Overheating Li-Pol Charger

[Overclocked]

No, Not the batteries The charger itself. Its based on the MCP73831 IC (Link to Datasheet). The Schematic and Board I used are included in the zip file (Eagle PCB files). This is the 2nd Time Ive redid this board, the first time the IC kept overheating as well. The 2nd time I reworked the PCB, and even before soldering, I put in a little Arctic Silver (Thermal Compound for CPUs) in the space underneath but this doesnt seem to alleviate the heat problem.

I am using a Supply of 4.34V and a charge current of 200mA. I am actually charging a LiFePO4 battery, hence the diode, to drop 0.3V. It still works and has actually completed charging on One battery. LiFePO4 batteries need 3.8V Max but are charged the same way LiPolys Are. But In any case, The Voltage measured at the LiPol Terminal Confirms this, so the battery isnt the problem.

On my Supply I have a current meter, And it swings, as does the voltage. it goes from 200mA down to 0 and back up. The Output voltage swings from 3.7 down to 2.7 (or so). Theres no way this is normal. The IC is Hot, Like a hot cup of tea you just stuck your finger in.

I dont know what to do to Resolve this. A small Fan works But My original Plan was to Charge Batteries via Solar. If I need a fan to cool the IC off, it would defeat the purpose. I could technically use a LM317 circuit (theres a few floating around the web) and use a PIC micro to monitor the Voltage and current But, Why over engineer something?

 

[Boncuk]

Please refer to the datasheet page 17 (6.1. Application Circuit Design).

If the chip gets as hot as 85deg/C I don't see a problem there, although I like cool electronics.

Boncuk

 

[Overclocked]

Please refer to the datasheet page 17 (6.1. Application Circuit Design).

If the chip gets as hot as 85deg/C I don't see a problem there, although I like cool electronics.

Boncuk

Thank you I changed the resistor and lowered the Charge current to 100mA. However I now have some more info that suggests what could be wrong. With NO battery, It still pulls 100mA and gets Hot. I find this odd and interesting. Another key fact is the fact that The indicator LED does not light up apon power up when its "finding" the battery (Rev 1 did this, and then the LED shuts off if there is no batt present). I suspect theres something going on with the PCB in this area, however, An Ohm Check and continuity check doesnt confirm this. Nothing Goes to ground in that Area. Its possible the LED is bad, but a diode check shows 0.770 Volts for a Green LED (In my experience green LEDs should be around 1.2-1.3v I think).

Im open to suggestions.

 

[Boncuk]

Hi Overclocked,

I guess the chip is already damaged. With no battery connected or charge complete it should not draw more than 200µA (maximum). The nominal current for that state is given as 53µA.

When no battery is present the status output should be in hi-z state.

Green LEDs normally drop 2.4V.

Boncuk

 

[Overclocked]

I guess I counted my chickens before they hatched-so to speak. I let it cool down overnight, plugged it in the next morn and The LED came on, Voltage output was Normal. Ok Still very Odd behavior. I begin to think about how rev 1 blew up (physically!). The input cap blew up or something and I had reused just the resistors and the output cap. I know the resistors were good-they had to be. For some reason I decided to check the output cap. Oddly enough the 4.7uF Cap read 10uF. Theres something wrong with this picture. Is the Cap too Small? Is the ESR too large (I was asking myself this) I used tant electrolytic's.

So, When I got home from work, I started up the iron, Went through my parts bin of scrap parts I salvaged and found a bigger Cap (10uF). I put it in, and all seems to be fine. I still dont want to count my chickens before they hatch but it doesnt seem to be heating up with a charge current of 100mA . Maybe I should use a more realistic linear supply instead of a switching psu. I will update this thread if it acts up again. Or Start from scratch. I really didnt want to give up on this thing and start with another IC. I mean, why design a IC with a certain operating range if its not supposed to operate there?

Now to start thinking of other things. Granted this IC is wired for Li-Pols, But Id like a LED to light up when it reaches 3.8V. Im thinking a comparator can do this quite well, but for some reason I still want a MCU to be involved some how to measure voltage and current. For now I'll stick with simplicity.

ADD: I see the mistake I made with the status LED in the schematic.