search

SAR-mango / SinESC

SinESC brings sinusoidal motor control to the RC drone and fixed-wing communities without common downsides.
GNU General Public License v3.0
18 stars 5 forks source link

Mosfet proposal #6

Open LukasHNZ opened 3 months ago

LukasHNZ commented 3 months ago

What about the https://www.infineon.com/cms/en/product/power/mosfet/n-channel/isg0613n04nm6hsc/ ? it got less resistance and lower gate capacitance. There are also other ones with a bit higher capacitance and lower resistance or vice versa. What do you think is the best tradeoff? Nexperia also got some very low resistance mosfets.

As I understood, a lower Resistance and faster switching would result in less Thermal losses, thus saving weight for cooling and increasing flight time both ways. Looking forward to hearing your perspective!

SAR-mango commented 3 months ago

Glad to see interest in the project even after several years! May I ask what brought you here?

The FET you suggest is a great option. It didn't exist when I did the parts selection (four years ago), but I found that considering the product of the gate charge and on-resistance helped me narrow down the options greatly. This FET excels in that regard.

When selecting parts for a new version of the ESC, I am going to consider the avalanche energy rating – this is crucial for durability. I may even switch to 60-volt FETs since 6S seems quite common these days (though I have sadly not kept up with the hobby since college began). I might also prioritize gate charge over on resistance, since I feel like I overestimated the peak and average motor currents. 55A seems overkill.

When developing the ESC into its current state, I mostly focused on the hardware. But when I revive the project, I am going to spend lots of time creating my own control algorithms. Partly because I now have some theoretical background, and partly because this is arguably more important than the hardware alone. I would be able to tailor the code for this exact application, cut out the useless garbage, and potentially stop using ST's HAL library. I also want to develop a self-tuning regime since I hate the idea of having to tune PIDs for each ESC/motor combination.

LukasHNZ commented 2 months ago

What brought me here is google! I recently saw a ton of videos regarding a fpv speed record, reaching 500kph for a few seconds, also the Redbull drone that could stay at high speeds for relatively long times. Then the recent hype around special rotor shapes like from lifeline or toroidal propellers. I was wondering what the technological limits are at the time and how it can be harvested! Some motor controllers on the market already use Vector control (FOC) - is this a synonym for the sinusoidal one you got or are there still differences? It also claims better efficiency, low end torque, less noise etc. but at a hefty pricetag usually.

If there is a significant difference between your approach and a FOC approach ill probably end up testing both, i was considering using a generic motor with any prop that usually goes with it, and then testing stationary thrust with a small rig with a load cell to find out how the real life figures are. With the current batteries energy density and the power hungry nature of UAVs every % of energy savings is very valuable- especially when going for record attempts or other extremes

SAR-mango commented 2 months ago

There are differences between FOC and sinusoidal control. FOC is theoretically superior. It involves determining the exact rotor position and energizing the motor phases such that the magnetic field is exactly orthogonal to the rotor, which maximizes torque (and thus efficiency) and smoothness. I say theoretically because measuring the rotor position is difficult and usually requires a sensor. Sensorless techniques also exist, but require lots of tuning, precise current measurement circuitry, and well-designed software.

SinESC performs sensorless FOC. In the readme I used "sinusoidal control" and "FOC" interchangeably, but they are not the same. For certain motor types, FOC results in a sinusoidal waveform. In these cases, one can do away with FOC and recreate the sinusoidal waveform "blindly". That is what sinusoidal control is. Nevertheless, I hope to develop the ESC to the point where it becomes ubiquitous within the hobby.

My expertise is EE and I hope to do my part to create the most efficient ESC, but there are probably better places to save energy. You should first optimize your overall avionics system and ensure that none of the subsystems draw a ton of extra power (example: choose an efficient video transmitter set to the minimum RF output power necessary). Then look into the mechanical and aerodynamics stuff – frame, prop airfoil, prop/motor combo, motors themselves. The greatest savings will probably be found there.