Hello,
I was doing some studies on how to use the 18650 on Mavic.
Battery rating:
According to the bat_current values from the flight log, it looks that Mavic usually needs around 8Amp. Maxumum it needs is 25Amp and I was able to use so much by increasing the maximum speed of the drone.
I have studied the battery model LG18650HG2 as it looked to me to be the best for our requirements. It can provide easily 20Amperes therefore it is suitable from this point of view (discharge rate).
Flight time can be calculated as follows:
Note: Referring to the current stored (mAh) in the battery is not correct. The drone uses power to stay in the air, not current. Power = tension * current. This battery does not have the tension (voltage) constant as with other Lithium battery chemistries, therefore the power it stores is less.
For example, if we would ignore the voltage drop, we would expect to have 3.8v * 3A = 11.4Watt.
In reality, this battery will provide 9.8Watts @10A therefore a 3S chain will provide 29.4 Watt.
DJI's battery has 43Watts @ 240grams
If we replace their battery with 3s2p we will have 58.8watt @ 300grams. We would expect 36% longer flight time if the weight would be the same. As we are 60 grams heavier (8% of the total weight) we would probably use 8% more power therefore we should achieve 33% longer flight time. This is gross calculation because the motor's efficiency starts to drop from 60% speed which can further impact our flight time.
Therefore, we be able to fly from 21 minutes -> 27.9 minutes
or distance from 13km -> 17.29km
The problem with LG's battery chemistry (H-NMC) is the voltage of this battery will reach 3.5Volt around 1000mA used @ 10A and 3.3volt at 1800mA. At this point Mavic will decide that the battery is depleted and end the flight, even though we still have 11.52watts left (39% of the battery).
I'm thinking that there are 2 quick solutions to this problem:
Solution 1: We might be able to configure the drone to fly with his battery. We can do this by changing some configuration parameters. I'm not sure if this can work even if we convince the drone to try because the motors might need higher voltage.
I suggest to review the following parameters:
g_config_bat_config_emergency_voltage 3000 -> 2600
raw_bat_level_1 3960 -> 3250
raw_bat_level_2 3600 -> 2800
g_config_bat_config_bat_limit_limit_vol_min 3300 -> 2700
g_config_bat_config_bat_limit_limit_vol_max 3700 -> 3500
Solution 2: We can put 4 batteries in series. We can see in the datasheet that the voltage drop on this battery is almost immediate therefore even if we have 16.8volts when is fully charged, it will drop very soon to 14.8Volts. You can review the file hg2.pdf (first entry on google) to visualise the voltage values. 4.2*4 = 16.8v is normally really high if the drone expects 11.4v. At 3.7v each cell we would get 14.8 which is still high but it might work. Looking over DJI's products, their E300 ESC supports 4s for a 3s drone. Granted, this is not the ESC used in Mavic, but I was unable to find the ESC datasheet for Mavic. However, I strongly doubt that they would design an ESC that will burn at 16volts.
Charging the battery at a lower voltage is not ok, because we are losing the whole advantage of power density of this cell therefore if we don't want to charge the battery at 100%, is better to stick with other chemistries.
If we put 4s2p, we'll get 78.4Watt but an extra 166 grams : 82% more power, 22% more weight. This should get us 50% longer flight time but we are getting dangerously close into the motors inefficiency zone.
Therefore I expect to get from 21min -> 31 minutes or from 13 km -> 19.5km.
There is a 3rd solution but is more difficult to implement:
It involves adding a DC/DC converter which costs money, weights a little as well and has a 87-93% efficiency. We could put any reasonable number of batteries with this converter as we can provide whatever tension we want. For example 5 batteries would get us the same weight and the original Mavic but after the conversion loss we'll end up with 9.8*5*0.87 = 42.63 watts which is less than the original battery. The advantage here is that we could put 6, 7, 8, 9 batteries and still provide the same output voltage.
Note: I haven't tried any of the above. This is just theory (educated guess). I am not sure if the reduction of flight time calculated above is correct. If it is not correct, we'll get longer flight times but I strongly doubt that we will be able to achieve 52 minutes of flight time.
If anyone tries to put 4 batteries in series, please let us know, so we don't risk burning too many drones....
A comment regarding what I've seen people to do here: Use in parallel of batteries with different chemistries. I don't think this idea is wise: As we noted, the H-NMC battery will quickly lose tension. The other battery connected (Li-Ion) is not losing the tension at the same rate; LiIon has a tension curve which is quite flat during most of the power provided. What will happen is: The Li-Ion will start charging the H-NMC, but it will do so at an unregulated current. Both are designed for high current output and the transfer of the current will be at the rate that H-NMC can absorb power. This will result in batteries getting hot (high current transfer tends to do that because of the internal resistance of the battery). The drone might even get starved of power if the Li-Ion cannot charge the H-NMC and power the drone in the same time. You could very well crash your drone doing this. If you don't crash your drone, I am sure that you'll damage the batteries. Another problem with this is the heat generated is from the power that we wanted to use in order to fly the drone, power that we now waste. And finally, note that H-NMC can only work with this drone configuration until 3.3volts therefore you can only use 60% of the power which it stores.
Regards!
