Predicting the time for replacing Li-Ion battery

[BjMag]

I have a bit unusual requirement regarding predicting appropriate time for replacing Li-Ion batteries.

In my application I will use two Li-Ion 18650 batteries (certified capacity 3300mAh) in series that will supply power to a device that draws almost constant 80mA (can vary between 70-90mA). The Device is connected 18 hours a day (or less) in room temperature (during this time the batteries are not connected to any charging or monitoring circuitry). Thus, it is expected that about half the battery capacity has been used. Then the battery is being automatically connected to a charging circuit every day during the remaining time slot of 6 hours. The max charging current available is approx. 500mA which means that fresh batteries should be fully charged in about 3.5 to 4 hours.

The charger circuit will apply the normal Constant Current/Constant Voltage scheme. The circuit (yet to be designed) will be able to measure the voltage before and after the charge (i.e. without the Device connected). I will also be able to monitor the battery cell surface temperature during the charge (as well as during use). The measured values can be received and stored through a Raspberry Pi4B I/O. In addition I could measure the voltage during load a second before I disconnect the Device and I can also measure the voltage a second after connecting the Device (if that is of any help).

I have made the assumption that using these kind of batteries in the specified application using the specified charging method they will last for a pretty long time (maybe 2 years or so). I’m looking for a way to alert the user that it is time to replace the batteries as they are not likely to be fully charged in the 6-hour time slot thus not being able to supply the device with current up to the specified 18 hours. I plan to notify the user with a flashing LED on the Device.

My question: is there a way to reasonably determine this using the measured values over time?

 

[3DDave]

Typically the charging circuit needs to be attached to the LI-ION battery full time as it also functions as the low-voltage cutout to keep from killing the cells via undercharge condition. Note that typically undercharge isn't the direct problem but if the cells are in series then one cell will get far out of sync with the other from under charge. Likewise the circuit cuts power in case of a short circuit.

There are battery monitoring ICs that do these tasks. I'd look at using those or at least getting the data sheets and duplicating the functionality.

For example:

https://www.ti.com/power-management/battery-management/overview.html

https://www.st.com/en/power-management/battery-management-ics.html

https://www.maximintegrated.com/en/products/power/battery-management.html

https://www.analog.com/en/products/power-management/battery-management.html

https://industrial.panasonic.com/ww/products/pt/battery-monitor-ic

See also:

https://www.cell.com/iscience/pdf/S2589-0042(21)00028-6.pdf

 

[BjMag]

Hi and thanks for your reply. I'm full aware of what you are saying. But in my application the charging circuit cannot be connected to the batteries during the time they are being used. This is an absolute requirement. So existing BMS ICs or circuits are of little help. I'm looking for another way to the reasonably determine the health of the batteries. As stated the batteries will be charged daily and the current draw is almost constant and also very low. If it is of any help I can also automatically separate the cells and take any measurements while charging them individually he keep them reasonably balanced.

I'm struggling with this issue and will have to skip my device development if this problem can not be solved. I have the this inquiry published in many places around the world and really hope to find an acceptable solution.

 

[itsmoked]

As 3DDave states you need a BMS.

The BMS must remain connected to the batteries at all times. Doesn't matter if you don't want the batteries hooked to the charger, that's okay, you still need the BMS connected to the batteries. It absolutely needs to disconnect the batteries before EITHER of them get below the the specified cut-off voltage. If they ever get below the cutoff voltage then they must be deemed "compromised" and should never be charged again - enforced by the BMS. That is, unless you're fine with being sued into oblivion. That's because it's now the Lithium-Ion standard realized because over-discharged batteries are more likely to fail catastrophically and if you choose to not do it that way then you choose much more liability.

The BMS also helps greatly with your stated mission because it will easily track battery health. It can also, if you choose, keep the batteries balanced with each other which probably doubles the average service life of a LiIon battery pack. It also disconnects the pack before either battery is over-discharged OR over charged which is, of course, another even more serious hazard.

The BMS will provide fuel-gauging which dumps out battery capacity. Of course battery capacity is what you're after with respect to "when to change out batteries". From day one the battery capacity will slowly diminish. You can provide notification when it drops to some percentage that you know will still carry the user for long enough to get the batteries replaced.

I'd suggest you work into your product an exchange program where you replace the batteries and reset the BMS and shoot the pack back to user. That allows you to keep your product reliable which will not happen if the users fish around and buy crap batteries to replace the tired ones. Your product will get a major black-eye because its reputation will be dragged in the mud for any reason it doesn't work - even if it's the customer's fault.

In this modern scene you can also care for the earth by correctly recycling the returned batteries whereas lots of them would otherwise end up in the landfill a terrible waste of a resource.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[BjMag]

OK. I understand! I will have no other choice than canceling my project. Thanks for your comments!

