LiFePo4 steady charging/balancing?

[oxcart]

I'm trying to make a compact emergency 12V starter pack for my prius C. A died 12V(4-6yrs on average) is the most common cause of stuck hybrids.
On paper it's doable as the 12V is only an auxiliary battery and the peak absorption(power up main ECUs, then close 2 HV battery switches) is something like 11A for 1s.

I've done a bit of homework: i selected a 4S LiFePo4(for safety) battery pack and i want to put a diode on the + of the pack, so after power up the DC/DC converter which powers the 12V bus isn't charging it(to my knowledge the car DC/DC has capacitors and everything and runs just fine without the 12V battery, there are no alternators).

Now the catch. I don't want to recharge the pack at home or to periodically check it or to keep it at 100% charge thus reducing its service life. Well, i'll eventually check it a lot since it's a project, but the point is i want it to be "stable" and plugged in, always ready.

A 3.33V/cell should be at ~80% SoC and with plenty of amps to power the ECUs, while aging well. I plan to use RC batteries which have the "power" XT60 connector and a 5-pin tap to charge and balance the 4 cells.

However
1)I can't find a "charge my Lifepo4 to 80%" charger, is it fine to a) use a schottky diode with a 3.7V/cell charger to have a 0.4-0.37V drop or b) to just hook the 4 cells at a regulated 3.33V/cell source as this https://www.pololu.com/product/2842?

More difficult. How do i keep the battery balanced, aka 3.33V for every cell? I can't find an off the shelf 4 cell charging/balancing circuit, only passive resistors, and it must not get hottie or self discharge when the car is parked.
Is there any clever way to get away with only 1 12V DC/DC or am i forced to use a single 3.3V DC/DC for every cell? Can they be wired in series?
Please enlighten me

[David Hess]

1)I can't find a "charge my Lifepo4 to 80%" charger, is it fine to a) use a schottky diode with a 3.7V/cell charger to have a 0.4-0.37V drop or b) to just hook the 4 cells at a regulated 3.33V/cell source as this https://www.pololu.com/product/2842?

While not ideal, using a series diode is probably acceptable given the rough environment (operating temperature range) for the batteries which will shorten their life anyway. 

Many regulators use a pair of resistors to set their output voltage and it should be possible to adjust one for the correct output voltage.  Pololu nicely proves a schematic where you can see these resistors as R1 and R2; shunting R1 will lower the output voltage.  You can also see on the schematic that the output is not isolated; see below.

Do lithium rechargeable batteries have a temperature coefficient of voltage for charging ?  I have never seen much about it.  Lead-acid cells have a temperature coefficient of about -3.3mV/C.

Is there any clever way to get away with only 1 12V DC/DC or am i forced to use a single 3.3V DC/DC for every cell? Can they be wired in series?

If the regulators have floating outputs, then they could be connected in series but that is a rare feature; usually the input and output share ground.  It is also easy to design and build a regulated inverter with as many outputs as needed but the awareness of this type of circuit has been lost with time; it is or was common when floating or isolated outputs are needed.  Pulse (gate drive) transformers are suitable for this if a standard transformer is desired instead of a custom one.

More difficult. How do i keep the battery balanced, aka 3.33V for every cell? I can't find an off the shelf 4 cell charging/balancing circuit, only passive resistors, and it must not get hottie or self discharge when the car is parked.

I have pondered the same problem in connection with stacked supercapacitors.  I have a couple of old Powerware online UPSes which run on 5 (!) 12V batteries and given the cost of replacement batteries, have considered replacing them with 34 to 36 series connected supercapacitors; that will not provide much run time but will be more than adaquate to ride out short power line disturbances which are getting more common year by year.

I think the easiest solution in this case is to switch the series resistors into the circuit *only* while power is applied to the charging circuit.  A multi-pole relay could be used for this or MOSFETs perhaps driven by diode-capacitor charge pumps to generate their needed Vgs while providing effectively floating outputs.  Then the resistors can be sized for good charge balancing without worrying about discharge current.

[onlooker]

I did not really understand the details of the problem that the OP is facing. But, maybe, the OP's problem can be solved easily by using a 4 cell BMS and a DC-DC step down convertor.

--A 4 cell BMS board for 10A to 20A peak ($5 to $12 on Ebay or Aliexpress).
Be watching out for
  1). the current rating (peak and continues);
  2). with or without balancing;
  3). the working/standby current of the protection circuit (e.g. 1uA ~ 500uA). 

If the cells are similar enough, the balancing feature is likely not required. The BMS will shutoff the battery pack when the weakest cell is charged or discharged to the protection setting (e.g. 4.2V and 3V). A BMS without the balancing capability can still protect the pack on a per cell bases.

--A DC-DC step down module for setting a proper charging voltage ($1 to $2). If the charging voltage to the BMS is too high, the BMS will get into the protection mode and shutoff. But, I think you can always set it lower to just charge the pack to roughly  80%.