Brainstorming Copter Launcher for a High Speed Plane

[Michail_Belov]

Actually, this is a good point, because if you incline the copter backwards, it will start generating backward acceleration which would slow the plane and require more forward thrust for the plane motor. That could be easily solved by making the wing angle with respect to copter bigger, not 10 degrees, but maybe something like 16 degrees, so during climb the copter would be at 0 pitch, but the angle of attack would be still around 10 degrees. However that could result in wing stall during the initial stage of QHOVER.

However, the extra power needed to overcome that thrust due to backward inclination is not that big. For a 10 kg copter and 6 degrees that would be rouighly 1 kilogram. However, we would be flying using partial thrust, maybe 3 or 4 kg, so backward thrust would be around 400 grams. For the plane motor that is nothing, it has a 5 kW motor which can generate several kilos of forward thrust easily.

[Michail_Belov]

STAGE 4: Release, once a safe altitude is reached, the release sequence is initiated by the pilot using some channel and a switch. Lua script would change the pitch forward by some 5 degrees and increase power to some setting which would be equivalent to sharp climb for the copter alone, wait for the copter to reach that pitch inclination (probably around 0.3 seconds), and then issue command to release the plane. The change of 5 degrees means a reduction of Cl by 0.5, i.e. the plane lift will be reduced by that amount. The plane would simply fall.

On the plane side, the release command sent by the pilot (through some switch) would put the plane into FBWA MODE, and once released, the pilot would have normal flight controls for the plane. Here is a small issue, namely FBWA would most certainly try to put up elevator and that could result in increased load on copter motors before separation and possibly a strike after separation. Ideally, the FBWA should be invoked after a short delay of maybe 2 seconds. I am trying to avoid the scripts on the plane side, so I am not sure if that is a big deal or not.

On the copter side, there should be some verification method to confirm that the separation happened so as to go to STAGE 5. One way of doing it would be to have a limit switch which would be triggered by the pin, a safer method would be to have an electric cable which would break the circuit physically once the plane is released. something like a servo extension hanging from the copter and plugged into two pins on the plane side which are short circuited. Once the plane is released and the servo cable is pulled from the pins, the copter will know that the separation is complete.

[Michail_Belov]

STAGE 5: RTL. The copter would fly back and land automatically. Once the separation is detected, RTL MODE will be entered, and the RTL should be set up correctly to assure automatic landing without the pilot intervention. One would have to pay special attention to potential mode changes for the plane which could cancel the RTL mode on the copter, i.e. the RTL mode should be continuously set by the lua script.

Also, the PIDs set up for empty copter would be set by the lua script as well.

Separation confirmation probably should be at least with two independent sources, because a false positive would be catastrophic. A sensor at the release pin AND the cable separation on the plane probably would be good enough.

In case that the separation did not happen, i.e. the plane became stuck, the pilot would fly copter back using the rudder and throttle (reducing the forward throttle would initiate a descent), and then force QHOVER mode manually and land.

[Michail_Belov]

Electronically, a small separate controller, something like Pi Pico could be used. It would communicate with FC lua using serial, and its only tasks would be to generate the PWM pulse for the servo to release the pin and two input switches to confirm the separation of the plane - copter.

[Michail_Belov]

On the plane side, there are very few issues. One is to delay the automatic switch over to the FBWA mode after the release. The other is to limit the power of the main motor and possibly the rate at which the power is changed. Punching the throttle to 5 kW will result in a loop and some interesting moments for the pilot… So, setting the maximum power to maybe 1000 W and power change rate to about 200 W/ per second would be very prudent idea.

[Michail_Belov]

OK. I was doing some hardware work and came up with the following:

[Michail_Belov]

This is my new copter… it has a diagonal of 106 cm, is based on 1 mm thick 30 mm tubing, and 3d printed frame with carbon fiber tow inserts (basically, the printed frame serves as a mold for making the frame which allows obtaining very lightweight frames, I impregrante carbon tow and then wrap it around the groves, which here are 5 x 7 mm both on top and bottom). Weightwise, frame, motors, esc, props, electronics, capacitar bank, cables, etc, weighs 1700 grams. Excluded is the release mechanism, which weighs 200 grams with servo and everything. And the compound 6S battery with total capacity of 8000 mah (Lipo) weighs close to 1400 grams. So overall, 3300 grams, about 10 % over the inital weight budget. The props are 15 x 4, and are very heavy, 45 grams each. I can replace them by 15 x 5.5 EOLO props which weigh 17 grams.

