Simple Cell Reverse Protection Circuit For Charger

[MrAl]

Hi there,

Here's a simple protection circuit for your charger.
It will work with batteries with at least some
power left in them (typical of Li-ion's).

Take care,
Al

Circuit:

If you cant get the green LED to light you'll have to use
another Red LED instead and label them clearly.

Just in case your charger chip doesnt like putting out
the required current for the green LED once the cell
is charged you can always wire the LED to the switch
if you use a double pole type switch. When the switch
is closed, the LED will be disconnected.

 

[MrAl]

Hello again,

Here's the next generation circuit...

It puts hardly any current drain on the cell
or the charger chip output.

Take care,
Al

 

[djpark]

MrAl,

I was about to write about the first led circuit, but now I see the second gen.

I suspected that with some load (green led) on the battery voltage output from the chip, it might drop the output voltage slightly and the charger may not stop charging forever. Also the current through the led (though a few mA) will be added to the total charging current in the eye of the charger chip. With your 2nd gen, the extra current would be below 100uA and I guess no problem.

While the red led makes a perfect sense, I am thinking, however, if we really need the green led. Also I feel that the series resistor of the red led shall be a bit smaller (perhaps 100-150 ohm) to give a real "Ooops" with a brighter light output.

Will it be possible to use a fet with low Rdson to switch off if the cell is inserted reverse? Or add fet to the input power side to switch on when a cell is inserted correctly? I am not sure how this would affect the charging characteristic.

Also is it possible to determine if it is safe to insert the cell reverse from the LTC4054 datasheet? I read about the short circuit protection, but no mention about the reverse cell.

-- dj

 

[Doug S]

[ QUOTE ]
djpark said:
Also is it possible to determine if it is safe to insert the cell reverse from the LTC4054 datasheet? I read about the short circuit protection, but no mention about the reverse cell.

-- dj

[/ QUOTE ]
A reversed cell will kill the LT4054. My solution has been to place a MOSFET between the LT4054 output and the cell wired to provide reverse polarity protection.

 

[MrAl]

Hi again,

dj:
Oh yeah :-) i realized less current on the output would
be better, but you're welcome to only use the red LED
alone and increasing the current to it would be great also,
that's a good idea i think i'll use that too on mine.

Yes, it's certainly possible to use a MOSFET to sense
cell voltage and determine if all is well before allowing
the connection to happen. I've even considered using
one of those low coil current (20ma) relays for this kind
of relatively low current level. The MOSFET would have
to be one of those super low R(on) types in order to avoid
messing up the charge characteristic. Even 0.1 ohms added
series resistance can extend the total charge time as
much as 100 percent, but a 0.008 ohm device would work
very well. Of course you'll have to think out a plan
for how it's going to detect and how it will reset when
the cell is removed. Maybe a mechanical interface of
some sort?
Perhaps Doug S would be nice enough to show us his circuit /ubbthreads/images/graemlins/smile.gif
I checked over the data sheet and couldnt find anything that
mentions cell reversal so you know what that means...
if it's not on there, they are trying to avoid the issue
and there's probably a good reason for that. According
to Doug S here it will blow out the chip, so it looks like
maybe it's time to do something to avoid that. On
my charger i'll probably just go with the switch and
two LED's for now...until we think up something better /ubbthreads/images/graemlins/smile.gif

Doug S:
Care to share your circuit with us?


Take care,
Al

 

[Doug S]

[ QUOTE ]
MrAl said:
Doug S:
Care to share your circuit with us?

Take care,
Al

[/ QUOTE ]

Al, I'll visit someone with a scanner [I don't have one] and scan the schematic of the chargers that I built for the Li14430 light project. They incorporate this feature. I'll email it to you since I don't have any webspace to host it. You may post all or relevant portions of it as you see fit.
BTW, I used a 0.030 ohm/typ N-ch in SOT-23 pkg. I cannot tell that it has appreciable affect on charge time. I use a reverse indicator LED as you suggest too. If the indicator is ignored, the charger will discharge the cell at the lower "precharge" rate where the IC thinks the cell voltage is below 2.9V.

 

[Doug S]

[ QUOTE ]
MrAl said:

I checked over the data sheet and couldnt find anything that
mentions cell reversal so you know what that means...
if it's not on there, they are trying to avoid the issue
and there's probably a good reason for thatTake care,
Al

[/ QUOTE ]

Al, the datasheet in effect covers this under "Absolute Maximum Ratings". The permissable voltage range for the BAT pin [relative to GRD] is -0.3V to 7V. As you and I know, whenever you see that -0.3V limit for a pin voltage there is a parasitic substrate diode lurking behind that pin that will conduct if the pin is driven sufficiently negative. BTW, I emailed you a schematic to your profile email addy. I assume you got it.

