Ever wonder if leaving your ESC on will drain your battery in a short amount of time? What about turning the ESC off? Is there still a risk? Does it depend on the ESC?
Well, I measured a bunch of ESCs to try to add some real numbers to these questions.
Types of switches
ESCs aren’t easy to turn on and off. The capacitors along with high currents during use means a regular switch won’t last. There are three major approaches that designers take to turn an ESC on and off:
Mosfet based “anti-spark” switch
This is built into the ESC, and consists of a bunch of mosfets along with a mosfet driver to turn them on and off based on the input of a momentary or latching external button/switch. Since it turns off the power right at the source, it should pull little to no current when powered off.
Buck based switches
There are different approaches to this, but the idea is that most buck converter ICs have an “enable” pin. Pulling this low means you can turn off the buck and thus everything behind it. The battery voltage needs to be converted to lower voltages to run 3 main systems on any ESC. ~10-12v to turn on the mosfet gates, 5v for accessories and some chips, and 3.3v for the STM32 and other chips. This is usually converted one after another (battery → 12v → 5v → 3.3v). It’s not easy to get clean power from 48v-3.3v, but much easier to go from 5v to 3.3v. So if a design only turns off the 3.3v buck, then the 12v and 5v bucks will still be active and powering whatever is behind them.
Even with the buck turned off, the capacitors are still connected to the battery terminal which means there will still be a spark on the first plug in. There is also some current drawn from the battery due to the capacitors self-discharging.
None - external loop key
Since the XT90s can handle the current, building in a loop key, especially one with a built in pre-charge resistor, will ensure that the ESC is fully turned off whenever unplugged. No extra electronics in the ESC are required and this method is fairly reliable. Depending on the location of the loop key, other devices like buck converters for LED lights or BMSs might still be powered on.
Testing setup
A loop key adapter was placed in between the power supply and the ESC. Then a multimeter in current mode was attached to the two open pins where a loopkey normally goes. Current was measured both off and on at a fixed 48v
Power draw
The mosfet based ESCs did very well when turned off. 6.2 MICRO amps is virtually nothing and represents about 0.0003 watts at 48v.
The buck based ESCs had some mixed results.
The SKP Solo was higher but still did pretty well.
The MakerX D60 is somewhere in the middle with a pretty reasonable draw of 0.65 mA when powered off.
The Uboxes were pretty bad. 1.5-3 mA when powered off. Also interesting about these ESCs, is the power draw gets worse the higher the battery voltage. At just over 16s equivalent, the Ubox V1 was drawing nearly 4.5 mA when fully powered off
Also I don’t know what’s going on with the iLogger, but that power draw is crazy for a small ESP32 based device.
Real world - what this means
First, if the ESC is left on, the power draw is significant enough that you shouldn’t leave it for more than a few days without checking it. A fully charged 12s4p P42A battery has 715 Wh at a 200mA discharge rate. Even best case scenario, that’s only about a month.
Note that this does not include accessories plugged in, BMSs, the self discharge of the cells themselves, or temperature.
How about when the ESC is off?
On the ESCs with mosfet based anti-sparks, you probably don’t have to worry about that being the cause of a battery failure. 20-200 years even at a low battery charge.
On ESCs with a buck switch, the Solo still does pretty well, needing 3 years if the battery is at 25% charge.
The Go-FOC D60 is a little worse. 8 months at 25% charge
The Uboxes though… 2-3 months at 25% charge.
