9

I'm building a charger for a set of LiPo cells. My circuit needs to make sure they get fully charged to their maximum level of 4.35V. At the same time, the circuit should protect the cells from overcharge. The charging stops when the 4.35V level is reached.

Tolerances on components might cause a tiny overcharge, for example 10mV. Can this be harmful for the battery?

From what overvoltage does the battery get damaged?

K.Mulier
  • 2,327
  • 3
  • 21
  • 40
  • 10
    Overcharging a Li+ or LiPo can lead to spectacular pyrotechnics with the bits of (what was) a battery slung around with significant energy. – Peter Smith Jul 16 '17 at 14:20
  • 4
    Exactly. That's why I would like to know how many millivolts tolerance I can accept on my measurement circuit (the one protecting my battery). – K.Mulier Jul 16 '17 at 14:23
  • Using 1% components (a tolerance stackup of 2% for the monitoring path) and a feedback device with sub-millivolt offset yields about 6mV without special measures; the measurement should really be set a few mV low for safety. – Peter Smith Jul 16 '17 at 14:31
  • 5
    Okay, but suppose that for whatever reason, I'm actually applying a voltage of 4.351 V instead of 4.350 V. Will it explode? Probably not. But what if I apply 4.352 V. Will it explode? Normally not. But what if ... You get the point? Up to what level can the cell safely go? – K.Mulier Jul 16 '17 at 14:36
  • 1
    Related: [Why is there so much fear surrounding LiPo batteries?](https://electronics.stackexchange.com/q/230155/5830) – user Jul 16 '17 at 21:28
  • 1
    +1. I think it is a healthy engineering approach to know what the allowable margins are for all components. Most answers/comments just spread a fear-mongering without any supporting data. – Ale..chenski Jul 17 '17 at 01:13
  • If it's really specced as 4.35V, anything less than 50mV is rounding error. If it's critical to the 10mV level it should be specified (including in your question) as 4.35**0**V. – Chris H Jul 17 '17 at 08:53
  • 1
    My research indicates that reducing the CV charge voltage even to 3.9V which reduces capacity but significantly increases total life equivalent to about 2500 cycles of rated Ah capacity. ALso choosing 50% DoD also extend lifetime Ah. But if you don't care about lifetime, you can get lower ESR and higher A at a higher temp albeit reduced Ah. Cadex has data on this. – Tony Stewart EE75 Jul 20 '17 at 05:42

5 Answers5

14

The charge (termination) voltage level of Li-ion batteries is specified by manufacturer. The specification is based on reasonably accepted number of cycles ("battery service life") a battery can withstand, say, 500, or 1000. This parameter depends on particular cell chemistry, internal construction, charging current, and is picked up by manufacturer for the best marketable value.

Higher charging voltage leads to slight increase in battery capacity, but it shortens the battery life. The manufacturer's recommended voltage is a trade-off between these two parameters.

Contrary to urban myths of "damaged" batteries, the dependence of "battery service life" on charge voltage is a smooth continuous curve. Certainly the lifetime dependence ends at some point with catastrophic failure, but fears of 10 mV overcharging are grossly overstated. However, 100 mV over 4.35 V (for Li-Po battery) might cause a problem, see, for instance, this publication from Texas Instruments, page 3-5. enter image description here

So, overcharging of 150 mV over the nominal 4.2 V leads to about 10% more capacity for first 50-100 cycles, but the service life shrinks from 500-1000 cycles to about 200. Extrapolating, the another 100 mV will result in maybe 30-50 cycles life. This means that 50 mV over the spec won't kill the battery.

The page 3-7 is also fairly informative. It says that 70-80% of capacity is coming during CC stage, while the tail (CV stage) makes up only 20-30% of capacity, so there is no much reason to wait down to 0.03C. Most TI chargers defaults to 256 mA to terminate the charge process.

For more insights and correct application of chargers, one might want to examine other materials as THIS ONE.

Ale..chenski
  • 38,845
  • 3
  • 38
  • 103
12

Charging Li-Poly and Li-Ion is not as simple as just pumping a voltage across it. It's a two-stage operation involving both constant current and constant voltage.

  1. Apply a constant current of (say) 1C (so for a 1200mAh battery that's 1.2A) until the voltage rises to 4.35V
  2. Switch to constant voltage applying 4.35V until the current flattens out (usually around 0.03C).

Here's the charge curve from a 1200mAh battery I charged the other day using my bench power supply:

enter image description here

As you can see most of the time it's waiting for the current to drop. The longer you wait with constant voltage the greater the capacity. I waited until exactly 0.03C.

Increasing the voltage over and above the charge voltage isn't going to force any more into the cell. Waiting longer for the current curve to become flatter is what you are interested in.

Operating any device outside of the published specifications can lead to undefined results. When operating something as potentially explosive as a battery outside of the published parameters can lead to explosive results.

