Safe, (easy) in-circuit rechargeable battery

Hi guys & girls.

I have been perusing this site for anumber of months - an invaluable resource.

I am currently in the process of building a wireless hand-held controller. The main components are as follows:

- PIC 16F876A (5V) (may add another, smaller micro, to deal with graphics, user interface, if necessary)
- Colour LCD panel (5V, backlight may draw up to 200mA when on)
- Wireless module based on Nordic nRF2401 (3.3V)

These all work together fine, but I am still powering it with a linear regulator and wall wart. I need this to be battery powered, and have "in-circuit charging" capability. I want the controller to be useable while it is charging. Small form factor would be a bonus, as would the use of common, relatively-easy-to-find components It seems my options are as follows:

- NiMH batteries. 3*1.2V = 3.6V. A boost regulator to step up to 5V, followed by a buck regulator, capacitor switcher, or a LDO regulator to step this down to 3.3V. This should also handle regular alkalines being inserted. Charging can be done with a resistor/diode within a few hours

- Li-Ion or Li-Po flat batteries. These look great, and the chance to make something with such a small form factor is alluring! But after reading lots of horror stories, I'm scared. Maxim and Microchip seem to have a few charging ICs which do the work, but I have never produced anything like this before -- are LiI batteries too dangerous for hobbyists to be playing with? The example schematics in the MCP73841 datasheet are clear enough (pick a few external components to comply with battery data sheet)-- but they all show a dedicated charging circuit, and make no mention of "in-circuit charging". Can the load still be connected to the battery when it is being charged thorugh such a controller? Since the charging IC measures and controls current sunk into the battery, I don't see how it could work if an external circuit were connected to the battery terminals.

Any alternatives, help, pointers, or personal experiences, would be much appreciated.

John

If it was me, I'd just use a 4 pack of AA NiMH batteries. The PIC will run just fine on full charge (almost 5.8v) down to practically dead (figure a bit above 3.6v).
The LCD should operate just fine in that range also. I don't know exactly which LCD you've got, but I know my parallel types run down to 3.5v ok, only problem is the screen wants to update just a bit slower and the contrast changes a bit.
As for the wireless module, again, haven't looked at the datasheet. I may very well be 5v tolerant. If not, power it with an LDO 3.3v regulator or maybe an inline 1N4004 diode to drop the voltage by about .6-.8v.

As far as charging, find a suitable voltage output wall wart, say a 9v charger. Take your battery pack capacity, say 4x AA NiMH rated at 1800mah. Put in a polarity protection diode in case you plug the charger in backwards, and add a series current limiting resistor to keep the batteries topped off. In this case, max battery pack voltage will be around 5.8v as stated above. An 1800mah battery pack probably shouldn't see more than 1/10C when almost depleted and 1/100C when charged up (to keep the battery pack living longer that is).
So, that being said...
9v input minus a diode drop 9-.6 = 8.4v.
An empty 4 cell battery pack is about 3.6v charging at no more than 1/10C (180mah).
(8.4v input - 3.6v battery voltage) / 180mah = 26.6 ohm series current limiting resistor.
When the battery pack gets fully charged, this same current limiting resistor will limit the charge current to:
(8.4v input - 5.8v ) / 26.6 ohm = 97.7mah -- a bit too much (about 1/20C).
This will give you about a 12-13 hour charge from completely dead. But if you're like me, you'll forget about the charger quite often, so I'd go with the other end of the spectrum:
(8.4v input - 5.8v charge pack ) / 18mah = 144.4 (give or take) ohms.
With the 144 ohm resistor, you'll get the 1/100C rate at the end of the charge. Completely safe for NiMH batteries, but the downfall is that it'll take a few days to put a completely full charge into the batteyr. I've been using the 26.6 ohm resistor method above in my mp3 player I built for about 2 years now on the same pack without a problem. You might want to use anything between 47-120 ohms, take your pick. It's all about ohms law...

Personally, I think you're overdesigning the project. Don't get me wrong, not a bad thing, but sometimes...simple is best... Stay away from Li-Ion for now. They're not really hobbyist friendly quite yet...especially not for a beginner anyways...or even a well-versed amateur like myself.

JDG