Professional Documents
Culture Documents
First draft
This document is about the design development of a man carrier drone. It sketches considerations on
how it could be configured, how it can be made. It is a contribution to an open source knowledge
development and can be used by other teams to make their unique choices. Herewith we as a team
and particular my self would like to thank all that have shared their knowledge by explanation on
skype, by youtube movies, by whatsapp, etc Also interesting to see that knowledge came from all
around the world: from USA to China, from India to Norway to Uganda. All in all quite a variety of
backgrounds and cultural differences.
Again thanks to the team and its supporters, Winfried Rijssenbeek, coordinator of The
HummingManProject, aiming to make a man carrier drone.
Part of my internal fun is to make shapes of beauty, I did that already as a 6 year old. Making
drawings of future cars. Another part of me is analytical: I love spreadsheets, phyisics and aero
dynamics. This project has it all!
Configuration
The making of a drone or multicopter that can lift a 80 kg is quite a challenge. It can be a wide craft
or a compact one. Or in between. Soon we realized that more compact results in lower efficiency and
hence higher energy needs to have the craft run. With a large craft we can have good efficiency low
energy requirements. Why is that so? It is related with the air stream or mass displacement of air.
Simply a craft that replaces air only in a rotor surface of e.g. 5 m2 has to push air mass trough that at
high velocity, so to compensate for the gravity force. Eg a craft weighing 150 kg with one rotor of 3 m
needs an average speed of 18 m/s.
Parameter Unit Value
Now if we take a HPV Sikorsky prize craft with Air mass density Rho kg/m3 1.25
No of rotors no 1
4 rotors of 12 m, we would need for the same
Diameter of rotor D m 3
weight Surface of rotor area A m2 7.068583
To speed up the air you need a lot of energy: it Weight of craft G kg 150
goes by the speed to the 3 power and by the G N 1471.5
surface. Finally a extra bonus is larger size G = .5 *V^2*A* Rho
props are more efficient due to viscosity Air mass speed at rotor m/s 18.3
diameter
effects. So the result is that the compact craft km/hr 65.7
needs e.g. 36 kW for hovering whereas the
HPV one only an estimated .8 kW
Parameter Unit Value
So we had to question our selves what is the Air mass density Rho kg/m3 1.25
good size? Well we decided that a good size is No of rotors no 4
the size that is compact can be trailered and Diameter of rotor D m 12
Surface of rotor area A m2 452.3893
really makes a this craft a 2.5 m by 3.75
Weight of craft G kg 150
meter. It can also be foldable. This will G N 1471.5
indeed result in losses in efficiency and high G = .5 *V^2*A* Rho
power requirements. We would think that Air mass speed at rotor m/s 2.3
this compactness is the unique feature of our diameter
km/hr 8.2
craft which can just hover and speed up like a humming bird.
Having decided that the craft needs to be compact, which indeed goes at the detriment of efficiency,
we now have to proceed with the design configuration.
Now in the configuration of a more rectangular surface of lift of our craft, we can use 3, 4 or 6 or
more rotors. It would be best to use a lower number of bigger rotors as this relates with the
efficiency. 6 rotors would be ideal for making the rectangular surface and in case of fall out, one still
can manage. It the case of 3 and 4 rotors, there is a unrecoverable balance. So we can start with 6
rotors or more and determine the parameters.
Weight of craft
It is important to count with a ultra light craft as the payload is 80 kg. We consider a 150 kg as total a
good max. The reason also being that a craft of more than 150 falls into another category as the so
called utralight airplane. We therefor need to have extremely low weight but safe components in
order to keep weight down so to add to the battery volume that in fact determines the flight
duration.
Propellers
E motor
Electronic speed control (ESC)
Cables
Belt reductors
Batteries
Air frame
Flight Control System
Propellers
Propellers depending on size between 0. 340 (32 inch diameter) and 1.1 kg (51 inch diameter). There
is a huge number of props available in the market from standard to special design, etc.. Materials
today are either wood or carbon fibre. Alum or other are not used in the Ultra light propulsion.