 

[MrAl]

Hello Doug S,

Oh silly me :-) I looked too fast i guess and missed that.
Yeah, if something drives the output pin below -0.3 volts
the chip wont like that at all, which also means my idea
of using the high current schottky might not work either.

I got your circuit in the email, thanks. I dont understand
how the MOSFET protects the circuit however...what is the
idea to connect the battery BEFORE applying power to
the chip?
I'll post the schematic if you like...

Take care,
Al

 

[Doug S]

[ QUOTE ]
MrAl said:
Hello Doug S,

Oh silly me :-) I looked too fast i guess and missed that.
Yeah, if something drives the output pin below -0.3 volts
the chip wont like that at all, which also means my idea
of using the high current schottky might not work either.

I got your circuit in the email, thanks. I dont understand
how the MOSFET protects the circuit however...what is the
idea to connect the battery BEFORE applying power to
the chip?


[/ QUOTE ]
Order of power application doesn't matter. Yes, it is a bit subtle but if you study it a bit it will come to you. Big hint: with the cell installed backwards the MOSFET will end up operating in its linear region.

 

[MrAl]

Hi again Doug,

Oh yes, very nice!

Im sure more people would like to use your protection
idea if you dont mind. I can draw up a schem of the
protection part alone and post it.

BTW, do you have anything on hand for parallel cell
protection?

Take care, and thanks,
Al

 

[Doug S]

[ QUOTE ]
MrAl said:
Hi again Doug,

Oh yes, very nice!

Im sure more people would like to use your protection
idea if you dont mind. I can draw up a schem of the
protection part alone and post it.

BTW, do you have anything on hand for parallel cell
protection?

Take care, and thanks,
Al

[/ QUOTE ]
Sure, Al, use that any way you see fit. Be sure to mention that for this to work the Mosfet needs to be a LL gate threshold type. That particular one I specified on the schematic I sent you turns on in the 1 to 1.5V range for the currents involved. If a high Vgs threshold type is used instead, the BAT output of the IC may rise above 2.9V enabling the full setpoint charge current. This will force the MOSFET to dissapate too much power, certainly more than a SOT-23 can handle.
Sorry, I haven't given much thought to the parallel charging issue. All of my parallel Li-ion cell applications involve cells that are assembled into packs. They, of course, are already self equalized before installation into the charger. As a practical matter, as long as two cells don't have grossly different Voc, I would not hesistate to parallel them in a charger. I'd say within 0.2V would be OK.

 

[MrAl]

Hi again Doug,

Here it is:

Take care,
Al

 

[djpark]

Thanks a lot Doug & MrAl. That is wonderful.

Any suggestion on the FET part number?

-- dj

 

[Doug S]

[ QUOTE ]
djpark said:
Thanks a lot Doug & MrAl. That is wonderful.

Any suggestion on the FET part number?

-- dj

[/ QUOTE ]
There are many available N-ch FETs that would do. On LTC4054 chargers I have built using this protection I used IRLML2502 which is an IR part in a SOT-23 package.

 

[djpark]

[ QUOTE ]
Doug S said:
There are many available N-ch FETs that would do. On LTC4054 chargers I have built using this protection I used IRLML2502 which is an IR part in a SOT-23 package.

[/ QUOTE ]

Thanks. It sure is nice with Rdson=0.045ohm.

I tried to get this part but can't find locally. The closest I can get is IRLML2402 with Rdson=0.25ohm and Si2302DS with Rdson=0.085ohm. Sigh..

 

[Doug S]

djp: that Si2302DS will be fine for your needs. The slightly higher Rds is of little consequence.

 

[djpark]

Thanks, Doug.

 

[MrAl]

Hi there,

Hmmm, i might be a little concerned with 0.085 ohms
series resistance. It could extend the charge time
significantly. There's a simple test however, without
buying one first...
Connect a resistor of 0.085 ohms (or close to it)
in series with the depleted cell and time the charge
cycle. Compare with the time to charge with no series
resistor.
If the internal R of the cell is high (0.3 or so) it
might not matter, but for my cells it would almost
double the charge time.

Also consider using two in parallel :-)

Take care,
Al

 

[djpark]

That FET is already in my possesion, the lowest Rdson I could get from local RS.

I haven't done the logging of the charge time with it. But I rmember seeing the charge current drop significantly when I added a DMM in serires.

But another time I noticed that the charge stopped with cell voltage lower than expected such as 4.1V.

I will need to try again to establish some consistent behavior.

-- dj

 

[MrAl]

Hi again,

Oh, it's worth a shot then :-)

Let us know how it affects the charge time please?

Take care,
Al