I'm thinking over new multicopter power distribution board which will integrate also 4 current sensing shunts, two dual current/voltage/power sensors, a 42 Vin 3A DC-DC Buck regulator. The current,voltage and power measurements will be carried out over I2C interface. I think it's time to place aside the analogue measurements as there are a lot of clever ICs for the purpose.
Some explanations:
1. R1-R4 are 4 current shunts. Each of them can handle 60A...or even more, but some more thermal researches should be done;
2. US1 and US2 are dual current/voltage/power sensors with I2C interface;
3. U6 and surrounding passive components are the DC-DC Buck regulator;
4. The PCB is 6 Oz double sided with glued heat sink ( optional...if needed ) on the bottom layer.
As this is just a preliminary idea, I'd like to know your pro and cons about the design.
Best regards
Nick
Comments
http://www.rcgroups.com/forums/showthread.php?t=2182068
If' you're pulling that much current, chances are you're going with at least a four cell battery, possibly higher. The 12V BEC would be useful.
Hi naish88,
If you imagine the I2C interface, you'll note that the drain resistors are connected to Vcc at autopilot side ( in most cases 3V3 ). As the only wires going from the power distribution board to the autopilot are GND, +5V, I2C_DATA and I2C_CLK, the only critical wire remains the GND. CAN driver will use the same GND and +5V from the DC-DC Buck + you'll need some cheap small MCU over the Power distribution board to serve the CAN interface. This worth just in case you use this MCU for all functions the power distribution board could be engaged with. The CAN could be used for almost everything - to drive the ESCs and to carry current/voltage/power data. Thus you'll have again just 4 wires between the power distribution board and the AP board - GND, +5V, CAN_H and CAN_L. Thus all ESC signal inputs will be connected to the power distribution board but not to the AP...... ( this idea is interesting to think over ).
The only thing I'm worried about is that the power distribution board is too noisy place and I have to choose either noise resistant MCU or just shield it.
I like the idea of 4 current shunts, as also someone suggested will be really good to have some more for fpv batt, gimbal.... The only thing I'm not really prone to, is the I2C bus. Analog is terrible, but If you use I2c and you share the same bus with the compass and other sensors, there is a great chance that noise can create issues. I will rater prefer to see something like the CAN bus, where all the new sensors are going to.
There's no problem to obtain 12VDC from 4S or higher, but is not easy to have 12VDC from 3S or less. It requires a DC-DC Buck-Boost regulator, but these regulators work up to 12V Vinmax or less. Just a few of them can work at much higher voltages, but the price is high enough to render their use.
If most of multicopters are using >4S, the solution is simple.
I would highly suggest an integrated 12v BEC that can handle around 2 amps. Most video gear is 12v and it would eliminate a BEC floating around. A 5v 2amp BEC would be great too. Backup power to the avionics and also for lights.
Very nice job :)
I would like to point out that, having multiple current readings (one per motor) is a great way to acquire real-time data and use it for failure prevention. Using graphs to see the data over long periods of time, one could probably tell if a motor needs cleaning or replacing by just looking at the current draw during hover @ 50% throttle for example.
Also, during flight, it would be possible to analyze the current readings of each motor and possibly detect a loose prop (or missing prop) and apply a sophisticated recovery algorithm perhaps :)?
This would of course require software and firmware modifications, but it would be worth it IMO.
i'd say 500mA to 2A if you need to power all the fpv stuff and gopro - it cames to about 1,5A
Hi Ben,
Yes, the idea of using I2C is to get ready readings for current, voltage and power from the current sensor. For such short wiring, it is not a problem to have noiseless I2C info. With the same cable length, I2C will be more noise immune than the analogue one.
About the additional 12V...have to think about...maybe I'll find some place for. What current it should be designed for....250...500mA?...or less?