Concept Phase

A project log for RC Battery Backup with Magnetic On/OFF Switch

A 2-battery power path system for radio control gliders. Power can be switched on/off by a magnet.

Bud BennettBud Bennett 01/21/2018 at 04:540 Comments

(I moved almost all of the content from the details page to this log, where it really belongs.)

I admit that I don't have all of the information that I need at this point -- the initial concept stage. I decided to launch this project after attending a local RC club meeting where one of the members announced that there was a component offered by Zepsus that performed this function for *only* $43.00 that was small and light (11g). The concept appealed to me but the price did not. I had some expertise in this area so I decided to rough out a concept for an alternative design.

Circuit Approach:

My first thought was to use a magnetic reed switch, but quickly changed to a Hall-effect device, to sense proximity to a magnet. The circuit must have memory to latch the state of the applied power after the magnet is removed. It must also automatically select the battery input with the higher voltage, or share current between batteries with the same voltage. And if the backup battery is a voltage that is significantly lower than the primary battery, then the switchover event when the primary battery fails should not cause a significant glitch at the output.

So here's what I came up with in about 2 hours:

Note that all of the components use SOT23 or SC70 packages or variants. These are much easier to handle than MSOP, which have a 0.5mm pad pitch, and are still pretty small.

U1 is an ideal diode power path controller that automatically selects the battery with the higher voltage to provide power. U2 is a micro power 3.3V LDO regulator. U4 is the hall-effect switch, and U5 is a D-type flip-flop that provides memory. M1-M2 are low RDSon FET switches that work with the LTC4412 to select the input source. M3 and U3 switch the power on/off to the load.

The only non-obvious circuitry is R4-C4 between the flip-flop Q and D pins. I expect that the hall-effect switch output to glitch when the magnet is drawn close and also removed. R4-C4 ensures that the flip-flop only changes state once during a magnet sequence of around 1-5 seconds.

I expect significant changes to much of this as I realize other requirements or better approaches...

Off-state Current Budget (typical):

U1 -- 20µA

U2 -- 2µA

U4 -- 8µA

Total -- 30µA

Cost Estimate: