A complicated question about LiFePO4 batteries, - and switching.

On Wednesday, June 20, 2018 at 9:27:33 AM UTC-4, OHM Ω http://aviation.derosaweb.net wrote:
Andy,

I have used the ideal diodes mentioned by Shaun McLaughlin (http://re-voltage.eu/ US$27 each) as I was given a pair to test with. They do what they say with minuscule voltage drops across them. Full disclosure, I am currently using two 1N5821 Schottky diodes in my panel right now (page 74 of my presentation) - small, simple, cheap, plentiful, robust, and easy to mount..

That being said - the availability of inexpensive ideal diodes has exploded in recent years mainly for use in systems like solar panel arrays. See https://www.ebay.com/sch/i.html?_fro...es%29&_sacat=0. I also like the increasing range of mounting styles available allowing for placement options behind our increasingly cramped panels.

One other comment: If you use diodes* (of any type) you can run two batteries in parallel safely. The danger was always been that without diodes, and with two batteries at different voltage output levels, that one battery could cross-charge the other. This could potentially happen without any current limit (!) until they equalize voltages. On a single bus with two batteries and diodes, the battery with the greatest voltage level would power the bus by itself until the separate battery's voltages equalized.

- John OHM Ω

* Teaching Moment - A diode only allows current to flow in one direction. In our case current will be limited to flowing from the battery to our avionics. As current flows through a diode there is a small voltage drop across the input to the output. We want to avoid, or limit, this voltage drop in our glider based power systems as we have no generating power source as found in airplanes.

I wonder about those "ideal diodes". They are presumably similar to what's in the batteries' BMS, I would guess power FETs that are switched on or off - and when "on" they conduct well with minimal (but not quite zero?) voltage drop. The "ideal diodes" must also have a sensing circuit that responds to the small voltage drop, amplify it and feed it to the gates of the FETs. (Seems like a Schottkey diode would be a good enough substitute.) The ones in the BMS can conduct in both directions (for charge and discharge), perhaps there are separate FETs for each direction.