Xoar http://www.xoarintl.com/rc-propellers/gas-props/PJD-Sport-Prop-Type-A-Laminated-Propeller/#des
Bambula http://www.bambula.cz/
ZDZ http://www.mejzlik.eu/extra-light/
Menz prop http://www.big-planes.de/profil/jc-super-prop/index.php
JC prop http://www.big-planes.de/profil/jc-super-prop/index.php
CFK prop http://www.big-planes.de/profil/jc-super-prop/index.php
Vess http://www.vesspropellers.com/servlet/StoreFront
Seidel http://www.seidel-props.de/index.php?option=com_content&task=view&id=26&Itemid=57
APC http://www.apcprop.com/v/index.html
TCF http://www.hobbyking.com/hobbyking/store/__11967__tcf_27x10_highprecision_carbon_fiber_sport_propeller.h
tml
Different
brands http://www.amr-rc.com/products-brand/engines-accessories/propellers/bambula/
Falcon
Hobby http://falconhobby.en.alibaba.com/product/811711058-218365617/Gas_beech_wood_propellers.html
PT-Models http://www.troybuiltmodels.com/items/PT3210.html
Biela http://www.troybuiltmodels.com/items/BP3616-3.html
HELIX http://www.helix-propeller.de/aircraft/7_2_propellerblaetter.php
Parameter Unit A B C D E F G
Power kW 8.00 15.00 12.00 8.78 8.82 13.00 25
Speed rpm 8000 2500 3000 2330 5390 4900 3300
Costs euro/unit 340 1200 1200 1700 1050 1700 2500
Voltage V 55 63 60 50 49 80 120
Amps A 145 238 250 175.5 180 163 208
Efficiency % 90% 90% 90% 90% 90% 90% 90%
Weight kg 2.52 4.1 4.1 3.3 2.5 2.5 5
Ratio Euro/kW 43 80 100 194 119 131 100
Ratio rev/kW 1000 167 250 266 611 377 132
Ratio Euro/kg 134.92 292.68 292.68 515.15 420.00 680.00 500.00
It can be concluded that in value for money A is a clear winner, with B Ret 30 5 a second best. Likely
in terms of kg and low speed the B and C are good alternatives. Thus we opt for the A.
There are ESCs for the RC community and industrial ones and more. The ESCs are in different ranges
from 50 to 300 Amp and normally a brushless motor is given a combination with a ESC. Eg the
rotomax 150 is paired with the Fat Boy ESC of Turnigy Hobbyking.
Cables
All the current from the battery passes through the ESC and there goes into the motor. The amps
passing through the cables require a thick cable if not the losses can be considerable. Again this
would tend to take higher voltages if this can be opted for. Cables are best alum. The reason simply
that although alum has a 56% lower conductivity it has a much lower weight too (2.7 iso 8.9 kg/l) so
it is wise to use alum conductors of bigger size (1.33 The copper wire diameter)! Wirelength should
be low: max 2 meters. Positioning of the ESC is best closest to the e motors. For the cables with a 2
meter length we aim losses less than 3 %. So with a 10 AWG copper conductor (5.6 mm2) and 2.5 m
we have such result with 55 V and 100 Amp. (http://www.calculator.net/voltage-drop-
calculator.html?material=copper&wiresize=3.277&voltage=55&phase=dc&noofconductor=2&distanc
e=2.5&distanceunit=meters&eres=100&x=46&y=15)
With high speed motors such as the rotomax 150 that have relatively low power output, a reduction
of the speed can give a better match between the propeller and the motor. This requires some
explanation. If we have a 150 kg to be lifted by 6 rotors we need 25 kg of lift per rotor and a max of
50 kg to cover for redundancy, balance, wind, etc.. So we need to match propeller and motor to the
best efficiency aiming for the motor not to exceed 10 kW. This can be done using software or
formulas to calculate trust of propellers based on their rotational speed.
10000
A disadvantage of a belt redactor is the weight enhancement in the belt drive itself but also due to
the more weight prop. Another uncertainty is if the prop behavior is sufficiently agile to respond to
the balancing power requirements. And possible vibrations. For calculation programs, Gates
(http://engapps.gates.com/IQInstallService/Install/DF-Pro%20Install%20311.exe)
For this craft however we intend to use direct drive system with 12 rotors in 6 contra rotating modes:
each prop is to be attached to a single e-motor, but in a contra rotating direction.
Batteries
This is the essential component of the craft: it is the energy source that must be able to provide the
draw of Amperes to the motor as well as the required flight time. Flight duration is the major issue
we have. We want to have at least 10 min of flight duration, which means with a 4.5 kW for 25 kg
trust per motor and 6 motors we have 4.5 * 6= 27 kW or if we have 12 motors with 1.9 kW reach to
produce 12.5 kg of trust we need 1.9 * 12 = 23 kW. If we have an motor efficiency of 90 % we thus
need 30 to 25 kWh of one hour for both options respectively or if 10 min flight we are requiring a
4.2 to 5 kWh battery capacity at minimum.
