I’m thinking about scratch building an instrumented battery charger. Why? Hobby. I can build exactly what I want.
Some of my projects never get past the back of an envelope. Most of the time, I complete the design but never order parts. Others have made it to working prototypes. I’ve never built more than two or three of anything.
We will see how far this gets.
Constructive comments very welcome!
Step 1 Requirements and design goals:
Must be something I can build at home.
Must be possible to model in Spice or LTSpice
Able to charge 1 to 4 cells.
Up to 4 independent charger circuits.
Multi Chemistry Li-Ion, LeFePo4, NiMH
Programmable or adjustable charge rate.
Maximum charge rate of at least 2 amps. 4A preferred.
Battery temperature monitoring. Charge termination on over / under temp.
V, Current, Ah logging
Internal resistance measurement
Computer interface. (Serial / BT / Ethernet to be determined later)
Modular design.
Will operate stand-alone without external computer or controller
I’m going to use a commercial battery charger IC. No point in making this harder than I need to.
To make selection easier, I am limiting myself to parts from Linear Technology. This cuts down on the number of data sheets and application notes I need to read. More importantly, many of their parts have LTSpice models.
LT 1511 and LT 1505 do not support LiFePO4. I will probably eliminate them early on.
LTC 4000 reference designs have a high parts count. I will probably eliminated it.
That’s all for now. I’m off to read some data sheets.
I’ve updated the requirements to explicitly mention independent charger circuits. It isn’t safe to run all four slots in parallel from a single charger. While this is obvious, I thought it was worth a brief explanation.
We don’t know the chemistries or the state of charge(SOC) of the 4 cells in the slots. For example LiIon and LiFePo4 have different CV voltages. The charger would end up over or under charging some of the batteries.
Even if all 4 batteries are identical types, we don’t know the SOC.
Imagine what would happen if a fully charged INR18650-30Q was in one slot and a discharged INR18650-30Q was in the second.
The two batteries would be in parallel. The batteries would be connected together with only about 40m ohm load. I = 2V/0.04 ohm = 50A.
50 amps is a bit more than the recommended 4A rapid charge current. I’m not sure what would happen but it wont’t be good.
Of course we could put a protection fuse or resistor between the charger slots, but independent chargers is much better.
Texas Instruments has too many choices. Lots of reading ahead.
bq24190 Lithium only.
bq25601
bq25890 and bq25892 Lithium only.
bq25898
The features look good but it comes in a microscopic DSBGA package. It has 42 pins in a 2.8 x 2.5 mm package. PCB layout and assembly will be challenging. I will probably skip this part because of the package.
bq24725a
bq24735
For later — I2C keypad scan IC. May be handy if I put a keypad in the device.
PCA9557 I2C and SMBUS I/O expander - 8 bit parallel port
ADP5350 maxium fast charge current is only 650ma. Otherwise it was looking good. It has a lot of extra functionality that could have been usefull. I’m going to keep that one in mind for future projects.
I will probably go with the LT4015 but am going to look at the AD5062 first.
Standalone or I2C control. In standalone, basic settings by jumpers
Thermal limiting and battery temperature limits.
Very low external part count.
Battery Isolation FET. Useful
USB or external adapter power.
Battery short detection.
Dead and weak battery trickle charge rates.
Optional JEITA battery temperature charging rates —> reduce rate if battery is too cold or too warm. Adjusts CV voltage too.
Programmable Charging termination voltage between 3.6V and 4.5V in 0.02V steps.
Programmable fast charge current between 50mA and 1300mA in 50mA steps
dead battery, trickle, watchdog and safety timers
programmable LDO for system power - I can use the 5062 to provide power the rest of the device.
Good application support including PCB layout guidelines.
Layout guidelines are sooo helpful.
The part is missing a few features. 1.3A charge current is below my design spec of at least 2A. 1300mA just barely covers the standard charge rate for some 18650 cells. I’d like to be able to test and use rapid charge rates too. An INR8650-30Q or 25R can rapid charge at 4A.
Panasonic NCR18650B standard charge rate is 1.625A
1.3A is not enough. I like the ADP5062 but it won’t work.
I would approach this from a different direction, battery charger ICs are designed for minimum cost low parts count commercial projects and can be fussy to operate outside designed goals. Starting with the goal of flexibility and maximum control I would use something like an arduino with software to control and measure current and voltage with basic power chips instead of battery specific.
I’ve thought about that. I agree that most of the battery charger ICs are designed to be part of a device like a tablet or laptop.
In the end, I may end up going with the flexible option.
Before I do that, I want to explore the more integrated parts to see if there is one I can make work.
I’ve already eliminated quite a few of the low part count ICs so because they don’t meet requirements.
Going the Arduino + power supply route puts the parts count up a lot and possibly the load on the Arduino. It means my code and external hardware is responsible for safety checks.
Some of the charger ICs have programmable temperature, voltage and time checks. They go into a fault mode if when any of the limits are exceeded.
Since I plan to have 4 independent chargers, part count is important.
==
I’m impressed by just how few external parts the new charger ICs need.
This is an SLA charger I built in 2012. It charges a single 6V battery. Look at the parts count!!
It is suppose to be a backlight driver.