amperage than measured (31.3 A instead of 26.5 A) and correspondingly too much mechanical power (1214 W instead of 818 W) in the "hover" column. The values in the other columns are off as well. Four of the six gauges are in the yellow (warning) range. And it warns against too high voltage (even though the motor manufacturer specifies 12s LiPo) and too high motor temperature (in the maximum case). At the top of the web page they claim 15% accuracy but the results exceed this limit by a wide margin and just don't match reality.
Maybe the constant-speed rotor feature doesn't work properly, but that is pure guessing and I could as well have misunderstood the tool. In any case, it seems to suggest a bigger, more powerful motor. They warn the user against too low calculated amperage and urge him to measure real amperage before flying. But now it seems they should also warn against too high calculated amperage. I'm glad that I did not look at the Kontronik website in the first place but inquired Mr. Illig from Scaleflying.de about a proper motor and then just bought it.
Estimate
Still in quest of a formal way to select a proper motor, I reverted to simple calculations. At least the first step – determining kV – seems reliable because it mostly depends on well-known (given) parameters. Moreover, kV is most determining for a motor since it tells how fast the motor spins depending on voltage, and – as inverse value – how much torque it produces depending on amperage. Once a kV is selected, the next step is estimating an upper limit of internal restistance R, which determines how much amperage flows depending on voltage. Both characteristics together determine a motor's speed, size, and power.
The first equation employed here is a simple quotient. Its numerator is motor rotational speed, the product of set rotor speed nset and gear ratio ig. The denominator is basically motor voltage. Battery voltage Ub is reduced by the throttle ratio TR defined above. It's 0.8 (80%) here because we assume 44.4 V nominal voltage for the 12s LiPo battery. In operation, voltage is always higher but TR still stays in the demanded 70% to 80% range then.
The square root of maximum motor efficiency ηm max is the factor which specifies how far peak-efficiency rotational speed is below maximum rotational speed. It's used here because we want to operate the motor at peak-efficiency rpm! Peak efficiency of the motor alone is estimated as 92%, but 5% are deducted for the resistances of battery, ESC, and cables with connectors. Another 2% are deducted since motor voltage is reduced by the TR.
The result (464) is much closer to 450 than to the next higher kV (500 or 520) in the Kontronik Pyro lineup of helicopter motors. And it is better to choose a slightly lower kV because that ensures that the throttle ratio TR stays in the demanded range. A higher kV would entail the ESC having to operate also below 70% TR and producing a lot of heat then. Correcting the TR for the selected kV is simple:
Even though the throttle ratio is relatively close to full throttle (100%) now, there's yet enough power reserve because power increases (at constant rotational speed) with voltage squared, that is also throttle ratio squared. Even in this worst case (82.6% TR at 3.7 V cell voltage), motor voltage can be increased by still 21% (to 100% TR) and power increases by still 46% then. And the motor's size will be more than just sufficient because the next higher size step in the motor lineup has to be chosen, so the power reserve will be even bigger.
In the instructions, the helicopter manufacturer recommends a motor with kV 450 and of 750 size. We have reproduced the first part now, albeit only based on the recommendation for 1100 rpm main rotor speed. If we had assumed 1200 rpm then required kV would have been 507 instead of 464. But we commit to kV 450 and have four size-options (700, 750, 850, 900) in the Kontronik Pyro lineup of motors, or even five if we accept also kV 480 (800) what is just as close to 464 as 450.
Also the next step relies on a given figure, the required (average) torque (moment) Mg req (g for gear output shaft) in hover flight. It has been calculated above based on logged telemetry data, which had been measured at 1100 rpm main rotor speed. In the equation for torque, it could be simply interchanged with resistance R. The selected kV has to be used now and the accordingly corrected throttle ratio TR. Now, estimated maximum motor efficiency ηm max and its square root replace the product of resistance and idle current, which is still unknown as long as no motor is chosen.
By definition, resistance R includes the resistances of battery, ESC, cables and connectors. They have to be subtracted out of the calculated 105 mΩ. I had assumed 20 mΩ for the ESC and heliCalc assumed the same. 38 mΩ for the batteries was assumed after repeatedly observing the charger display when charging for the next flight.
47 mΩ is an upper limit for the motor's internal resistance. It allows for just enough amperage to achieve the required torque at the designated throttle ratio TR and with enough power reserve. In the Kontronik Pyro lineup of motors, even the smallest size with kV 450 – 700 – has a lower internal resistance so 750 size is not needed and the power reserve is even bigger. That's why the throttle ratio TR has to be adapted again, now to the smaller resistance R which allows for the required amperage (which in turn is transformed into torque) at a slightly lower voltage. The actual values of Rm and I0 are taken from heliCalc:
This equation for TR is a ratio of voltages. While the denominator is full battery voltage, the numerator is the voltage needed for the required torque. The first term in parentheses is the voltage effective in the motor, that is producing torque by allowing for appropriate amperage. The second term is the "voltage drop" caused by internal resistance and friction, and the third term is the opposing generator voltage generated by mutual induction at set rotor speed.
The actual power reserve can be calculated directly now, not by analogy to the throttle ratio. Half of the equation for torque Mg is canceled in the ratio and numerator and denominator differ only by throttle ratio TR. Again we recognize battery voltage (full as well as reduced by TR), "voltage drop", and generator voltage. The two latter terms are subtracted leaving the effective voltage producing torque. This effective "rest" is surprisingly small, especially the reduced voltage at throttle ratio for hover flight. Accordingly, the power reserve is surprisingly big:
It's even bigger than calculated above. There, I had used the self-measured I0 (2.3 A instead of 2.2 A specified in heliCalc) and higher battery internal resistance Rb (60 mΩ instead of 38 mΩ as per the charger), like for an inferior battery with lower C-rate. The product of both values R⋅I0 is noticeably increased (from 0.198 V to 0.258 V) but is absolutely small so that can't explain the big difference between power reserves. However, the absolutely small increase of "voltage drop" requires a correspondingly small increase of throttle ratio TR (from 0.815 to 0.828), but that means a relatively big increase of effective voltage required to produce hover torque. Now this TR as well as the power reserve are exactly like calculated above:
This does not prove that the calculation is correct since I just "reversed" the calculation above. But it shows how sensitive the calculation – and the drive – is to changes in the given constants. It may well be that an inferior motor, ESC, and battery (each with higher internal resistance and the motor with higher idle current than here) would require a 750 size motor to have enough power reserve. (Once setting full pitch in panic demanded a 3.5 power reserve.) Higher kV would be required for higher set rotor speed only.
Generally I think that a calculation based on specified and assumed values can't be really correct. At least I always adapt, or calibrate a calculation with measured (telemetry) values, like shown above. Still it may be off the mark, for instance when it's cold so the battery has higher internal resistance than usual. And an even bigger problem than calculating the drive is calculating the rotor, that is required torque derived from reasonable rotor speed and pitch angle as well as required power reserve.