Batteries for Breakfast: Getting my battery charger NOT to fry my batteries

I’ve been playing a lot with the EFJ 5100 series portable radios. The radios are excellent units by my standards (I will admit to not yet having used an XTS 5000). I have a couple of original EF Johnson battery chargers for the 5100 series. They’re moderately strange looking items, allowing you to place the radio in as normal for charging on the rear battery contacts, but also a slightly unconventional method of strapping the battery to the charger.

But that’s not my pet peeve with these units. What really frustrates me is how hot the batteries get when it charges them. They come out nearly hot to the touch. I know these so called “rapid chargers” are pretty hot stuff, but I have my qualms with units that are fixed to “mach 6” and literally cook your batteries while you wait. You’re helpless to charge them any other way unless you use staples and alligator clips to charge them with your “all in one” 5 amp battery charger (because everyone has one of those in their back pocket just like me…)

So finally I’ve grown frustrated and decided to try and bust one of these things open. It wasn’t too bad on the disassembly. Two screws obscured by rubber feet, and two by the silver label. It popped right open without much fuss. There was a PCB quite tethered with an umbilical cord of wires to the upper half. Interestingly, clamping a battery to the unit seems to engage an internal thermistor as opposed to the built-in-the-battery one used on the contact based charge.

Inside it revealed a very friendly looking (save giant gobs of glue) through-hole PCB with lots of resistors and filter capacitors and transistors and one inductor of some sort. The brain of the unit is the TI bq2004, a fast charging IC for NiMH and NiCD battery types.

A lot of the circuit is pretty complex (especially the analog current control circuitry, made difficult to reverse engineer as some of the copper traces are obscured and some enter beneath the giant inductor glue blob with no obvious return). What was easily decipherable was the selector pins for determining the charge rate. By grounding, floating, or tieing them to Vcc, it’s possible to select the charging rate from a list of selected values. In this case, the selected value happens to be the second lowest value available (a c/2 charging rate) and the slowest rate which utilizes the top-off charge to ensure the batteries are full when the green light pops on.

The current detection is all based upon the reference voltage value received by a control pin in the IC. By constructing a circuit that will translate the current into a voltage, the IC can adjust the current via pulse width modulation to achieve optimal charging rates. Several example circuits are given, but it wasn’t particularly clear to me what the circuit on the PCB was doing (as some more complex designs can tie into the thermistor, allowing control of current based on temperature). The presence of a 1.1 Giga-Ohm resistor puzzled me further.

However, it looks like the solution will be to drop the charge current to the lowest setting (c/4 setting) and see if there is an improvement. Obviously the circuit was designed for a different charge rate, but it’s my hope that this will improve the negative symptoms. This will involve grounding the selector pins on the IC with two resistors while trying best not to disturb anything else in the circuit. If this is unsuccessful, replacing the current control pin with a more restrictive resistor value will be tried, hopefully not obliterating some of the artistic qualities of this circuit. A picture of this PCB is below:

I’ll have more updates once I get soldering.