If it were safe to charge with a voltage above the specifications the specifications would show that. Usually in the form of an absolute maximum specification.

Majenko
  • 55,955
  • 9
  • 105
  • 187
  • 2
    Thank you, this is very interesting. But how much millivolts can the "constant voltage" be higher than the specs? The specification for my battery is 4.35 V, for your battery it's 4.20V. – K.Mulier Jul 16 '17 at 14:21
  • 3
    You shouldn't exceed the rated charge voltage. It's impossible to say how much over you can go before it explodes - it could be 1pV or 250mV. Anything over what it's rated for is potentially dangerous. – Majenko Jul 16 '17 at 14:23
  • 3
    Yes good point : follow the instructions / specifications - that's why they were provided. – Solar Mike Jul 16 '17 at 14:39
  • 2
    Note: your datasheet for the batteries should give you the exact specs for what the batteries are built for. – Cort Ammon Jul 16 '17 at 20:07
  • "Operating any device outside of the published specifications can lead to undefined results." Well, nature has certainly defined the results. They are just unknown by you. – Matt Jul 16 '17 at 23:22
  • 1
    Look, ma! A troll! Can I feed it? Can I? – Majenko Jul 16 '17 at 23:25
  • 1
    Actually, the process has four stages: pre-charge (when the cell voltage is below certain threshold), CC charge (70-80% of capacity), CV stage (20-30% of capacity), and trickle charging. – Ale..chenski Jul 17 '17 at 00:46
  • 1
    You should never let the cell get below the threshold so it needs a conditioning charge. Trickle charging isn't a charge stage, it is merely something you can optionally do to keep it charged after charging is done. – Majenko Jul 17 '17 at 00:48
  • 1
    The discharge threshold is also a pretty much arbitrary parameter too. There is no hard "the threshold". – Ale..chenski Jul 17 '17 at 00:51
  • A decent battery contains protection circuitry to prevent over discharge, so the battery manufacturers obviously believe there is a threshold. Personally I take their opinion as gospel. – Majenko Jul 17 '17 at 01:08
  • 1
    @CortAmmon, datasheets for batteries usually contain only "recommended", single value for charging voltage. The "exact" specs do not help in defining engineering tolerances for charging equipment. I suspect these manufacturers have no clue themselves, the actual data are hidden deeply in R&D departments from where these manufacturers stole the cell design. – Ale..chenski Jul 17 '17 at 01:25
2

At 4.35 V cutoff, the battery is already being damaged for an ordinary LiCoO2 cell. You need to cutoff at 4.15 V or 4.2 V maximum.

The amount of energy capacity between 4.1 and 4.2 V is only 1.2% of the total energy capacity. Between 4.2 V and 4.3 V it is even less, probably 0.6% of total capacity. There really is no point to go above 4.2 V, given that you only increase the energy content by 0.6% and it will quickly and permanently reduce the energy capacity of the cell, even if you do not cycle your cell often.

Tranceducer
  • 47
  • 1
  • 9
  • As can be seen from the TI chart in the Answer above, the storage capacity gained by an extra 0.1 V after 150 to 200 cycles becomes negligible. It is best to keep the limit voltage to 4.2 (relaxed I believe) if you plan to use the cells for more than 150 cycles. This is also true for calendar life in addition to cycle life. – Tranceducer Jul 17 '17 at 01:17
  • 2
    there are different uses of batteries. Say, in RC competition (racing cars, airplanes, or whatever), with 100-200 A discharge currents, the battery might be used only few times, and the charge capacity maxed out might define whether you win or lose. – Ale..chenski Jul 17 '17 at 01:31
  • Nowadays, some LiPo batteries claim to operate at 4.35 V, using special technology. Does it actually mean their lifespan is severely shortened? – K.Mulier Jul 17 '17 at 07:36
  • It's called LiHV, see this comparison and test: https://oscarliang.com/lihv-lipo-drone-battery-hvli/ – Mark Jul 17 '17 at 07:44
2

schematic

simulate this circuit – Schematic created using CircuitLab

This is a simple constant current constant voltage design using 2 LM317. I have to build one myself because it's very hard to buy a 4.35V charger in the market. Circuit is over simplified, do remember to add the caps and protection diode.

Set the U2 output to 4.35V during no load using R3 pot before charging. At full charge it will not exceed 4.35V. Also do check the manufacturer spec on the battery, mine mention 4.35V +- 0.03V charging voltage.

Jason Han
  • 1,720
  • 9
  • 22
0

Constant current switching to constant voltage and then shutting off at 4.2, as chargers work, isn’t necessary. All that matters is u don’t put in too many amps st a time and don’t go over or under voltage. Even the amps u can safely put in can greatly vary depending on soc. If it’s getting warm charging back off and will last longer. Ironically blasting w very high current can actually heal cells increasing energy density and destroying dangerous dendrites. But that’s going a step further.

hummina
  • 9
  • 3