Author Topic: Is Li-ion 18650 battery protection necessary for low power solar devices?  (Read 1339 times)

0 Members and 1 Guest are viewing this topic.

Offline brych34Topic starter

  • Contributor
  • Posts: 15
  • Country: wales
Hi everyone

I'm currently working on upgrading a solar powered LoRaWAN air quality monitor. It contains an STM32 that runs on a 3000mAh 18650 battery charged by a CN3791 module. I've been trying to find how dangerous it is to overdischarge the battery but most sources seem to focus on loads consuming 200mA or more.

Originally I had a DD04CPMA module (like this one https://www.ebay.com/itm/254261521570) to protect the battery. It cuts off the power when the voltage falls below 2.5v. However, it seems to have broken - it now cuts off the power a split second after it is turned on, no matter the battery voltage.

I was wondering whether I need to order a new one, or whether I could potentially get away without one? The STM32 is set to stop transmitting and go into 'hibernation' with all the sensors off when the battery voltage goes below 3.6v. In this mode it consumes around 2.6mA. If there's around 800mAh left in the battery when it goes into hibernation it should be able to last for around 13 days, which is hopefully long enough for the solar panel to get some good weather and charge the battery up.

Unfortunately the STM32 continues to consume current down to 2v, at which point the 3.3v regulator voltage drops sharply and the current consumption drops to 0.4mA.

The device will be installed on the side of a building near a car park. Do you think it's safe to discharge the battery from 3.6v downwards in hibernation mode with no protection?

Thank you!
Brychan
« Last Edit: July 26, 2021, 04:33:49 pm by brych34 »
 

Offline Nusa

  • Super Contributor
  • ***
  • Posts: 2416
  • Country: us
Depends on what definition of "safe" or "dangerous" you're looking for. Trickle discharge until death isn't going to make the battery go BOOM, if that's what you mean. If you mean safe for the battery, each deep discharge cycle is going to take large chunks out of its lifetime.

On the other hand, a charged 3000 mAh battery charged to 4+ volts with a 2.6 mA load should have no problem lasting for days. Unless it's no longer a 3000 mAh battery (do a capacity test on it). Or possibly never was, depending on the vendor purchased from (do a capacity test when new too!). Or the charging system isn't working, for whatever reason. How long did the system work before failing?
 
The following users thanked this post: brych34

Offline Siwastaja

  • Super Contributor
  • ***
  • Posts: 8090
  • Country: fi
Overdischarge itself is safe, it's the subsequent charging after such overdischarge event which is risky. Overdischarge may (or with luck, may not) increase self-discharge by many orders of magnitude due to copper dendrites, partial internal shorts. These may hotspot during charge.

Non-broken charger ICs do not allow starting the charging of overdischarged cell, so if your charging circuit is up to its task, there is no safety problem, just cell replacement and disposal problem.

Are there any way to reduce the sleep current of your design? Enable pin on the regulator? Any hack you could use to turn off the power?

Putting things into perspective, no, even if you overdischarge to say 1.0V then charge with a modest current, it's very unlikely to burst into flames. But doing this nevertheless can't be recommended, ever.
« Last Edit: July 26, 2021, 05:53:07 pm by Siwastaja »
 
The following users thanked this post: brych34

Offline AmnevaR

  • Contributor
  • Posts: 40
  • Country: ua
If you mean safe for the battery, each deep discharge cycle is going to take large chunks out of its lifetime.

Can confirm. I have a few 18650 from old laptops. Once they have been discharged to 2V or less, there is no coming back.
After charging those cells they just lose their charge overnight.
 
The following users thanked this post: brych34

Offline tunk

  • Frequent Contributor
  • **
  • Posts: 968
  • Country: no
I've seen 18650s discharged to well below 2V turn out fine.
 
The following users thanked this post: brych34

Offline Siwastaja

  • Super Contributor
  • ***
  • Posts: 8090
  • Country: fi
I've seen 18650s discharged to well below 2V turn out fine.

Yes, even with high probability, but this does not come completely without risk thus can't be endorsed for a product, hobby or commercial.

Below maybe 0.5V, chances start to get very slim.
 
The following users thanked this post: brych34

Offline brych34Topic starter

  • Contributor
  • Posts: 15
  • Country: wales
The device will be a bit inaccessible, so it sounds as if it's best to get a protector just in case to avoid having to replace the battery often.

The capacity test shows the battery is just over 3000mAh, so that should be ok. The device worked for around six months before failing, presumably because the protector broke.

Unfortunately I can't think of any more ways to reduce the power. All the LEDs are gone, the sensors are cut off by N-channel MOSFETS and the STM32 and RFM95 are both put into sleep mode. A voltage divider for measuring the battery voltage is using around 0.2mA but I don't think it's worth going to the hassle of replacing them and recalibrating the voltage measurement. I suppose I could add a third MOSFET to cut off the RFM95 but it would be messy because I didn't design it into the PCB.

Do you have any protector modules you recommend? I'm a bit hesitant to get another DD04CPMA seeing as the last one died.

Thanks for the advice!
Brychan
 

Offline varunkoganti

  • Contributor
  • Posts: 11
  • Country: in
Yes,
Li-Ion batteries are extremely sensitive, you need battery management system(BMS), even if you are using 1 cell.
 

Offline mikerj

  • Super Contributor
  • ***
  • Posts: 3231
  • Country: gb
I've seen 18650s discharged to well below 2V turn out fine.

Me too, but I think it depends quite strongly on how low the voltage is and how long they sit in this over-discharged state.  I always put them in a vented metal box and charge them at very low current to recover them.  If they self discharge quickly after then I scrap them.
 

Offline brych34Topic starter

  • Contributor
  • Posts: 15
  • Country: wales
Actually I've been playing around and it seems that if I connect PA15 to DIO1 of the RFM95 instead of PA12 the hibernation current goes down to 0.37mA. If there's 800mAh left in the battery when it goes into hibernation it should now last for around 90 days.

I think that's long enough for the sun to come out and charge the battery back up, so I'm going to try putting it up without a protector. If it fails I can always add a protector later on (with a new battery!).

The only thing I'm not sure about is at what voltage level it should go into hibernation. I was thinking maybe when the battery is at 3.6v but I'm not sure.

Thank you,
Brychan
 

Offline Siwastaja

  • Super Contributor
  • ***
  • Posts: 8090
  • Country: fi
Sounds like a good plan to me. You can go to slightly lower voltage and re-do your 90 days math, of course, if you want more capacity out of the cell. Maybe 3.55V or so, but no less than 3.50V. Also be aware of the tolerance of the circuit which measures the battery voltage.

If you look at the typical discharge curves, cell voltage seemingly drops sharply near empty, but that is under significant load such as C/2; it's actually the ESR which rises sharply! If you have very tiny load, say like C/20 or less, some 3.35V is already empty, equivalent to 2.5V under load! So with low current, better have 3.35V as 0%. 3.65V being 50%, quite small changes make a big difference.
 
The following users thanked this post: brych34

Offline perieanuo

  • