Li-Ion Charger IC for $1.66

[Doug S]

[ QUOTE ]
CM said:
Klaus, I had 10uF on hand so I just used that. No reason other than that. I agree the part is large but it's not going into your pocket so size doesn't matter. (OK, again, knock it off you guys /ubbthreads/images/graemlins/grin.gif)

Cris

[/ QUOTE ]
In fact for the board itself, bigger will be better since as a linear regulator power will need to be dissapated. The IC would benefit from a square inch or so of copper. Beyond that, added area has diminishing returns unless you also go to heavier copper such as 2 or 4 oz.

 

[CM]

Yes, copper, especially for the ground pin is desired. If you want to run this off a 9V or 12V supply, I'd suggest dissipating the power elsewhere besides the controller, probably using a 7805 regulator that can dissipate the I*V or even a beefy resistor in conjunction with that. Inefficient yes, but you don't want to fry the IC. Ideally you would run this from a 5V supply. I plan on using a brick type supply since it has +5V that can source several amperes of current. Probably overkill, but I already have the part. If you need help with power dissipation issues, PM me and I can help.

CM

 

[Doug S]

It occurs to me that this IC is also a candidate for use as a very low parts count linear constant current regulator for driving LEDs. This idea was kicked around awhile back for a different Li-ion charger IC. See this thread for the general idea:


Charger IC as a linear CC driver

 

[highlandsun]

That thought crossed my mind too, but these parts require 4.3V or more input, and spit out 4.1V output. The output is a bit high for continuous drive of an LED, don't you think?

 

[Doug S]

[ QUOTE ]
highlandsun said:
That thought crossed my mind too, but these parts require 4.3V or more input, and spit out 4.1V output. The output is a bit high for continuous drive of an LED, don't you think?

[/ QUOTE ]
Howard, the output voltage will not exceed 4.20V but will droop to whatever lower value is needed to not exceed the current set by the program resistor. The IC will continue to operate down to an input of 3.55V [typical] before the UV lockout kicks in. This would be a suitable linear CC regulator for 4 alkaline, NiCd, or NiMH cells. Once the UV lockout threshold was reached, dropout would take the form of the regulator cycling on/off at a very fast rate resulting in decreased average current.

 

[CM]

I wonder if using an LDO wired as a current source would work just as well and cost less. Any thoughts on this? We must also keep in mind that these methods (linear) is not very efficient, though they are the simplest methods of regulating current. Don't you guys just love these brainstorming sessions? Lots of "out of the box" thinking going on here.

CM

 

[LED_ASAP]

You can make a constant current driver easily from LDO's, but to properly charge a Li-Ion, you need extra circuitry to stop the charging when the battery is full. IMO it's better and simplier to use specialty IC's in the charger.

 

[Doug S]

CM, the LDO voltage regulators that I have looked at do not present an easy why of converting to CC regulation without losing the advantage of LDO. These Li-ion charger IC's, on the other hand, present an easy way of setting the current. In the case of the one under discussion in this thread, a single resistor sets the current value. I must admit that a LM334 based circuit is still my favorite for CC linear regulation. These LTC4054 circuits, however, would be very easy to fit into a PR style bulb base. I can fit a LM334 circuit into a PR base but it is a lot more work when doing hand assembly.

 

[CM]

LEDmodsASAP,

I think the topic inadvertently changed from Li-Ion chargers to CC drivers for LED's. Sorry for straying off topic. And to continue the off topic discussion...

DougS,

I see where you're going now. Yeah, this will fit into a PR base rather easily. Hmmm. Just a sot-23 and a chip cap and you got a LED CC regulator.

Maybe a new topic is called for? /ubbthreads/images/graemlins/grin.gif

CM

 

[koala]

I see a potential exploitation for that Li-Ion Charger!!! /ubbthreads/images/graemlins/grin.gif

 

[CM]

I got a breadboard ready to go. Just need parts. If it works, it should work great with a single Li-Ion cell especially if you have a low Vf LED. It should have the efficiency of a resistored DD, which is good of you're not dropping too much voltage, but you get constant current regulation until Vbat <= Vf. Would be interesting to see what happens after that. You do need to pay attention to the IC's power dissipation since it is a linear circuit. Linear Tech just needs to make a switcher for charging that has the same simplicity as this one. Anyone know if they do?