 

[IRstuff]

18650's are like the LI equivalent of AA batteries; as such, they're used in lots of applications where there are no BMS' and they're removed from, my laser pointer, as an example, and placed in a stand-alone charger, and the charger would basically refuse to charge such a battery if it were below some preset cutoff voltage. It would seem plausible that your charger would simply refuse to charger the battery, if the discharged voltage were below, say, 3V. Note that 2 years seems to be a bit optimistic, even in the best of circumstances, per link below.

https://www.instructables.com/How-to-Prolong-the-Life-of-an-18650-Battery/

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert!

https://www.youtube.com/watch?v=BKorP55Aqvg

faq731-376 forum1529 Entire Forum list

http://www.eng-tips.com/forumlist.cfm

 

[3DDave]

The removable ones are supposed to have a BMS chip; naked li-ion cells are dangerous.

What heckin laser pointer uses an 18650? At least in the USA any laser with greater than a 5mW output is illegal to be sold as a laser pointer. They are legal to buy and own but the FDA regulates them and doesn't permit anything above 5mW to be described as a laser pointer for sale.

 

[IRstuff]

On the advice of counsel, I invoke my 5th amendment rights. ;-)

Spoiler

They're are green lasers, which are pretty cool in that you can see their beams diffracting off air molecules and reflecting off dust particles in the air.

Actually, the "tactical" flashlights also use 18650 cells. In any case, the OP stipulated 18650s, so it's ostensibly no different than running down my flashlight. plucking the battery out and plugging it into a charger.

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert!

https://www.youtube.com/watch?v=BKorP55Aqvg

faq731-376 forum1529 Entire Forum list

http://www.eng-tips.com/forumlist.cfm

 

[3DDave]

I worry about metal cased devices with 18650s in them. In the event of thermal runaway they are a pipe bomb. I know it is rare, but the way they are typically built without a pressure vent is bothersome. Like I said, there are naked 18650s for OEM incorporation and protected 18650s that are for flashlights and so forth.

I think some of the vaping community have lost fingers from unprotected 18650s looking for more current draw than the protected cells can manage. I see this report - someone was killed by 18650s in a vaping device:

https://www.nbcnews.com/health/heal...gerous-deadly-e-cigarette-explosions-n1032901

A reddit thread suggested they are pulling over 100W from a pair - 7V, so nearly 15 Amps?

Perhaps you know this, but I've seen references indicating that the some green diode lasers are even more hazardous when they are cold, like freezer temps. They use a frequency doubler as a following element to a near IR diode and when they are cold the crystal doesn't function, allowing a lot of near IR through, the kind of invisible light which at sufficient power can damage retinal cells.

https://www.eevblog.com/forum/dodgy-technology/video-dont-use-a-green-laser-in-the-cold!/

 

[IRstuff]

They use a frequency doubler as a following element to a near IR diode and when they are cold the crystal doesn't function, allowing a lot of near IR through, the kind of invisible light which at sufficient power can damage retinal cells.

A correctly designed transmitter should have a wavelength filter to block out the pump wavelengths you refer to. Nevertheless, that infrared wavelength is slightly less hazardous than the green output, since the green is more dangerous than even a red laser. Nevertheless, the pump output power is substantially higher, although most of it is supposed to be absorbed by the pump medium.

It's pretty cool, I think, that green lasers actually contain similar devices and operations as the military laser designators for "smart" munitions like Hellfire. "smart" because they're actually pretty dumb

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert!

https://www.youtube.com/watch?v=BKorP55Aqvg

faq731-376 forum1529 Entire Forum list

http://www.eng-tips.com/forumlist.cfm

 

[3DDave]

The green is bright enough to warn the user. Dropping into IR is what gets people to try looking directly into the laser to see if there is something wrong. When it's not converted into green, it's not getting completely absorbed.

"Correctly designed" is exactly what most laser pointers are not. An IR blocking filter costs money.

 

[IRstuff]

Without the filter, a green laser will be MUCH more dangerous, since the pump light isn't collimated, so there'll quite a bit that's coming out off-axis from the main green laser beam

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert!

https://www.youtube.com/watch?v=BKorP55Aqvg

faq731-376 forum1529 Entire Forum list

http://www.eng-tips.com/forumlist.cfm

 

[3DDave]

I would expect that the crystal is large enough to intercept the entire output of the IR diode so no collimation is required.

Still - when it's bright green people don't look at it, which is the same reason few stare into the Sun, but get retinal damage using neutral filters that let IR through. It's not a matter of being more dangerous - it's when it isn't obviously dangerous that causes problems.

 

[IRstuff]

I was referring to the output of the medium; while the lasing energy is self-collimated, by virtue of being bounced back and forth between the two ends, and output directly into the frequency converter, the pump energy doesn't do that, and can spray out of the lasing medium pretty much in any direction, and can sneak around the frequency converter and get to the output port in a wide angle beam, although there are physical limitations due to the depth of the exit aperture itself.

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert!

https://www.youtube.com/watch?v=BKorP55Aqvg

faq731-376 forum1529 Entire Forum list

http://www.eng-tips.com/forumlist.cfm