PSAS/ news/ 2004-06-09 - APS (Avionics Power System) design: Li Ion battery charger choice

Folks: Ivan, Ted, Tim, Andrew, Kiah

Agenda: Choose a battery charger IC. Figure out more about battery charging.


We were looking at a bunch of battery charger ICs: see ChargerIcResearchLv2 for a list of the chips we considered. We decided to go with the LTC4006 as the chip of choice until we find out that we can't use it, in which case we switch to the MAX745. We chose the LTC4006 because:

  1. It's autonomous - don't have to have firmware in the PIC to charge the battery (vaguely) safely (note that the PIC should be doing some kind of state estimation so that cycling the input power to the charger doesn't continuously charge the battery).
  2. Small parts count (like the MAX 745) and in an SSOP16 package.
  3. Current limits the current from the input power source by folding back the current to the battery.
  4. Uses a MOSFET instead of a diode for the input power switch.

Note that using the LTC4006 does require a diode to seperate the pack from the input power to the avionics... but we can use a "switched diode" like the LTC4412 mentioned on the ChargerIcResearchLv2 page. Note to Ted: we'll probably want to draw the avionics power system power from the left side of the battery current shunt so that we can also use that shunt for the PIC's monitoring of the battery current (both charge and discharge) .

Tim and Andrew came to the stunning conclusion (love that word, "stunning") that we'll be wasting ~ 0.5 watt (!) per node if we stick with linear power supplies. This just is too much, and there may be a problem with some of the older current nodes which are using a 78L05 type of linear regulator. So this almost forces us into using a switching supply, so it looks like we'll just use Tim's recovery node front end for all future PIC nodes.

To Do:

Ivan Research the power input to power output side of the LTC4006
Ted Research the battery side of the LTC4006
Andrew Redesign the older parts of the APS: the umbilical interface, the RR signal, power switches, etc.
Tim Dig up the via strip design of the recovery node and redesign the crowbar circuit, 10MHz xtal divider for new board, and anything else that needs to be done.