DJI Mavic, Air and Mini Drones
Friendly, Helpful & Knowledgeable Community
Join Us Now

Building a smarter Mavic Battery

gigelu

Member
Joined
Jul 20, 2017
Messages
22
Reactions
17
Age
54
Hi!

I intend to start developing a better Mavic battery from ground up and licence it as open source. This will include PCB, firmware and battery case. I could not find any other similar projects therefore I have started my own.

Project can be found at: github.com/kopern/smarter_mavic_battery

I welcome volunteers to help me on this project as my time is limited and I am not likely to be able to finish this project by myself.

The improvements will be:

1. We will use a higher density battery than the ones used by the Mavic battery:

The original Mavic battery is using LI-ION Polimer cells with a weight for 202grams and provide 43.6 Watts of power. This translates into 215Watts/kg. I was able to fly 15km with this kind of battery and still have 12% energy left)
LG 18650HG2 battery has 11.5Watts in 48grams which translates to 240Watts/kg, which is 11.5% improvement in capacity for the same weight (16.72km range should be achievable)
NCR18650GA battery has 11.9 Watts in 48grams which translates to 247Watts/kg, which is 14.8% improvement in capacity for the same weight (17.22km range expected)
300Watt/kg batteries are expected to be soon avaialble. These would result in 39.5% improvement therefore we should be able to fly 20.9km)

Technical considerations:

Due to the low voltage of the LG HG2 / NCR GA batteries, we cannot use 3 batteries in series. Logical would be to try to use 4 batteries in series but at the moment I am not sure if Mavic can work properly with the high voltage provided by these batteries while fully charged. My intention is to use a DC/DC switching converter to keep the voltage in the parameters of the original batteries.
This will offer us the possibility of using 5 or 6 batteries as well to increase the flight time, although the weight will be increased.
I expect to test a 4x2 battery configuration as well in order to provide a backup should one of the battery chain fail during mid flight but this will increase the weight of the drone considerably (184 grams).
We could also look into using 3x2 battery chain but we need to provide a step-up converter in order to keep the drone flying for the whole voltage range of the battery.

2. I intend to add at a later date a gps tracker. We will use this tracker to transmit the GPS position of the Mavic drone in order to be able to recover it should it fly away for any reason.
As a technical solution I find inconvenient the GSM transmitter due to the requirement of 3rd party services and GPRS subscription. My intention is to build from ground uo a tracker which uses LORA technology instead of GSM.
The chip that I inted to use is SX1276. This chip is able to transmit data up to 48km line of sight, without the need of a GSM transmitter or GSM network.

3. All hardware and software will be open source therefore the cost of this battery will be considerably lower than the original battery.
 
Regarding the case, my intention is to allow the 18650 batteries to be removable. I am not happy with the idea of spot welding due to the high cost of the welding machine. At the same time, the manufacturer of the product will not need to go through the pain of complying with shipping regulations and restrictions of a product that contains batteries as the customer can purchase the batteries by himself/herself.

If you are interested to contribute, I need help with:

1. reverse engineering the protocol of the mavic battery (maybe it can be extracted from the mavic firmware?)
2. write the firmware for ESP32 in C
3. design the a printable case for 5 x 18650 batteries to fit in the Mavic.

Please PM me if you are interested to contribute. Thank you.
 
Last edited:
Please find attached pictures with the batteries' weight. Using 5 x lg hg2 batteries, the weight of the aircraft is increased with only 28 grams which IMO is negligible. We should be able to fly 20km easily and using the new propellers from Mavic Platinum this distance should be even longer.
 

Attachments

  • 5xlg hg.jpg
    5xlg hg.jpg
    286.6 KB · Views: 262
  • original battery.jpg
    original battery.jpg
    280.5 KB · Views: 260
... if I use a step down converter, I won't be able to achieve 100% energy conversion efficiency; 92% is the best I can hope for. This will cancel the advantage we get by using a higher density cell :-(

Maybe I should try using 4 cells in series first. Hopefully I will not upset my Mavic ....
 
Great project ! Unfortunately can’t help you.... but I Follow the evolution !

Good luck and keep us posted of any progress !
 
I have discharged the original battery and the lg hg2.

The results are:

lg hg2 - 45.2g, 9,986mah 221 w/kg
original - 67.3g, 12,192mah 181 w/kg

I have calculated that with:
4 cells, we'll fly 16.74km
5 cells, 18.69km
6 cells, 21,72km
7 cells, 23.38km
8 cells, 28.66km

4 cells and 8 cells, do not require dc/dc conversion therefore the flight range is improved compared with 5,6,7 cells.
 

Attachments

  • batt compare.png
    batt compare.png
    3.2 KB · Views: 138
  • flight estimation.png
    flight estimation.png
    10.6 KB · Views: 134
  • Like
Reactions: Dw911 and lolo780
Can these cells take the current draw from the Mavic and continuous thrust in sport mode? Or do they get pretty hot? I think that discharge is about 10A per cell max with this type of battery, but I'm not sure what the mavic draws under max thrust... from previous experience, I'm guessing that is not enough current and you need closer to 20Amps ish.

How did you get on?

Ta
 
Last edited:
you can get 18650 cells that are good for 30amps draw each with peak of 40amps... its usually a balance between max amps draw and capacity but 3000mah cells with 30 amp draw are pretty common
 
I have feeling that DJI will go down this route...
Higher density batteries (mAh), higher voltage configuration maybe 4s. Lower kV motors. Current limitation of ESV/motor thrust output to ensure batteries are not heavily taxed. 5 - 10 mins more flight possibly traded against slightly faster flight.
 
Regarding the case, my intention is to allow the 18650 batteries to be removable. I am not happy with the idea of spot welding due to the high cost of the welding machine. At the same time, the manufacturer of the product will not need to go through the pain of complying with shipping regulations and restrictions of a product that contains batteries as the customer can purchase the batteries by himself/herself.

If you are interested to contribute, I need help with:

1. reverse engineering the protocol of the mavic battery (maybe it can be extracted from the mavic firmware?)
2. write the firmware for ESP32 in C
3. design the a printable case for 5 x 18650 batteries to fit in the Mavic.

Please PM me if you are interested to contribute. Thank you.
The only significant joining compound would be the IC chip/circuit involving the lock/hibernation and deprecated states but switching to lithium ion from polymer we wld have a decrease on response time per the quick discharge of the polymer Batts I like the idea I'll look deeper into my circuits and try to help any way I can
 
The only significant joining compound would be the IC chip/circuit involving the lock/hibernation and deprecated states but switching to lithium ion from polymer we wld have a decrease on response time per the quick discharge of the polymer Batts I like the idea I'll look deeper into my circuits and try to help any way I can
I'm currently still trying to use my batteries I never got to fly all 3 Batts were failed. I'm trying to bypass but the code is impressive by dji but there's a way...always is
 
Lycus Tech Mavic Air 3 Case

DJI Drone Deals

New Threads

Forum statistics

Threads
130,985
Messages
1,558,595
Members
159,978
Latest member
James Hoogenboom