6S LiPo Battery Value and Capacity Comparison

Flight time depends mainly on the drone weight. In order to get flight times above 1 hour on a single charge (as shown here), the weight has to be minimized. Typically the most dominant weight hereby is the battery.To identify the best battery I created a 6S LiPo Battery Value and Capacity Comparison.

While there are several arguments speaking for 6S, it added agility to the PhoDrone build and fixed the issue of sporadically maxing out motors.

Even though the 12000mAh battery shows the best value and density, its 266Wh are too high for taking it on an airplane as carry-on (FAA limit is 100Wh / battery, with airline approval two batteries up to 160Wh). Due to this reason the 6600mAh (147Wh) has been selected. According to the Frontier airline hotline there is no upfront approval required.

  • Value can be calculated by Capacity / Price)
  • Density can be calculated by Capacity / Weight

The diagram above is sorted left to right by Ranking = Cap /Weight * Cap /Price

See also 4S LiPo Battery Value and Capacity Comparison.

Small Batch Prototyping for 3D Printed Parts

3D printing enables rapid prototyping but lacks production-ready qualities such as smooth surfaces, temperature resistance and food safety. These attributes however can easily be added by combining 3D printing with casting silicone and urethane plastic for small batch production runs.

As long as you don’t need hundret thounsands of parts, this method should get you up and running without paying a fortune for a silicone injection molding setup.

When selecting mold making materials for 3D printed parts, it is important to be clear about the intended material for mold and cast .

Some points to consider are:

  • Silicone such as Dragon Skin can be extremely difficult to stick to surfaces, except silicone itself. 
  • Urethane plastic such as Smooth Cast 310  is exotherm. This means it gets hot while it’s hardening.
  • Some 3D printing filaments are more temperature-sensitive than others.
  • Some 3D printing filaments are easier to post-process than others.

Having this in mind, the proper materials can be chosen basing on the desired material of the positive.

For PLA the following works nicely:

Final PositiveMoldSeperateSealRelease
Urethane PlasticSiliconeMonster ClayXTC-3DEase Release 200 
SiliconeUrethane PlasticSculptex oil based clay XTC-3DUniversal Mold Release

Please note that urethane plastic is exotherm:

  • This requires a non-wax based clay to define the two-part mold. Otherwise the surface in contact with the clay will be full of air bubbles due to the melting wax-based clay.
  • Hot glue is not a good choice for the mold as the glue melts again and the mold leaks. Better use glue + Acryl or Wood + Screws (+ glue)

Applying multiple coats of XTC-3D with sanding (400…1500) in between is highly recommended.

Degasing the silicone and urethane before pouring significantly reduces trapped air bubbles and increases the chance of success.

Other useful items:


1Medicine Cups to mix XTC-3DAmazon
1Mixing Cups for silicone and urethane plasticAmazon
1Craft SticksAmazon
1Safety GlassesAmazon
1Vacuum Chamber for degasingAmazon
1Vacuum Pump for degasingAmazon
1Clay Sculpting ToolsAmazon
1Hot Glue GunAmazon

Long Flight Time (1h), Aerial Photography Quad, 770mm 17inch

RC Groups


The center board has been glued with HK CA glue (HC-50-175). A small frame with screws through the four beam holder holes kept the sandwich layers aligned.
Note: Better do large areas with this glue outside as it generates a fume.
Using nylon spacers and screws, the top and bottom sandwich were assembled.


The beams were machined with a 1/32″ flat end mill and cut to length in the last run.
M3 nylon screws lock the beams and act as breaking point when crashing.


To attach the electronics to the center board, I chose two different approaches. Components which will not be used in every flight have been attached using Velcro. The other components were glued directly to capton tape I put to the center board. The ESCs were wrapped in capton tape to prevent short circuits. I hope the heat dissipation is reasonably low with the 30A ESCs.

Sandwich Glue

  • HK CA glue (HC-50-175)
  • Loctite Vinyl, Fabric & Plastic Flexible Adhesive
  • HK 30min Slow Cure Epoxy
  • 3M General Purpose 45 Spray Adhesive

Sandwich Bend Test

  • 100x10mm x 1mm CFK, 3mm Fill, 1mm CFK

CFK 1mm, 3k plain weave carbon, orientation parallel (0°/90°) to centerline
Fill +/-45°

Bend test showed best adhesion and no separation up to 6kg for HK CA glue (HC-50-175). All other separated well below this (usually at around 3-4kg). However glueing is tricky for large areas such as the center board.


During the first flights, the flight time was reasonably high at a bit over one hour with 1555 props.


Arms folded away during the first crash (I’d love to call it unplanned rapid decent though :-)) and the M3 nylon screws broke as planned, preventing damages to the rest of the drone.

Flight Time

With this setup, flight times above one hour have been achieved, even including the 180g 3-axis gimbal and GoPro.

In the diagram below various test flights for a 15″ 5.5 and 17″ 6.0 propeller were aggregated to show the effect the prop choice has on flight time.

Update: I’ve updated the battery with a 6600mAh 6S and the quad flies a lot more stable now.

Previously the flight recorder showed motors occasionally maxing out during sharp moves or wind gusts.

I’ll do some test flights and check the flight time. First flight was 2500mAh for 15min

Bill of Material

rctimer 5010 360kv
Multistar Lipo Pack XT90
Flyduino KISS
1760 Folding Props
Flight Controller and Power Module
GPS witd Compass
Damping Balls
Folding GPS Antenna Base/Black
Lipo Voltage Checker
FPV Camera
Telemetry Radio
Skyfpv Video Switch
LC Filter
FPV 3 Axis CNC Metal Brushless Gimbal With Controller For DJI Phantom GoPro 3 4
Silicone Wire
FPV Antenna
Goggle Battery

4S LiPo Battery Value and Capacity Comparison

Flight time depends mainly on the drone weight. In order to get flight times above 1 hour on a single charge (as shown here), the weight has to be minimized. Typically the most dominant weight hereby is the battery. To identify the best battery I created a 4S LiPo Battery Value and Capacity Comparison.

  • Value can be calculated by Capacity / Price)
  • Density can be calculated by Capacity / Weight

The diagram above is sorted left to right by Ranking = Cap /Weight * Cap /Price

See also 6S LiPo Battery Value and Capacity Comparison.

Raspberry Pi 3B+ Enclosure 3D Model


Cloud Server

The above setup is used for a Nextcloud setup. A Nextcloud is a self-hosted productivity platform and allows you to host your cloud data such as pictures of your phone yourself.

It is a full replacement for DropBox and Google Drive for me, with the benefit of protecting my privacy. Additional features enable video chat and many other fun things.

They installation script can be found on github.

Backup Server

The above setup is used for a backup server which regularly backs up all Raspberry Pis in a rolling backup on a daily and monthly basis. The installation script can be found on github.

I have several PIs installed in my home. One of them acts for example as Nextcloud server and resides in a DMZ.

In such a setups you don’t want servers to actively push into your private network for security reasons.

So you have another server/PI residing in the private zone pulling the data such as pictures and logs in.

The backup PI uses CRON jobs and scripts to do a backup every day and month. It hereby keeps specific data for 12 month in a rolling backup.

Bill of Material

1 1TB HDDAmazon
1 Hard Drive AdapterAmazon
1 Raspberry Pi 3Amazon
1 Memory CardAmazon
1 Power supplyAmazon
1 Y-Cable ExtensionAmazon
1 #4×1/2 Pin Head Phillips Zinc ScrewsHomeDepot
1 #4×3/8 Pin Head Phillips Zinc ScrewsHomeDepot

Not used by me but recommended in setups with a fan: Nocturna Fan or WINSINN  Fan.

If you don’t need a hard drive, you might find this case attractive as it includes a fan already.

Parabola Tool

This tool will help you to design a parabolic trough and give you the information you will need to build a simple one on your own.
Download ParabolaTool_v190

ParabolaTool v1.9:

Compiled without source code guarding. Exe will start correctly on machines without additional DLLs.

ParabolaTool v1.8:

Export functions are available now.

ParabolaTool v1.7:

By ” Print 1:1 ” functionality you are now able to plot the parabola at its real size and to use it as a template for cutting the beams or to cut a solid beam.

Reflector Width:

You can enter the width of the reflector here. (Real/plain width of the mirror)

Supporting Beams:

To stabilize the parabolic shape, it is useful to place some supporting beams along the curve. The middle is always fixed to x=0 and the outer one to the highest point of the reflector.

Focal Point Height:

Parabola Tool calculates the curve by selecting the damping of the parabola. The height of the focal point is the most important variable you have to select to get good results. The higher this value is, the lower are the shadowing effects. Keep in mind that this will also increase the vulnerability to inaccuracy. A good starting point should be 30cm-1m for 1m x 2m base frames and 3/4 ” absorbers.

Parabola Width:

Is the shortest distance between both ends of the reflector. You can use this as a starting point for the width of the base frame.

Parabolic Trough

Currently, I’m building a low cost 1000mm x 2000mm parabolic trough (used for example at Desertec) for experimental purposes, which is split into four 1000mm x 500mm segments. Polished stainless steel is used as reflector material.

The tool used to calculate the support structure can be found here.

Up to now, there are no building instructions here, but I will update the site from time to time.
If you are interested in more technical drawings please write a mail. If there are enough requests, I will upload further material.

Update 2011-05-11:
A technical drawing (1:10) is now available: parabolic trough_technical_drawing

Update 2011-05-08:
Today I did a series of tests, which should view the capability of the new mirror mounting.

Here are the results:
To get 700ml water to boiling temperature, one of the four modules (500mm x 918mm) took 7min.
If the absorber is preheated, the water already starts boiling in less than 5min.

The next test setup was filled with cooking oil to determine the maximum temperature of the module.
The highest temperature was 138°C after 4-5min with the poor, quickly black lacquered aluminum tube.

In my opinion it’s not that bad and I will continue to optimize the absorber.

Further steps are:

  • Try to get a ” cheap ” vacuum absorber tube. (DONE)
  • Try different materials for the tube.
  • Aluminium has good thermal conductivity (235W/(m·K)) but it is cooled down too fast by the fluid inside the tube and shadowing effects.
  • In my opinion it woud be better if the tube acts as a capacity and could store the heat, which could then be transmitted more steadily.

See also the Parabola Tool if you want to build your own.