Custom Efoil control board with Anti-Spark, cooling control & Kill switch

So a good friend and college of mine Adrian, has a wealth of electrical knowledge. Naturally, Ive been bending his ear for the last few weeks asking for some help on the electrical setup.

Anti Spark and kill switch being important. Adrian pointed out that relays for 100+ amps are big, and in extreme cases can fail. He suggested solid state relays (mofsets) and offered to draw up a circuit board design to incorporate 12v (for pump) 5v for receiver, PIC chip for pump control, Anti Spark and kill switch.

Adding a passive AMP meter onto the board and using VESC protocol to send Battery AMP and voltage to compatible receivers (Maytech V2) for non VESC like the seal 300A

Below is the logical layout.

Q1 is 6 x IRF4905
Q3 will be an IRL2203N
116Amp transistor to run a 0.3A pump
Q4/5 will be 2N3904
PIC sets the anti Spark timing for the smaller mofset Q2

Here is the PCB design

The idea is if the kill switch (manual boat/jet ski style) is off, everything is safely quickly disconnected.
Using just one 12V UBEC, the board offers connection to receiver and a 12v pump.

Ild like your feedback and let me know if i have missed anything vital, any glaring mistakes or feedback appreciated before i make a 400 euro smoke bomb.

1 Like

The chosen PMos is not capable enough. 6x IRF4905 still makes 3.3mOhm which is far too much. <0.5mOhms should be the goal. Also i would use NMos and cut the GND path for the ESC.
A good candidate for the NMos would be an IPT007N06N or IPT012N08N, use 2,3 or 4 in parallel.
The problem with shutting down at high current is the voltage peak created by inductancy of the battery cables. Either you switch it off slowly or you limit the voltage by a surge diode, a very powerful zener diode. Also you do not need an extra transistor with resistor for precharge. Just switch the main switch on slowly, e.g. by a DAC or a PWM out.

Thanks PowerGlider. He takes your point with those 4905’s it’s just what he had to hand. He had suggested they might not be man enough and said there are better ones available.

Will look at cutting the ground using n rather than p. ( that’s where it started but he had a reason not to do this)

I’ll update as the plans progress. But Adrian’s now working on telemetry For those of us NOT using VESCs. The open source nature means that effectively we can get some of those features on any ESC.