Help modifying a cheap tablet for battery-less operation

[kizmit99]

Hi All,
I have a small (older) android tablet that I use to run TouchDRO in my shop.  This tablet has a single Li-ion battery inside it, similar to this:


Over the years I've been using this tablet I've had to replace the battery several times.  I think the underlying issue is that this tablet remains in the shop and it can get below freezing in there during the winter (honestly not sure that's the reason they fail, but fail they do).

The last time one failed I made an attempt to make the tablet work without the battery installed, and simply run off the USB charging cable.  I didn't do any in-depth analysis at that point, just that the tablet wouldn't operate with no battery at all, and seemed to work if I replaced the battery directly with USB power.  In practice, this has been somewhat hit-or-miss in how it works, it sometimes works ok, sometimes it can fight me for quite a few minutes before it will eventually boot up.  Below are a couple of pics of how I made the conversion and what I'm working with:


This is the battery protection board from one of the replaced batteries:


Close up of the battery charger on the main board:


I've decided I'd like to make this work better and did a little research on what I'm dealing with...  The battery protection board is just that, a very typical battery protection circuit.  Very similar (likely identical) to the following circuit:


The datasheet for the battery charging chip (5056) on the main board is attached below.  I have not traced how it is wired up in this application, but I would be surprised if it was anything other than the typical usage described in the datasheet - with the exception of the ~CHRG and ~STDBY pins, which I suspect the processor may be able to monitor...

In that last image you can see two additional (red and black) wires directly above the battery wires.  I believe these are likely driving the backlight for the display, and I suspect, given their location on the board, that their power is being derived directly from the battery - which would explain the tablet not working when no battery is present at all.

So, given all that -- I'm hoping someone might be able to suggest a way to easily convince the main board that the battery is present when in fact it isn't...

I've considered:
* Just use a resistor pair to generate 4.2V at the BAT pin of the 5056.  I suspect this would make the 5056 think the battery was fully charged and have it stand-down, but I also suspect the backlight wouldn't work.
* Feed USB power directly to the battery input pins on the board, remove the 5056 chip and tie ~CHRG and ~STDBY high (in case they are monitored).

I suspect there's a better solution, and hoping someone might point me in the proper direction to find it.

Thanks for looking, and thanks in advance for any suggestions...

[TheMG]

I think the underlying issue is that this tablet remains in the shop and it can get below freezing in there during the winter (honestly not sure that's the reason they fail, but fail they do).

Is the tablet ever connected to a charger while it is below freezing? Charging li-ion batteries in freezing temperatures is no good for them at all.

[kizmit99]

Is the tablet ever connected to a charger while it is below freezing? Charging li-ion batteries in freezing temperatures is no good for them at all.

Good to know -- but I usually have the usb power off when I'm not in the shop.  It's always possible I forgot to hit the power switch sometimes of course...

[fzabkar]

Replace the battery with a 4.2V DC-DC converter and rewire the DC-in jack so that it powers the converter rather than the tablet's power circuitry. In that way the tablet will think that the battery is fully charged. Also disconnect the charging electronics from the battery, if necessary (probably not).

You could use an external DC-DC converter (probably better). Just remember not to plug in the original 19V (?) adapter.


Edit: I just noticed that you have already considered a similar approach. Sorry.

[sleemanj]

>  In practice, this has been somewhat hit-or-miss in how it works, it sometimes works ok, sometimes it can fight me for quite a few minutes before it will eventually boot up.

Your problem is almost certainly capacitance, or rather, lack thereof.  When it attempts to boot it draws a fairly significant current, which pulls the voltage down, and causes the device to think the "battery" was dead.  A charged battery will have plenty of capacity right there, your power supply is over relatively long wires.

Remove the battery, and protection board, on the main board where the battery used to connect put a nice big electrolytic capacitor, and your 4.2v power supply wires. 

[Lona_Joseph]

Just make sure it's getting atleast 3.7v

[kizmit99]

Thanks for the suggestions...

I did a little more testing today and believe I've confirmed that my assumption about the backlight being driven from the battery power is incorrect.

I also tested with just a capacitor across the battery terminals on the main board - no better than my original hack (which already had a cap, but also the battery protection board).

Also tested driving the battery terminals with a bench supply - again, no better, but I could see the battery charger chip affecting it (the power supply).  At various voltages (from about 3.5 to 4.3V) the current driven from the PS would vary quite a bit (from almost nothing to a couple hundred milliAmps), sometimes in a pulsing pattern (with a constant voltage).

I also tried a single 16450 Li-ion cell across the battery terminals.  Now this was very surprising, but it still booted erratically (meaning it takes several power-button presses to finally get it to boot up, and there doesn't appear to be a clear pattern for getting it to boot consistently).

I'm beginning to suspect that the erratic start up behavior I'm trying to eliminate may just be the way this tablet starts up from a power off state (erratically).  Everything I've tried seems to behave about the same.

I decided to just break down and buy a few small Li-ion batteries in a form factor that will fit in the case and live with the more frequent than desired dying of the cells.  I purchased a multi-pack from AliExpress that are smaller capacity than the original battery (so I would expect severely limited on-battery life) - but I only paid about $8 for 4 or 5, compared to about $15-$20 for a 'proper' replacement...