Avoiding sparking when connecting companion board

When the smart SOC board is attached to the battery, it can produce a decent spark:

• High (30-60V) battery voltage
• Uncharged input capacitor on DC/DC converter
• Inductance of the battery wires

Without a limiting resistor, the current spikes to nearly 50A with a 42V low impedance supply! It also causes high voltage oscillations. A 15 Ohm series resistor solves the problem:

With a series inductance, the system can oscillate. Tricky.

Good article from RECOM on limiting the inrush current.

RECOM RPMH's soft start

Power consumption

The effect on the energy consumption of the board is insignificant. With the 0.5 mA that the Yamaha board is pulling, the voltage drop over 15 Ohm is only 7.5 mV, and thereby the power 3.75 uW.

Experimental validation I

A 10 uF electrolytic capacitor makes a nice spark when connected to the desktop power supply set to 42V.  Same when shorted out. However, with a 22 Ohm inline resistor, it is impossible to get a spark.

Experimental validation II

The aid was to try to kill one of the SoC boards forcing them to spark. On a first attempt, we connected the amperimeter in series with the DC-source and tried to measure the current when the spark appeared. In reality, the amperimeter owns a low resistance value that was protecting the circuit all the time, making it impossible to kill.

When measuring the amperimeter resistance, we got a value of ~1.2 Ohm, and since this was enough for protecting the circuit at all, future developments will include a resistor a bit greater than this (10 Ohm for instance).

Experimental validation III

Finally, we have verified that using a 10 Ohm series resistor at the B+ terminal of the board, makes the system much more reliable, leaving no way to kill it from sparking.