This all-analog, microcontroller free (!) project got designed in response to cries for help from a diehard model plane enthusiast on the Elektor staff. He likes to fly large high powered models. One problem he ran into was self-destructing power connectors when connecting the battery pack to the plane (i.e. the motor controller). Every time the damage was due to heavy sparking, due in turn to high inrush currents. Those were expensive sparks as it turned out because the connectors are a 6-mm diameter, gold plated types. Clearly, an inrush current limiter is called for to ensure a controlled, spark-free initial current flow rather than a thump and a small explosion. Such a regulator did not drop from the skies, however, and took some time to develop at Elektor Labs. Below is a digest of how the project evolved from doodling to a working model keeping everyone happy.
Good guidance was found in Motorola Application Note number AN1542 . Using rough concept sketches (Figure 1) an inrush current limiter got designed for 37 volts battery power and a 200 amp load in normal operation. To achieve a low overall Rds(on) it is best to use a couple of MOSFETs in parallel. After an LTspice simulation run, the pain appeared to be not in the amps but in the load capacitance responsible for the inrush current, so the circuit got designed for the worst-case scenario. Still, there were concerns about the safe operating area of the MOSFETs. To test the water, measurements were carried out on a small 10-A 3-phase BLDC motor driver, and that turned out to have “just” 120 µF input capacitance. A bit later a bigger motor controller turned up specified for 120 A, and this was found to represent an input capacitance of 13,800 µF (13.8 mF) at an ESR of about 2.7 mΩ.
Moving towards a practical circuit the type IPB017N06N3 MOSFET from Infineon was chosen mainly based on the promise of 1.7 mΩ of ‘on’ resistance per device, not forgetting the relatively low cost and ready availability from the distributors.
Now the question remains: how many MOSFETs do we need? Back to the LTSpice simulation, now using the IPB017N06N3 model, some component values were in need of tweaking. Also, a heatsink was found—cheap, standard size (1/2-brick) and with predrilled holes. Looking at the schematic in Figure 2 there are some marked differences with the version proposed in AN1542. Motorola shapes the current into a square wave, causing a sudden current and power surge which slowly dies out. By contrast, the circuit shown here has the current increase slowly, resulting in a sawtooth-shaped current. Consequently, the power dissipation graph (PFET) looks like an inverted parabola. Figure 3 shows the basic waveforms—arguably they respect the safe operating area of the MOSFETs far better than AN1542.
A TVS (transient voltage suppression) diode, D3, helps to protect the MOSFETs in case of accidental polarity reversal. An early prototype was tested with a 15,000 µF (15 mF) capacitor with and without a resistive load, connecting to a 40-V supply through the X-Treme circuit. Everything seemed to function as expected, although, with no resistive load connected, the under volts lockout did not function correctly on the falling edge. As a final test, the circuit was used with a BLDC controller driving a 10 kW motor, unloaded, drawing 8.5 amps at the continuous speed and 20 to 30 amps when throttling. Tests were done at 37 V and 48 V, doing ‘cold starts’ several times over. Although cables and connectors got noticeably warm, the MOSFETs and the rest of the circuit remained cool. No “thump” sounds were heard (so customary from high-current loads), or exploding capacitors.
This flagged the go-ahead for the design and production of a single-sided (!) TH/SMD circuit board—the component layout is shown in Figure 4. The value of R1 sets the trip voltage hence is dependent on the battery voltage. The interdependencies are listed in Table 1. In practice, the circuit should not be used with battery voltages lower than 12 volts. Fortunately, that’s a rare occurrence in high-power (BLDC) motor applications—you can easily see why. The circuit board potentially carries extremely high currents, both ‘surge’ and ‘continuous’, meaning you have to strengthen all MOSFET source and drain PCB tracks, and the whole length of the BATT– and BATT+ PCB tracks, with pieces of 2.5 mm2 (13 AWG) solid copper wire, preferably two in parallel. Most of this plumbing work is in the area covered by the heatsink later. If you find 1.5 mm2 copper wire (16 AWG) easier to juggle with, that’s fine also but do three or even four pieces in parallel. Also, apply generous amounts of solder along the tracks and the copper wires—it’s a bit like Plumbing-4-Beginners. If for some reason your board comes with a solder mask on the above-mentioned tracks, remove the masking material and expose the copper by scratching with a sharp hobby knife. The pre-tin and install the helper wires.
The battery and load connections K1-K2 and K3-K4 must be made using high-quality terminals of your choice, preferably gold plated. Get the best you can find, round or flat (‘FastOn’ / spade type), whichever you prefer, as long as you solder them straight to the PCB tracks. Remember, every milliohm counts in this circuit and you do not want to lose motor power or torque during takeoff, now do you. To prevent polarity reversal, consider using a ‘socket’ (female) and a ‘plug’ (male) connector on the + and – battery lines. The same can be done on the + and – output lines. The MOSFETs are flat on the board, and the heatsink is on top of them with thermally conductive sheet material held pressed in between. The heatsink is secured with four corner M3 bolts or screws, with two M3 nuts on each bolt acting as standoffs, i.e. between the board surface and the flat side of the heatsink. The total standoff height is approximately 5 mm. The bolts should be lightly tightened so as to barely compress the heat conductive sheet material. Although we’ve talked mostly about motor controllers for R/C models here, the circuit is suitable for any 12-40 V DC load that represents a very low resistance initially, including big electrolytic reservoir capacitors and lamp filaments. AN1542:
 IPB017N06N3 datasheet: