Lithium battery charging mythology?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I have seen in a number of places dire warnings about the charging regime for lithium chemistry batteries.

These suggest that once a lithium battery has reached full charge the charging voltage must be removed, else the battery risks bursting into flames or exploding. It has been suggested that the cause of this is solid lithium plating out onto the electrodes.

Reading the National LM3420 lithium charge controller datasheet, I find only the following 'Once the charger is in the constant voltage mode, the charger maintains a regulated 8.4V [*for 2 cells] across the battery and the charging current is dependent on the state of charge of the battery. As the cells approach a fully charged condition, the charge current falls to a very low value.'

*My parenthesis.

...As the cells approach a fully charged condition, the charge current falls to a very low value.

This is what one would expect.

Nowhere in the datasheet can I find any suggestion that the charge controller removes the charging voltage once the battery is charged. In fact the LM3420 is merely a precise voltage reference to be employed in a standard constant voltage/constant current circuit as per the LM317 datasheet, it is intended to permit the construction of such a charger without the use of a trimpot for final adjustment of the output voltage.

Can anybody direct me to a manufacturers datasheet or other reliable source which suggests that using such a circuit presents a safety hazard, or can I safely assume that the dire warnings seen elsewhere are merely further evidence of the internet myth-generator at work?
 
I have seen in a number of places dire warnings about the charging regime for lithium chemistry batteries.

The dire warnings are quite true. :) I do a lot with high power LED flashlights that use lithium 18650 and 26650 cells. Here is what happens when a lithium ion cell is overcharged:

https://www.youtube.com/watch?v=SMy2_qNO2Y0

There is a photo over on candlepowerforums.com of a shorted lithium ion cell in series with others in a flashlight, causing polarity reversal in the shorted cell, which resulted in the aluminum flashlight tube exploding and taking out an entire sliding glass door. Luckily the owner heard a "pop" and dived behind a sofa in time. :)

This writeup is pretty accurate with current thinking on charging lithium ions

Charging Lithium-Ion Batteries ? Battery University

Note the second paragraph and table 2.

Some older charging chips for lithium cells did use to try to "maintain" the cell at 4.2 volts at end of charge. That didn't work out as the article explains. Current thinking in the flashlight community is best to cut off entirely at 4.2 volts, which is what the newer chips do. I see the datasheet on that national chip is dated 2000.

A good example is the venerable Ultrafire WF-139 charger that a lot of flashlight folks use for 18650 cells

Ultrafire wf-139 Lithium Ion battery charger 14500, 17500, 17670, 18500, 18650, 18670

The original version of that charger had a chip that "maintained" and went up to 4.3v at end-of-charge. It was famous for killing 18650 cells in quick order. :) The latest version out for the last couple of years has a chip that cuts off entirely at 4.20V. Works great. With lithium-ion that extra 0.1V really, really matters. My Sony laptop here (using 18650s in the battery pack as many laptops do) has a "battery preservation" setting that I use, which charges the cells to just 80%. From that article you can see in table 2 that is 4.1V, just 0.1V less.

I hope this helps!
 
Last edited:
I agree... Flat charging at the top to maintain the cell at a constant voltage is not a good odea. Let them self discharge dow a ways and then top them off again. THis is easily done in a micro. I do the same sith SLA and NiMH batteries... Same algoritm of CC, CV for top off and then let them set until it's time to charge them up again.

Tony
 
It's a relatively new technology.

Most of the chargers that I am aware of used in the R/C hobby stop charging when the battery is full.
Given charging is more aggressive to get faster charging times.

These batteries are more sensitive over and undercharging. Things get more complex if individuals start to drift over charging cycles, making balanced chargers ideal.

It appears your chip is not a balanced charger.

If you have a chip and follow the guidelines then I suppose you answered your own question.

For an application like this, I myself would prefer something that charges the cells individualy and shuts off when charged.
These batteries have excellent shelf life for a rechargeable, I can't see a real advantage to continue charging vs the risks of over charging.
 
Hmm. 'Battery University'. I'd seen that site already. Here's a quote:

'Charging and discharging batteries is a chemical reaction, but Li-ion is claimed as an exception. Here, battery scientists talk about energies flowing in and out as part of ion movement between anode and cathode. This claim has merits, but if the scientists were totally right then the battery would live forever, and this is wishful thinking. The experts blame capacity fade on ions getting trapped. For simplicity, we consider aging a corrosion that affects all battery systems.'

This is not a standard of discourse calculated to inspire confidence in me. It's about the level of hogwash I expect from somebody selling hi-end cables. The site is just an advert for a book, and on the basis of what I gleaned, not one I would buy.

A reversed-charged cell in a flashlight blew up. What's that got to do with the price of cheese? The YouTube explosion was deliberately provoked by overcharging.

If I was of a nervous disposition I might be affected by these 2 pieces of 'evidence', but they're irrelevant, and worse, PREJUDICIAL.

