- In the beginning the main target was to have a flexible board with just soldering pad’s to connect the cables, but it makes the board a bit bigger in size.
Furthermore, some people just don’t have the skills to solder a wire (REM: why somebody in god’s name own an MR if he can’t solder a wire?). So it was
decided to install DF-13 connectors with two different cable length, to adjust the distance to the FC (10cm or 20cm). Any other length possible on request and
without increasing costs, but might be with a delayed shipping date.
Why use a Hall sensor ?
- The measurement over a normal shunt resistor is not accurate at lower current (<3.0A). For a Hall sensor the measurement starts at 0.5A with an accuracy of
+/-0.5A over the whole range up to 200A !
- A shunt resistor create heat due to the voltage drop, the hall sensor has only an internal resistance of 100uOhm, so there is no power loss.
- Due to the heat created by a shunt resistor and the power cable, the measurement of the current is not linear and depends on the temperature. This is not
happened to a hall sensor, a temperature change (created by the main LiPo cable) will not influence the measurement.
- The current flows only through the hall sensor and NOT through the PCB. Most other current measurement boards has the main cable soldered to the PCB and
then it goes to the shunt resistor -> these boards can’t handle over 60A constant current ?
- Hall sensors are very expensive, compared to a normal shunt resistor and not everybody out there wants to spend the money to top up for a good measurement
system. So the sales quantity and profit will not be within the target.
- Pixhawk has internally a 3-way power selector over an ideal diode chip. The 3-ways are USB, power connector (6pin) and the Output PWM rail on the back of the
FC. So it is possible to power up the FC with either one of this power sources, but how do we know which power source right know is powering up our FC if there
is USB, a PM module as well as an backup BEC connected to the output (ESC/Servo) rail ?
The answer is easy: Whichever voltage is higher by 0.25V to any other power source is selected as the internal power supply, as long as this voltage do not
exceed 5.70V !
The result in practice on the field can be different, as there are many components connected to the FC like, GPS, Servos, opto ESC’s… etc., the power
consumed by the system is not stable, which means the supplied voltage is not stable as well. The reason for this is the loss in voltage due to small power
supply cables and maybe many connectors.
To prevent the internal ideal diode to switch too often between different power sources, we choose a bit unusual high voltage (5.35V) as a main power supply.
Which means only if any other power supply (USB or PWM rail) is in the small range of 5.35V+0.25V=5.60V and the maximum voltage of 5.70V, then the diode
would switch over to the other source.
- To reduce the resistance in the power line and increase the safety, or should we ask, why does the DF-13 power input of the Pixhawk / APM has +/+/I/U/-/- ?
There are also two wires, for positive and negative, used to reduce the risk of failure.
- A switching power supply can be a very “noisy” part in the power supply chain and it is very difficult to shield the coils (1.5MHz) from the current measurement
board. So it was decided to keep the two away from each other.
- Many people complain that the UBEC seems to be a bit big, but fact is that he is only 22mm x 17mm. What makes him BIG are the safety capacitors at the
input and output !
We all had the issues before that any ESC burned out due to the “hammer effect” in the supply lines, but do we consider that the UBEC is sitting on the same
voltage source ?
Does anybody ask himself so far why suddenly his BEC burned out ?
Why does some people add some capacitors onto the ESC’s to reduce the risk of failure, but in the same time they forget that there is also anywhere an BEC in
the supply line which might need some protection too ?
How good is it if your ESC’s survive a voltage spike, but your BEC didn’t and the MR crashes ?
If you can answer some of the questions by yourself, then you will also figure out why this UBEC is a bit bigger than others.
- That's one good question and a bit difficult to answer, so lets look at the datasheet of the hall sensor itself:
100A Sensor -1.3% / +2.4% Output error over full scale Noise output signal 12mV
200A Sensor -1.2% / +1.2% Output error over full scale Noise output signal 6mV
- We can see here that actually both sensors are the same, except that the 100A version has an higher internal signal amplification at the output stage, which
results in double output error as well as noise.
- So what does that mean if me measure with both sensors exactly 50.0A ?
100A Sensor current reading 49.35-51.2A - Output voltage ca. 2.0V (with Vcc 5.0V) - Noise 12mV
200A Sensor current reading 49.40-50.6A - Output voltage ca. 1.0V (with Vcc 5.0V) - Noise 6mV
According this short calculation, we could always use the 200A sensor as he his even more accurate, but we have to look at the output voltage too.
As lower the voltage in an analog signal wire as more sensitive is the signal to external interference from RF and EMI and the wire to our FC is not shielded. So this will cancel out the advantage of the 200A sensor at low current measurements.
So the result is now, that we should use the right sensor for the right MR, which means if your hovering current is 20-50A, then choose the 100A sensor. If the current is 50-xxA, better go for the 200A sensor. Furthermore, keep the DF13-6P cable between the current sensor and the FC as short as possible and try not to run it in parallel with the main battery cable. This is actually the main reason why I send out the sensor boards with 2 cables, one in 10cm and one 20cm.