Smoothing potential 30kHz noise in glider battery “switcher”?

Hi all,

I've created a circuit that switches individual items of avionics (for my glider) between two 12V batteries with a common ground. The idea is that I can switch items between battery A or B depending on how the batteries (and my power requirements) hold up over the course of a flight.

There are five individual instruments. In total they nominally draw from 100 - 500 mA, but some (the radio in particular) could spike up to 5A. wiring is specified for at least 5A, and each output has a 3A fuse (aside from the radio, which has 5A).

I originally started with the following circuit (the diagram shows 2 functional blocks of the five) -- 5 groups of P-channel FETs. Instruments are switched on, and between batteries, with individual on-off-on toggle switches at the gate of each FET. Capacitors at the gate provide enough debounce to prevent power cutting out when switches are moved from position A, through off, to position B.



This circuit works fine, but I neglected to notice the bypass diodes on each FET. This means that, if one battery depletes significantly more than the other (or if, for example, one battery is not connected), current from one battery circuit makes it into the other. I'd like to avoid this.

The obvious solution is two diodes at each output, but as battery capacity is at a premium in a glider, I'd like to avoid the voltage drop. I'm considering adding**broken link removed** to each output, as they have the lowest voltage drop I can find. However, the datasheet suggests that the voltage drop spikes periodically, with a frequency of about 30kHz / duty cycle 99%.

I'd like to avoid sending this spike off down wires to the instruments -- not that they can't handle it (they're designed to accept unregulated 12V from batteries), but to avoid the potential associated noise, which could impact sensitive radio / FLARM antennae.

Can I just stick capacitors between each output and ground to help smooth the voltage dips? Or will I need to add an inductor in series with each output, and a damping circuit as well?


Or am I barking up the wrong tree altogether? Is there a simpler way to do this? Or maybe I just shouldn't worry about the leakage between the two battery circuits?

Any and all advice much appreciated.

The voltage will drop slightly during the charge pump refresh, as the Mosfet turns off and the body diode starts to conduct.
That should not be a problem for your equipment.

Now, if you want a droop-free output, you may try the Linear LTC4358.
It employs the same charge pump principle like the TI unit, but it has an external connection to add a hold-up capacitor to prevent to avoid that.

Please note: I have not used this circuit myself, but from reading the datasheet it holds promise.

Or you could opt out for a little more complex, Oring-diode controller, which drive external Mosfets.
This last solution would be ideal, as it automatically performs the switching between battery banks, and have many protection features.

Thank you very much, that's very helpful.

The LTC4358 looks great (albeit very expensive). Looks like I can probably get away without the hold-up cap as well as the voltage only droops on switchover.

Ideally I could get away with these and do away with the P-Channel MOSFETS altogether, although I'd still like to be able to switch each gate on/off.

Some sort of automated switching sounds great, but it's probably easier to balance the instruments between the two batteries towards the end of a long flight manually. In reality, once I've found the best distribution, I expect I won't do much switching, unless one battery becomes almost depleted... at that point I'd switch essential instruments to the good battery and non-essential ones to the dying one (or turn them off altogether).

Besides the radio, and I would assume a GPS unit, what other instruments do you carry on your glider?

Basic instruments are mechanical/pneumatic (airspeed, altimeter, mechanical variometer)

Electrically-powered instruments are electric variometer, glide computer (inc GPS), FLARM anti-collision system, radio, and USB power (for GoPros etc). Transponders may eventually be mandated. For cloud flying a gyroscopic or solid state "turn and slip" or artificial horizon is helpful.

At the end of the day, it's probably only airspeed that is really critical (and even without it you can fly by noise/attitude if you know the glider well enough), but everything else is nice to have, especially when flying cross-country.