I have many LED flashlights, Fenix, Ultrafire, Aurora with Cree, Seoul Semi and Luxeon LEDs. I stopped visiting candlepowerforums because I couldn't take those guys gloating over blowing more money on a pocket light than lots of 3rd. world dwellers have to live on for a year.

Anyway, enough of that...

Does anybody know of an integrated charge control chip that employs a more complex regime than the one I quoted from National (now TI)? Or a circuit in commercial production?

Yes, it's true that it's possible to use a uproc to switch charging on and off and it's possible to employ balanced charging, I fly model heli's so I'm familiar with such implementations, but it's all expense and board real estate. In the case of the heli's the cells are so abused, charge and discharge >> C, that balanced charging offers tangible benefits, but in a portable electronics device the current is << C. In many smaller heli's unbalanced chargers are employed. I have 2 sitting on the windowledge.

I have other features I am trying to cram into limited space in a product, and a budget to boot, so I'm going to go with something authoritative, like a datasheet, where somebody's commercial a$$ is on the line, not a bunch of received folklore from a gaggle of flashlight freaks with more money than good taste.

dtproff, you recommend the same for SLA and NiMH, but nobody else feels it necessary, so I think you are simply overcautious.

Has anybody got anything else?
 
Here is a good explanation of what to do on float charging. Float charging lithium ion cells - 2/1/2006 - Electronics Weekly

To my point... the final paragrapgh is telling:
"Float charging Li-ion cells is possible, but the voltage-temperature-lifetime-capacity trade-off is complex. And Broussely warns: “Certainly discuss this with your battery manufacturer. It is very application dependent.”"

The applications I did they were all wide temperature range (-20C to +50C). I used a thermistor to change the cell voltage for the top-off range. These were intended for maximum life, maximum holdup time applications.

I will follow up with a final comment... there are a lot of ways to do the same thing, some have advantage over other ways of doing things. I usually assume that the person chose a topology for a reason and rarely do I disagree with the approach unless there is a safety issue involved. So, when asked about my opinion that is what I give...

Good luck and please don't take offense to the comments above.

Tony
 
Don't trickle charge Li-Ions. They don't recombine like Pb/NiCd/NiMh and can't "burn off" excess charge.

I'd go with a chip that has both a C/x cutoff and a charge timer. You want the latter for safety reasons - I've personally seen a failed Li-Ion cell keep accepting current and get hot despite the termination voltage not being reached, and you don't want your charger to keep dumping energy into a bad cell in that circumstance. Similarly if you're charging multiple cells in series and one of them fails, a timer will limit the overcharge to the other cells.

For smaller chargers, I've used the Linear LT1731 in a number of different designs and I'm happy with it. It has both aforementioned features, reasonably priced, and only needs a few R's/C's and an external pass transistor to run.
 
Don't trickle charge Li-Ions. They don't recombine like Pb/NiCd/NiMh and can't "burn off" excess charge.

I'd go with a chip that has both a C/x cutoff and a charge timer. You want the latter for safety reasons - I've personally seen a failed Li-Ion cell keep accepting current and get hot despite the termination voltage not being reached, and you don't want your charger to keep dumping energy into a bad cell in that circumstance. Similarly if you're charging multiple cells in series and one of them fails, a timer will limit the overcharge to the other cells.

For smaller chargers, I've used the Linear LT1731 in a number of different designs and I'm happy with it. It has both aforementioned features, reasonably priced, and only needs a few R's/C's and an external pass transistor to run.

Normally you use a timer for each section of the CC/CV/Top=Off in order to figure out if you have a damaged Cell. Li/NiMH are Harder than SLA for the reason they stay flat on the discharge cycle so long. If you hit the end of the timer for any of those sections then you need to terminat the charge cycle. I have yet to find an off the shelf part that does everything I need it to do. The state machine is fairly easy but the compensation for temperature and charge acceptance can be problematic.

Tony
 
Normally you use a timer for each section of the CC/CV/Top=Off in order to figure out if you have a damaged Cell. Li/NiMH are Harder than SLA for the reason they stay flat on the discharge cycle so long. If you hit the end of the timer for any of those sections then you need to terminat the charge cycle. I have yet to find an off the shelf part that does everything I need it to do. The state machine is fairly easy but the compensation for temperature and charge acceptance can be problematic.

Tony
An off-the-shelf part that does all those things would probably cost too much to make to be marketable, and in today's world of throwaway electronics and heavily marked up replacement parts, probably wouldn't have many buyers anyway.

You can get adequate temperature compensation of battery voltage by using a charge controller with a voltage divider for the battery, and sneaking a NTC thermistor into the divider. I did this for a solar powered sensor application using a LT3652 charge controller and a single LiFePO4 cell, worked wonderfully.

To do a per-state timer, I'd probably manage an off-the-shelf charger using a microcontroller.
 
An off-the-shelf part that does all those things would probably cost too much to make to be marketable, and in today's world of throwaway electronics and heavily marked up replacement parts, probably wouldn't have many buyers anyway.

You can get adequate temperature compensation of battery voltage by using a charge controller with a voltage divider for the battery, and sneaking a NTC thermistor into the divider. I did this for a solar powered sensor application using a LT3652 charge controller and a single LiFePO4 cell, worked wonderfully.

To do a per-state timer, I'd probably manage an off-the-shelf charger using a microcontroller.

Micro is how I roll it each time. Fortunately I have only had to do the design once for each of the chemistries over about a 15 year period. Always good to know that it hasn't changed since the last time I touched them about 6 years ago.

Nice talking to everyone on this.

Tony
 
OK, thanks, you guys, for your input, particularly to dtproff for the links.

I hope I didn't cause any offence by my somewhat robust rejection of some of the material advanced, but I really needed to cut through to the nitti-gritti here, a load of handwaving wasn't going to cut it.

To summarise for anyone not wishing to work through all the articles and datasheets, it would seem that there is no real impediment to trickle charging, where the main issue is close control of the termination voltage, with the possibility of introducing a greater safety margin and extending cell life by choosing to charge to e.g. 4.1V or 4.0V per cell instead of 4.2V. It is not an absolute requirement for safety that the charging voltage should terminate at some point, but some protection against failed voltage or current regulation should be present.

'Protected' cells may provide protection against overcharging.

I have decided to stick with the LM317/337 circuit I have already designed for a number of reasons.

1. I am already using protected cells.

2. I don't want a terminated charge, because I want the device to be able to draw power from the onboard mains PSU or from the batteries without any switching complications, allowing the batteries to provide reserve current capacity in the event of demand peaks, the average power of music being << than the peak.

3. I don't want the complication of a switcher and I particularly don't want its accompanying noise. The LM317/337s are intrinsically relatively quiet devices, considered acceptable in many audio applications.

4. It's a bipolar supply and many chargers are intended for single sided operation, which would probably complicate the circuit design.

5. The uproc in the system sleeps for the very large majority of the time. Running it full-time to monitor the battery might mean additional measures were required to suppress digital noise and would require at least one track to cross the D/A groundplane split.

Thanks again.
 
I have other features I am trying to cram into limited space in a product, and a budget to boot, so I'm going to go with something authoritative, like a datasheet, where somebody's commercial a$$ is on the line, not a bunch of received folklore from a gaggle of flashlight freaks with more money than good taste...

I hope I didn't cause any offence by my somewhat robust rejection of some of the material advanced, but I really needed to cut through to the nitti-gritti here, a load of handwaving wasn't going to cut it.

...it would seem that there is no real impediment to trickle charging, .
Good luck. I have no dog in this game. But after everyone has tried to help you, and posted links out the wazoo, you come to this conclusion...

Umm, and you are being *paid* to design this, and have the balls to complain about free advice from folks in a DIY hobby audio group trying to help you?
 
I am not going to second guess his choices. I have said my piece. There are people here who design power supplies for a living and have had really large quantities of chargers ship from full scale production lines.

That being said, we gave the advice and all engineers have trade offs they have to make in a design to get something done. I wish him luck but I have to admit, I would not want to buy the widget he is designing because I think there will be field failures.

No offense meant, just my opinion.

Tony
 
From the Electronics Weekly article:-

'Now Li-ion cells have been in common use for several years and researchers have had time to perform a wide variety of life tests, it has been realised that, with care, simple float charging is also possible. This opens the possibility of using Li-ion cells for stand-by applications with 100 per cent of capacity available at any time.'

This is what I have chosen to do, taking the safety precautions recommended, in order to obtain a performance/weight/volume ratio unobtainable otherwise, at an unmatchable cost both in design effort and BOM.

That's my job. Nobody can say I haven't done my due diligence.

'For life goes not backward nor tarries with yesterday.' - Gibran

'I got a baaad feeling about this one' - Hudson, in Aliens
 
Lithium -formula batteries are great power storage and delivery devices ... but anytime you're playing with Lithium metal care must be taken. There are a number of low-charge and overcharge conditions that can lead to abnormal situations; any gassing at all and the risk of exposure to air or moisture creates potential disaster. Lithium metal oxidizes rapidly with such exposure, burning at a temperature in the many thousands of degrees F. Lithium metal in exposure to air will burn through a glass test tube in very short time.

They were banned on aircraft after some fires brought down (safely) cargo aircraft. There are now rules that allow carrying with strict limits. Cargo shipment of lithium batteries carries special rules on packing and package size.

"Smart" lithium-ion battery packs such as those used on laptops have circuitry built into the battery pack to kill the cell permanently (breaks the connection to the terminals) should they become undercharged to a point where they are no longer chemically stable should electricity be asked of them or given to them.

With single "dumb" cells this kind of safety is found in your charger alone. There is real danger in a casual attitude to Lion pack charging circuits. All safety precautions should be given appropriate consideration.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.