Maker Pro
Maker Pro

Replacing NiMh in phone

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
I have a number of devices which use a pair of NiMH batteries.

Clearly the charger is rubbish because the batteries get quite hot whilst on charge (which they are almost 24x7)

My thought is to replace a pair of these batteries with a single 10440 LiFePO4 cell with a "charger" in the space used for the other cell.

Here is my idea:

attachment.php


U1 is a "voltage detector which has a low output when the voltage exceeds 3.2V and high when it exceeds it (there is actually a 5% hysteresis, see here).

Q1 is an 8V P channel mosfet that can pass about 10A at Vgs = 2.5V. (See here) The purpose of this is to prevent the battery from charging once the battery voltage exceeds 3.2V.

D1 is a schottky diode which provides a discharge path when the mosfet is off (with higher performance than the body diode).

OK, so the idea is that the existing charger limits the current to the battery, and the voltage detector stops the charging when the battery approaches full charge. D1 reduces the voltage drop across Q1's body diode while the battery voltage remains above about 3.2V.

Does anyone see any problems with this? (edit: apart from the fact that the first version had the mosfet around the wrong way!)
 

Attachments

  • BattRepl.png
    BattRepl.png
    12.6 KB · Views: 115
  • BattRepl2.png
    BattRepl2.png
    10.2 KB · Views: 272
Last edited:

OLIVE2222

Oct 2, 2011
690
Joined
Oct 2, 2011
Messages
690
Steve,

You can probably use the existing charger as supply for the Lithium charger but Lithium batteries are far less forgiven than NiMH. They need a preconditioning phase, a current regulation phase and (a very precise) voltage regulation and charge termination phase, sometimes a safety timer and a t° sensor too.

Also the nominal NiMh pair voltage is 2V4 (can be 3V if alcalines are allowed in your phones) the nominal Lithium voltage is 3V7 and the last charging phase even need a higher voltage. So if the existing charger can deliver 5V or more a dedicated charging IC followed by a LDO around 2V5 will be better a the task.

Olivier
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
The charging base itself provides around 9VDC.

The current to charge the batteries appears to be limited at the moment, and I'm prepared to accept whatever current that is. I'm also prepared to ignore any pre-conditioning phase as I just hope the batteries don't get that far discharged.

And the charge will terminate way before the maximum voltage for the cell.

I have to do some tests to see what maximum voltage it will try to charge the batteries to though -- it may not even get to my cut-off point.

The main thing I want to prevent is overcharging the lithium cells.

I was going to add an undervoltage cutoff, but the portable phones themselves seem to totally stop working at around 2.5V anyway.
 

OLIVE2222

Oct 2, 2011
690
Joined
Oct 2, 2011
Messages
690
The whatever current you are prepared to accept can be too high as AAA NiMh are in The 800mAh range (for recent one's). If the NiMh are charged around C/3 it' s OK, if it's higher you can exceed the maximum charging current, see:

http://www.batteryspace.com/prod-specs/Spec_for_10440.pdf

Don't know what real battery life and capacity you can expect with such "quick a dirty" circuit knowing that the new setup deliver around 1.3Wh (optimally at 3.7V) compared to the 2Wh (0.8Ah x 2.4V) of the previous setup.

Some Lithium batteries also include an over current/under voltage protection at around 2V5.
Are the phones stop working themselves at 2V5 when powered with a Lithium battery or when powered with a lab power supply ? should be with Lithium battery knowing that 2V4 in the nominal supply voltage !
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
I believe the batteries are constantly trickle charged in the phone. That's probably why they die so often.

The circuit I have there will terminate the charge on the battery as soon as it reaches about 3.2V, so they're not going to get overcharged. 3.2V is very low for a normal LiPo, but for LiFePO4 it seems to be OK (about a 50% charge).

The cells are about 250mAh, so I'm expecting to store about 100mAh of useable charge.

If the charge current is too high, or if the voltage is too low, I'll probably just modify the phone (In which case I might just as well incorporate a proper charger circuit).

The main question is whether the circuit will cut off the charge appropriately, yet allow discharge. I'm pretty sure it will.

The phones stop working when the NiMH have a terminal voltage at about 2.5V at rest. (Yeah, tell me about how much charge they can actually use -- it's the bit between charged and overcharged, but hey, they make millions of these, they must be good, right?)

I'm going to run some more tests using a bench supply later. There are some interesting details like the voltage on the battery terminals being about 1.5V when the phone is on the charger and no batteries are inserted. Clearly it detects the presence of batteries before trying to charge them. The phone works fine on alkaline cells, but I haven't tried putting them in the charger to see if it's smart enough not to try to charge them -- I suspect not.

I've been doing a lot of reading about LiFePO4 cells in the last couple of days. I'm seriously considering building a quick and dirty battery capacity tool using a microcontroller and a constant current source/sink so I can graph their voltage vs state of charge and measure actual capacity.

My initial tests are going to use very cheap Chinese cells, because I'm aware of the risks of killing cells. If I can keep them happy, I'll be happy to use other cells.
 

OLIVE2222

Oct 2, 2011
690
Joined
Oct 2, 2011
Messages
690
The battery detection feature can be an asset, knowing that the battery should appear as charged I guess that the provided current will be the trickle one, safer for the LiFePO4.
yep not sure that he can distinguish charged NiMh from new alkaline cells (nor the color or the brand :) ). A kind of battery analyzer will be indeed useful, especially because your are working with half charged cells, however discharging curve of LiFePO4's is very flat.
I am confident that you circuit will cut the charge as expected, I already use a circuit close to yours with a MAX6326 and a P-FET for another application and it's works fine.

if you succeed to keep the phones alive during the whole day under normal usage it will be just fine according to the few means involved.

Olivier
 
Top