Today I had my first hover test which went OK, got 22 A at 23 V and 3100 grams, about 10 % more than predicted by the Ecalc. At 10 kg, the current should be around 120 A.

To exclude any possibility of overheating I placed the 60 A x 4 ESC directly under the prop tip for cooling. Next days I will try to load the copter and obtain more data on power and temperatures for heavy load.

[Michail_Belov]

The release mechanism also works very well, the spring has a tension of 7 kg, i.e. based on the estimates for pin friction etc., it is good for release of at least 15 kg load. Maybe I overengineered the whole thing here. It has a safety latch, and a cable which pulls the latch (not shown on the photo). I used a 32 gram servo, 4 kg cm nominal, but the force necessary to release is very low, I would say that even 12 gram servo would work here.

[Michail_Belov]

One very interesting use for this copter is a hover relay station for BLOS flights. It can accomodate a 21 DBi patch antenna, 30 x 30 cm, about 600 grams with servo pointing in elevation, and with the retransmission antenna of 14 dbi (6 x 12 cm), again with elevation pointing. The idea is to keep the copter on the direct line between the plane and the GS, so no azimut pointing on copter is not necessary. Copter would have to point to the plane, and if the plane moves to a side, then the copter would follow it on a circle, but always on the line which goes from plane to GS…

As to the RC control, a mLRS 900 Mhz on both the plane and copter would provide sufficient RC control range and data link,for obtaining the plane coordinates to control the relay copter.

The link between GS and copter would probably remain as 2.4 Ghz ELRS (without telemetry).

Another option is to mount a Dragon Link TX onto the copter and DL RX on the plane, and use mLRS for both RC and Data link to GS from copter. But I do not have many spare DL Txs, so not sure if I want to risk it.

For this, however, I would have to change both the motors and props, this is not exactly cheap, I am thinking about something like 26 inch props, and efficient motors, something like SunnySky V5208, so that motors/props would run close to 500 USD.

The motors on the copter as of now are from Hobbyking Xrotor 3115-1050 kV, and I bought the whole set surprisingly cheap for just 70 USD. Hope there is no some catch there… But so far they look good.

The setup with Sunnysky V5208 and 26 EOLO props is about 2.4 times more efficient than the current setup, and I hope that I could obtain a hover time close to 90 minutes with 700 grams load, with the total AUW between 4500 and 6000 grams, obviously with liions.

[LupusTheCanine]

European radio regulations limits are mostly based on EIRP* so unless you are extremely power limited (almost never on drones) there is no benefit to transmiting using directional antenna.

* EIRP - equivalent Isotropic Radiated Power.

[Michail_Belov]

As far as Relay Station goes, from my experience, using a 1.2 W VTX gives an acceptable video for a range of 40 to 60 km., and very good video for 20 to 40 km if transmitted to a 21 dbi antenna and RC305 receiver.

As far as the other part, that is the copter to GS link, the problem could be that a too high VTX power on the copter could affect the reception of the main antenna. So I plan to try to run it at 1.2 W as well, but possibly I may have to reduce it strongly, down even to 200 mW, depending on the real world result.

The normal setup I am using now for copters is a 3W VTX on the copter (usually run at 2 W) and a patch antenna (Mapple Leaf, advertized as 17 dbi, probably more of 14 dbi) on the ground, mounted on a simple tracker with servoes. This gives me very good image to 10 km., and quite acceptable to 16 km., which is realistic limit for safe copter flights in the mountains.

So having the same patch antenna on the ground and on the copter gives me an extremnely good video link. In addition, making the directional antenna and lua script to point correctly to the GS is extremely easy, and weightwise is about 100 grams including everything.

The relay copter mostly would stay close to GS, probably around 2..5km in most cases, but it could be very high, definitively above 1 km and up to 2 km., to clear the mountain ridges. So it would be looking down at an angle of 30 to 45 degrees, and unless you compensate that with an inclined antenna, the video quality would suffer.

[grekiki (Gregor Kikelj)]

I’m thinking, during the vertical climb, won’t the wing generate a ton of torque if it has an angle of attack? I mean the CG is generally at around CM or slightly ahead, but during the climb, the CG of the system will be significantly ahead which might make pitch control difficult. And you can’t even do much with the tail since it has so much less surface area and therefore lift capability.