I was doing some studies on how to use the 18650 on Mavic.
Battery rating:
According to the bat_current values from the flight log, it looks that Mavic usually needs around 8Amp. Maxumum it needs is 25Amp and I was able to use so much by increasing the maximum speed of the drone.
I have studied the battery model LG18650HG2 as it looked to me to be the best for our requirements. It can provide easily 20Amperes therefore it is suitable from this point of view (discharge rate).
Flight time can be calculated as follows:
Note: Referring to the current stored (mAh) in the battery is not correct. The drone uses power to stay in the air, not current. Power = tension * current. This battery does not have the tension (voltage) constant as with other Lithium battery chemistries, therefore the power it stores is less.
For example, if we would ignore the voltage drop, we would expect to have 3.8v * 3A = 11.4Watt.
In reality, this battery will provide 9.8Watts @10A therefore a 3S chain will provide 29.4 Watt.
DJI's battery has 43Watts @ 240grams
If we replace their battery with 3s2p we will have 58.8watt @ 300grams. We would expect 36% longer flight time if the weight would be the same. As we are 60 grams heavier (8% of the total weight) we would probably use 8% more power therefore we should achieve 33% longer flight time. This is gross calculation because the motor's efficiency starts to drop from 60% speed which can further impact our flight time.
Therefore, we be able to fly from 21 minutes -> 27.9 minutes
or distance from 13km -> 17.29km
The problem with LG's battery chemistry (H-NMC) is the voltage of this battery will reach 3.5Volt around 1000mA used @ 10A and 3.3volt at 1800mA. At this point Mavic will decide that the battery is depleted and end the flight, even though we still have 11.52watts left (39% of the battery).
I'm thinking that there are 2 quick solutions to this problem:
Solution 1: We might be able to configure the drone to fly with his battery. We can do this by changing some configuration parameters. I'm not sure if this can work even if we convince the drone to try because the motors might need higher voltage.
I suggest to review the following parameters:
g_config_bat_config_emergency_voltage 3000 -> 2600
raw_bat_level_1 3960 -> 3250
raw_bat_level_2 3600 -> 2800
g_config_bat_config_bat_limit_limit_vol_min 3300 -> 2700
g_config_bat_config_bat_limit_limit_vol_max 3700 -> 3500
Solution 2: We can put 4 batteries in series. We can see in the datasheet that the voltage drop on this battery is almost immediate therefore even if we have 16.8volts when is fully charged, it will drop very soon to 14.8Volts. You can review the file hg2.pdf (first entry on google) to visualise the voltage values. 4.2*4 = 16.8v is normally really high if the drone expects 11.4v. At 3.7v each cell we would get 14.8 which is still high but it might work. Looking over DJI's products, their E300 ESC supports 4s for a 3s drone. Granted, this is not the ESC used in Mavic, but I was unable to find the ESC datasheet for Mavic. However, I strongly doubt that they would design an ESC that will burn at 16volts.
Charging the battery at a lower voltage is not ok, because we are losing the whole advantage of power density of this cell therefore if we don't want to charge the battery at 100%, is better to stick with other chemistries.
If we put 4s2p, we'll get 78.4Watt but an extra 166 grams : 82% more power, 22% more weight. This should get us 50% longer flight time but we are getting dangerously close into the motors inefficiency zone.
Therefore I expect to get from 21min -> 31 minutes or from 13 km -> 19.5km.
There is a 3rd solution but is more difficult to implement:
It involves adding a DC/DC converter which costs money, weights a little as well and has a 87-93% efficiency. We could put any reasonable number of batteries with this converter as we can provide whatever tension we want. For example 5 batteries would get us the same weight and the original Mavic but after the conversion loss we'll end up with 9.8*5*0.87 = 42.63 watts which is less than the original battery. The advantage here is that we could put 6, 7, 8, 9 batteries and still provide the same output voltage.
Note: I haven't tried any of the above. This is just theory (educated guess). I am not sure if the reduction of flight time calculated above is correct. If it is not correct, we'll get longer flight times but I strongly doubt that we will be able to achieve 52 minutes of flight time.
If anyone tries to put 4 batteries in series, please let us know, so we don't risk burning too many drones....
A comment regarding what I've seen people to do here: Use in parallel of batteries with different chemistries. I don't think this idea is wise: As we noted, the H-NMC battery will quickly lose tension. The other battery connected (Li-Ion) is not losing the tension at the same rate; LiIon has a tension curve which is quite flat during most of the power provided. What will happen is: The Li-Ion will start charging the H-NMC, but it will do so at an unregulated current. Both are designed for high current output and the transfer of the current will be at the rate that H-NMC can absorb power. This will result in batteries getting hot (high current transfer tends to do that because of the internal resistance of the battery). The drone might even get starved of power if the Li-Ion cannot charge the H-NMC and power the drone in the same time. You could very well crash your drone doing this. If you don't crash your drone, I am sure that you'll damage the batteries. Another problem with this is the heat generated is from the power that we wanted to use in order to fly the drone, power that we now waste. And finally, note that H-NMC can only work with this drone configuration until 3.3volts therefore you can only use 60% of the power which it stores.
Regards!
Last edited:
