Homemade Charger for NiMH AA's

[BrianG]

As some may know, I recently got a bunch of Turnigy AA LSD (low self-discharge) NiMH batteries, but wanted a portable charger for trips, or when I don't feel like dragging out all my R/C charger equipment. Most of those chargers you find at Wal-Mart/Target/etc for AA batteries are timed (at best), so can overcharge or undercharge the cells. The charger below is pretty simple, will charge two AA batteries at once, and uses a temperature cutoff. Basically, it's a constant-current charger with thermal delta-peak detection.

Here is the schematic and parts list:

• R1: Thermistor rated 100k ohms at 77*F
• R2: 15k ohm 1/4w resistor
• R3: 10k ohm multi-turn trimmer pot:
• R4: 1M ohm 1/4w resistor
• R5: 1K ohm 1/4w resistor
• R6: 10k ohm 1/4 resistor
• R7: 470 ohm 1/4w resistor
• R8: 2.4 ohm resistor made from five 1/2w 12 ohm resistors in parallel.
• D1, D2: 1N4001 diode
• D3: Green LED (Vf=2.1v @ If=25mA)
• D4: 3A barrel diode
• Q1: 2N3906 PNP transistor
• Q2: 2N3904 NPN transistor
• Q3: TIP120 NPN Darlington transistor
• SW1: Normally-closed momentary switch
• C1: 0.1uF disc capacitor

Circuit operation:

R7 and D3 (LED) together puts ~2v at the base of Q3. That makes ~0.75v at the emitter of Q3, which generates a constant battery charging current of 0.313A through R8. Diode D4 prevents the battery from backfeeding into the rest of the circuit. The LED (D3) lights up when charging and shuts off when charging is complete, so it is used for charging status as well as charging bias.

A thermistor is positioned to physically touch the batteries. As the batteries reach full charge, they start heating up. This causes the resistance of the thermistor to decrease. The voltage divider consisting of R1 (thermistor), R2, and R3 generates a higher and higher voltage at D1, until it can overcome the diode and turn on Q1, which brings the LED (D3) voltage to ~0.8v shutting off the charging transistor (Q3).

When Q1 is turned on, Q1 and Q2 transistors basically latch each other on and will keep the charger shut off until power is removed. This is to prevent the circuit from charging the batteries again once the thermistor cools. You use the switch (SW1) to cut power momentarily which resets the latch circuit. C1 just makes the circuit stable when resetting the latch.


Some assembly notes:

The voltage supply is a regulated 6v DC voltage. It is important for the voltage to be regulated for consistent temperature cutoff. If you do not want to use a regulated supply, you could regulate (using a zener diode) just the thermistor circuit (R1, R2, R3). However, if the supply voltage is much higher than 6v, there will be a lot of wasted heat on the charging transistor. So, it's best if the supply is a regulated 6v for best operation.

The charging transistor (Q3) must be on a heatsink exposed to the outside air. The hottest it will get is when discharged batteries are hooked up. As the batteries charge, their voltage will increase which means there is less voltage across Q3, and less heat. Worst case Q3 power dissipated is ~0.85w, and will go down to around 0.45w near the end of the battery charge.

The 313mA charge rate will take about 7 hours to charge a 2200mAh cell. That may seem like a long time, but ~0.3A seems to be the best charge rate for these Turnigy AA LSD cells.

To calibrate the thermal shutdown, I started with the trim pot resistance at the minimum value. Then, while charging, I monitored the cell voltage. When I saw the voltage peak and then start to decline (which signifies end of charge), I adjusted the trim pot slowly until the charge circuit shut off.

Some pics of the finished product:




I had a hard time finding the right size enclosure, so parts placement is a little odd. The switch is very big because it is the only cheap one I could find. The momentary normally-closed switches RadioShack has are absolute junk.




It's kinda hard to see, but I glued the thermistor such that it touches the cells when they are in the holder. And to improve thermal conductivity, I molded some thermal epoxy on the thermistor to conform to the cells. Once it cured, it worked pretty well.




The inside. Yeah, it's pretty packed in there.

On the far right is the input power jack. I was going to put the whole power supply inside (using a bigger box), but this enables me to use 12v from a car cigarette outlet, or a 120v "wall-wart" power supply. I put a 1000uF capacitor for a little better filtering in case the supply is a bit "noisy".

On the top right is a modified cheap Chinese $5 R/C BEC. I removed the jumper lead and feedback resistor, and installed a 2.2k ohm resistor to provide a consistent 6.14v. The BEC allows me to use pretty much any DC input from ~8vDC all the way up to ~35v. Pretty flexible. And, typical of a switching BEC, the higher the input voltage is, the lower the required current. The following is a list of power supply requirements at various voltages: 9v @ 300mA, 12v @ 220mA, 18v @ 150mA, 24v @ 115mA, 30v @ 90mA. Any of these will function perfectly.

The messy pack of resistors on the far left is the 5 paralleled 12 ohm resistors to make the 2.4 ohm current-programming value I needed. There simply wasn't enough room on that tiny circuit board for them. The rest of the components are on the board though, but it was a tight fit. The blue "box" with the round brass button on the bottom of the board is the trimmer pot. I use it to initially tweak the circuit until the charge stops at the right temperature.

To be clear: I do not plan to market this thing. I just figured people might want a small, cheap, and simple charger.

[magman]

Brian...that is just to cool....I am glad there is one electrical expert on this forum. I think NASA may be in need of your services

[Bondonutz]

Your like a Mad Scientist of the electric RC world Mr B., awesome stuff !

[magman]

Originally Posted by Bondonutz

Your like a Mad Scientist of the electric RC world Mr B., awesome stuff !

Definitely a little "mad" for someone to like electronics like that

[rchippie]

Originally Posted by magman

Brian...that is just to cool....I am glad there is one electrical expert on this forum. I think NASA may be in need of your services

Im the one who told him how to make it .

[BrianG]

lol, thanks guys, but this isn't exactly rocket science.

Hippie: You're gonna have to prove that statement. Here's a test for you: Why did you select the TIP120 as the charging transistor?

[rchippie]

Originally Posted by BrianG

lol, thanks guys, but this isn't exactly rocket science.

Hippie: You're gonna have to prove that statement. Here's a test for you: Why did you select the TIP120 as the charging transistor?

I choose the tip120 transistor because thats what i had laying around my apartment .

[BrianG]

Bzzzz! Wrong! The correct answer: because it is an inexpensive power device with very high current gain (because of it being Darlington).

[rchippie]

Originally Posted by BrianG

Bzzzz! Wrong! The correct answer: because it is an inexpensive power device with very high current gain (because of it being Darlington).

Says you .

[BrianG]

Update:

Made a couple of changes to the original schematic:

- Decreased the supply voltage (BEC output) to 5.8v instead of the 6.14v I was running, which results in less wasted heat.

- Reduced R2 from 10k to 6.8k ohms.

- Added another thermistor (R9) to the PCB to track changes in the enclosure's ambient temperature.

- Used a slightly larger (and prettier) heatsink, and added some cooling holes to the enclosure. I found the BEC was getting a little too warm for my liking since it was not getting any air.



Other than that, the charger seems to be working swimmingly.

[JERRY2KONE]

Originally Posted by Bondonutz

Your like a Mad Scientist of the electric RC world Mr B., awesome stuff !

Well I have to agree with Bondo on this one. You are quite the mad scinetist here on RCM. Nice work BG. I still think you should have done a run of 12 or so for us to purchase from you. Oh well. It looks like one of those Radio Shack projects that we did when we were kids. Bring on the "Erektor Set" kits. Thanks for sharing BG.

[BrianG]

lol, hardly a mad scientist! It really is quite simple. I posted this project at an electronics forum where I am a newbie when it comes to this stuff. The thread started out about transformer selection, but ended up debating LSD cells. So, RCM is not the only place where topics get sidetracked.

I want to start dabbling in PIC microcontrollers next since there is only so much you can do in the analog world. Just gotta find an easy to locate (and cheap) PIC, and a programmer board. Just think of all the things (read: damage) I could do with a processor! Mwa ha ha!

[JERRY2KONE]

Originally Posted by BrianG

lol, hardly a mad scientist! It really is quite simple. I posted this project at an electronics forum where I am a newbie when it comes to this stuff. The thread started out about transformer selection, but ended up debating LSD cells. So, RCM is not the only place where topics get sidetracked.

I want to start dabbling in PIC microcontrollers next since there is only so much you can do in the analog world. Just gotta find an easy to locate (and cheap) PIC, and a programmer board. Just think of all the things (read: damage) I could do with a processor! Mwa ha ha!

See you even sound like a mad scientist, only you do not realize it because it comes natural to you. ha ha ha ha

[BrianG]

Made a V2 charger today. Similar concept except a little better:



The same basic constant current scheme, but the thermal shutoff is accomplished via an LM339 comparator. The LM339 actually contains 4 independent comparators, but I am only using two of them. I used a diode in the positive feedback loop to accomplish the latching. Resetting the latch is done via a NO momentary switch. Also, this comparator has "rail to rail" output capability, which simply means the output can go "almost" as high as the supply and "almost" as low as ground. This is due to the open-collector output, but does need that pull-up 1k resistor. Most other comparators/op-amps have push-pull outputs and cannot swing the voltage nearly as much.

Operation is pretty simple. The 470 resistor and LED provides the bias for the TIP120 transistor creating a constant current of around 310mA under normal operation. The two 47k voltage divider resistors set up a reference voltage of 1/2 supply voltage to the inverting input of the LM339. The thermistor, 56k resistor, and 10k pot set up another voltage divider to the non-inverting input of the LM339. When the thermistor is cool, the LM339 output is low, which means the 2N3904 transistor is off, so the constant current circuit functions. Once the thermistor heats up, the voltage at the non-inverting pin of the comparator exceeds the reference voltage and the LM339 ouput goes high (through the 1k pull-up resistor), which turns on the 2N3904 transistor pulling the collector to ~0.1v thereby shunting the LED close to ground, and shuts off TIP120. The 1N914 diode provides an isolated feedback loop to keep the LM339 latched. This avoids having the charge come back on one the thermistor cools. Using the reset switch brings the + pin back low resetting the LM339.

Improvements in this circuit design:

- Fewer components per charge bank. Last design had 17 for one charge bank, this one has 29 for two banks.

- Uses a normally-open momentary switch which are much easier to find than normally-closed types.

- The use of a comparator IC makes the temperature trip point independent of the supply voltage, so there is no need for a tightly regulated supply.

- Used a Schottky diode for battery isolation, which provides a lower voltage drop than a regular diode. This allows me to use a lower supply voltage for better efficiency.

- More effectively shuts off charging since it brings the base of the TIP120 to ~0.1v (last design brought the base to ~0.8v). The last design still had a ~1mA charge current even when the TIP120 was off. This design still has a charge current, but is miniscule being around 45uA (that's 0.000045A). I don't think the TIP120 could ever be completely shut off (short of removing power) since there are a couple integrated bias resistors as it is a Darlington transistor.

I gutted an old RadioShack charger to make this version:

[What's_nitro?]

Nice... I bet it was easier to use a case with a battery holder built-in!