Appart from the capacity requirement we also might have an issue with the large power draw.
Batteries have limits on the amperes you want to draw out of them: Fortunately Lipo batteries are
quite exceptional in this respect: they can release a lot of energy in a short time. However we will
have to calculate this aspect.
Finally the batteries need to supply a required voltage and the Lipo batteries that are best in both
energy and power density (Wh/kg and W/kg) have 3.7 volts per cell. They work with S numbers: 4S
Lipo thus means 4 * 3.7 = 14.8 V and of course when discharged this voltage drops slightly.
Finally a last aspect is the combination of batteries and location. Batteries when discharged can
produce heat and this heat needs to be released.
Air frame
Max weight aimed at is 5 kg. The airframe is to be strong and rigid. It should be able to withstand at
least 3 G. It can be made of alum 6061 tubes (e.g. 1.5 Inch and 1.2 mm wall), or from carbon fiber
tubes of different sizes. It would be interesting to have curved shapes, but this requires more
techniques and might make certification of the tubes/craft difficult as the bending might reduce
material strength.
As concerns redundancy, we plan to have a multiple sensor navigation sensor redundancy concept
(NAVKA-concept) in terms of:
However it is important to make best efficiency on this option to have a good flight duration (10 to
12 min). For this the think that 6 pods are required. If we take the 150 cc rotomax we cannot use this
with 6 units as it requires too much power. Each motor would then need to produce 50 kg of trust
and this is not possible with such motors (required 20 kW). We need to have more that 6 units, so 8
or 10 or 12. We again can do the exercise with more motors: e.g. 10 motors requiring 30 kg trust
per motor. This will require ca 7 kW/unit, which can be done. But better is even 12 so that a
completely balanced system is made. This is important for the balancing control as well.
On each pod we can mount 2 brushless electric motors, in a contra rotation mode. Therefore we
have 12 e-motors requiring 25 kg of trust per unit. Looking at this graph below of a prop 32 x 12 we
see the below. Using the Adamone software (www.adamone.rchomepage.com/calc_thrust.htm)
with corrections based on similar props, trust and energy are given below.
Rev per min Rev/min 2000 2500 3000 3250 3500 4000 4500 5000 5100
Power required by prop Watt 0.6 428 839 1455 1853 2318 3470 4953 6808 7228
(Kp) 9
Trust delivered (Kt) Gr trust 1.1 4477 7153 10491 12411 14501 19195 24581 30668 31970
We can see that 25 kg of trust per motor, needs 4.9 kW. For hovering we need 12.5 kg and the mech
power for the prop is 1.9 kW. Of course it is 12 of such e motors, thus requiring for the hovering 22.8
kW. We have used this same program for many prop sizes: see spread sheet. 28 x 12, 30 x 12.
Speed of the motor Rev/min 2000 2500 3000 3500 4000 4500 5000 5500 6000 6500 7000 7500
Power delivered to prop Watt 2842 3552 4263 4973 5684 6394 7105 7815 8526 9236 9947 10657
With the motor and propeller together we can see that the prop required power will exceed the
delivered power of the motor at ca 5000 rev per min.
15000
10000
5000
0
0 2000 4000 6000 8000
Revs per min
Airframe design
The airframe design is one of the most beautiful activities where all aspects come into play: strength,
minimum weight, but also looks. The frame definitively should look attractive, however for this
prototype we only need to show proof that it can be done. For the frame we also can set a weight,
strength and vibration target: e.g. Weight max 5 kg, strength eg. 3 G, vibration rotors and motor
frequency not coinciding with the own frequency.
Our most beautiful frame is shown below. But we will have to see if it can be realized as the bending
of aircraft tube alum is normally not recommended.
So with a 12 rotors we will have other wiring configuration. Also we plan at the start to have the
rotors in the middle section placed more apart. For the 12 rotor frame with only straight tubes we
can follow a straight version of this frame or make it slightly more 3 dimensional.
Battery design
Battery design requires a number of considerations but foremost is kW per weight and kWh per
weight. We calculated 2 battery types for this: one with extra kW/kg power due to nano part
technology and the other one standard. As can be seen the nano part system allows less batteries to
be used 26 units. However for the duration both produce equal energy and have a need for 68
batteries. In the below table we can see that the first calculation is to find out how much energy is
required: This comes to 5066 Amph at the working voltage of 55 V. Then we need to see how much
packages of 4 S cells we need: on the safe side we would take 16S, so 4 packs. Then we proceed to
check the requirement for the Amps that can be delivered as it is a short time that the batteries are
drawn. Here we see a substantial difference between the first type of batteries and the second (nano
tech). The nano tech batteries can deliver much more amps so are more interesting. Finally we need
to calculate the requirements in flight duration so to devide the amph required by the flight duration
by what is in the lipo cells. So in this case 92 Amp by 5 . This results in ca 18 units that again need to
be multiplied by 4 to get to the required working voltage.
