Bodmer wrote:lm317 wrote:I actually don´t see a prob with the voltage divider.
Agreed, on it's own it is a simple circuit that is easy to make and calculate the values with no design flaws when considerend in isolation. The complications arise when it is connected directly to the battery in the more complex setups, for example where an independant BEC is used or a BEC within the ESC.
The diagram below illustrates a bad setup. This diagram has been simplified to illustrate a point, the heavy black lines are the 0V wires. The 0V line to the motor is electronically commutated normally but for simplicity it is shown as a direct connection. The line shown as carrying 1.5A is the 0V line of the voltage divider sensor:
Parasitic paths.jpg
There are multiple reasons why this is particularly bad:
1. A parasitic current path exists from the ESC through the receiver back to the battery. In a setup using 14AWG wire for the main battery and 26AWG wire for the receiver and voltage monitor circuit the parasitic current could be 2 Amps or more if the motor current is 50A. If 20AWG wire is used in the receiver leads the parasitic current approaches 5A!
2. There is a parasitic path via the sensor return that bypasses any common mode choke in the power feed from the BEC/ESC to the receiver (typical in switched mode BECs). The common mode choke now makes supply noise much much worse since it couples noise spikes on the battery lines straight into the power feed to the receiver, this could easily reduce the operating range of the receiver and hence cause loss of control at a distance......
3. The parasitic current path can upset switched mode BECs that use a significant level of common mode output filtering, affecting their output regulation and causeing "pulse skipping" that typically results in very brief brownouts in the receiver supply, again likely to cause loss of control when the servos start working hard.
4. If the battery to ESC has a connector and if the +ve line to the battery connects first (even very briefly during a 2 pole connector mate) the receiver is likely to experience a brief voltage surge of the full battery voltage and very high transient currents will be routed via the receiver while the bulk decoupling capacitors at the input of the ESC charge up (brief capacitor charge curents in excess of 100Amps are not unusual and this charge current will try to go via the voltage divider 0V line!!!!!!)
So don't be surprised if things mysteriously fail with this setup.......
Here is a setup that deletes the parasitic path:
Divider monitoring.jpg
In this circuit the OV line from the resistor divider is NOT connected to the battery. The parasitic current path through the receiver is now removed and the battery voltage is sensed relative to the 0V that is routed via the BEC. The capacitor fillters the ground noise resulting from transient voltage drops in the 0V line. The 4K7 resistor limits the discharge current from the capacitor into the analogue channel input to the microcontroller in the case where the capacitor is charged and the receiver is off. (Note: even a very brief current of more than 2mA into the receiver pin will destroy or reduce the life on the receiver.)
The problem with the high power electrical systems that are used in model aircraft is that with the high currents and high transient voltages (that are invisible in a simple electrical analysis of the system) make wires look like resistors and inductors. Also low parasitic capacitances start to couple significant currents into sensitive electronic components.
ATB,
Bodmer