CM

 

[Doug S]

[ QUOTE ]
CM said:
I got a breadboard ready to go. Just need parts. If it works, it should work great with a single Li-Ion cell especially if you have a low Vf LED. It should have the efficiency of a resistored DD, which is good of you're not dropping too much voltage, but you get constant current regulation until Vbat <= Vf. Would be interesting to see what happens after that. You do need to pay attention to the IC's power dissipation since it is a linear circuit. Linear Tech just needs to make a switcher for charging that has the same simplicity as this one. Anyone know if they do?

CM

[/ QUOTE ]
CM-
One consideration is that the UV lockout will prevent extracting all of the available energy from a Li-ion cell. The UV triggers on 3.55V falling [typ] but does not reset until 3.8V rising [typ]. The 3.55V isn't too bad, especially for more current production [as compared to older designs] cells. The 3.8V level is a bit of a problem since there is a fair amount of capacity remaining at Voc=3.8V. One workaround is to leave the IC energized full time [it only draws a few uA] and do the ON-OFF switching at the output of the IC. If you get an IC whose UV threashold and/or hysterisis values are on the high side of typical, considerable energy of a single Li-ion may be unextractable.

BTW, regarding switcher based chargers, both LT and Maxim make them but not with the 3 component simplicity of this IC.

 

[CM]

Hmm, I overlooked that one. Thanks for the headspace check.

I did see some switcher based chargers and dismissed them from the parts count standpoint. No free lunch, again.

CM

 

[zmoz]

Hi guys...it's been a while since I've had time to visit this forum. This thing looks like something even I could build - I've been wanting some way to cheaply charge bare lithiums for a while now. There is one thing I am confused about - is this meant only as a charger, or will this also protect the battery from over discharge while in use...like the circuitry that is on a cell phone battery?

 

[CM]

zmoz, it's strictly a charger. The beauty is the low parts count and ease of construction (guess you saw my picture above) I've been endlessly on the lookout for a good home brew on-the-cell discharge protection like the Copias and Pilas. Been tempted to take a Pila apart to see what's under the hood but I don't want to render a $31 cell worthless. My guess is that they have an IC, two FETs, some resistors and cap on the (-) terminal.

 

[BobVA]

What about a "low battery" light in lieu of a battery protection circuit? Cobbling one up from op-amps that would fit in a small light is probably a non-starter, but I'm wondering if there might not be some one-chip SMT solution?
It wouldn't be idiot proof, but better than nothing....
Cheers,
Bob

 

[Doug S]

[ QUOTE ]
BobVA said:
What about a "low battery" light in lieu of a battery protection circuit? Cobbling one up from op-amps that would fit in a small light is probably a non-starter, but I'm wondering if there might not be some one-chip SMT solution?
It wouldn't be idiot proof, but better than nothing....
Cheers,
Bob

[/ QUOTE ]
I would rather go for idiot proof. It can be done for a dollar using only 2 components in SOT23 packages. Use a voltage detector IC and a mosfet. Both Seiko and Panasonic make suitable voltage detector ICs that operate on only 1uA. They are both good but the Seikos are cheaper, about 50 cents from Mouser. For another 50 cents you can add overcurrent protection in the form of a self-resetting PTC fuse.

 

[CM]

Doug, the Seiko's are the ones I'm looking at /ubbthreads/images/graemlins/grin.gif Cell mounting can be tricky and I don't want a fragile setup. THe circuit part is easy, the construction of it on the cell itself is the real tough part. So far, the Seiko's look like the best solution. Part of the reason for wanting to disassemble a Pila is to see if that's what they're using in there. I bet it is.

CM

 

[Doug S]

CM,
I have never tried to build a Pila single cell equivalent but I have built a dozen or so multicell packs with this approach with excellent results.

 

[zmoz]

Might it be possible to find an old cell phone battery or something of similar voltage and use the protection circuit from that?