Now there is a minor adaption we need to do. The 4 batteries of 4S produce 16 S and we have per
rotor point 2 motors and 6 rotor points. So the values of batteries need to be devided by 4 and by 6
to make good pairs so 24 batteries would be available per rotor point or 48 per motor, but we need
68 so if we would have 3 parallel sets we would need 72 batteries. If we make a central battery
system we can do with 68 as this is dividable by 4. We should understand that the above calculations
are done for the average hovering flight, not for redundancy. But with the selection of the nano tech
batteries we have that build in with a factor 68/26.
The above shows that the batteries are the essential component with most weight. The battery
location is to be ventilated.
The making of the craft starts with the ordering of the components and testing them. The testing is
done in many steps.
Eg the frame testing: first the connectors should be tested; the tubes connectors glue needs to be
tested with the connector. Once the frame design has been made we can test it by simulation. Then
after building we can test it by adding weight (static testing)
Eg the motors and batteries and ESC; here we can see the duration tests of a single motor prop ESC
and battery unit. If these work according to anticipated we can later test all at once.
Eg the propellers trust can be tested by the pod configuration of it: in the contra rotating mode.
Amps (A), trust (kg), V battery, duration, vibration peaks, revs min, etc.. can be measured
The control system can be tested by a connection to a smaller system, but ultimately needs to be
tested on the craft.
The craft is made in a clean environment with good climate control. This to avoid condensation on
components or effects on the parts.
The craft in its entire shape will first be tested in a building. It will be connected to the ground such
that hovering capability can be tested. The weight of the payload is not a person yet. Ultimately
when all tethering diseases have been solved, an open air test can be done. Again in careful steps.
Remember safety is a prime concern.
Notes
The craft calculations can be found in a spreadsheet attached.
Flight control and navigation for scalable and arbitrarily dimensioned UAVs
and manned multicopters article
J. Zwiener1 and R. Jger1
1Hochschule Karlsruhe Technik und Wirtschaft (HSKA), Karlsruhe, Germany
E-mail: jan.zwiener@hs-karlsruhe.de
E-mail: reiner.jaeger@hs-karlsruhe.de
The developed flight control package (algorithms, libraries and hardware) developed by the
authors can be used for the navigation and control of multicopter UAVs (see fig. 1) and
manned multicopters (see fig. 2). The multicopters are scalable in respect to their purpose of
use and their
Size
Payloads
Sensors equipment
Possible scenarios are:
3D mapping and geosensing
Film industry
Search and rescue of people
Agriculture UAV
Facility management and monitoring UAV
Wild-life protection
Transport UAV
Fire-Fighting air vehicles
ABC sensing UAV for emergency event
Environment-friendly and silent air taxis
Making use of further developments of the NAVKA R&D team, the UAV and manned aircrafts
are equipped with automatic starting and landing. Further sensors carried by UAVs
(cameras, laser scanners, radar and others) can be geo-referenced by using the navigation
state-vector of the navigation. The mathematical model and concept of the navigation
algorithms are open for the integration of further sensors, so indoor flights; flights in tunnels
etc. can also be enabled.
The main advantage of the developed flight control is the redundant sensor configuration,
both for the navigation (mainly GNSS and MEMS sensors) and for control the configuration.
The availability of the navigation state is essential for a safe flight control. Therefore the
sensors are treated in the mathematical model, algorithms and software as being distributed
in a redundant configuration on the aircraft; see [1] and [2].
References
[1] Jger, R. et al., SIMA-Raw Data Simulation Software for the Development and Validation
of Algorithms for GNSS and MEMS based Multi-Sensor Navigation Platforms. In: FIG Working
Week, Rome, Italy, 2012.
[2] Zwiener, J. and Diekert, J., Multisensorfusion zur autonomen Navigation und
Objektgeoreferenzierung. In: Geomatik aktuell Przise Navigation und mobile
Geodatenerfassung Out- und Indoor Band 7, Reihe B, Hochschule Karlsruhe Technik und
Wirtschaft, 2012. ISBN 9783890631066.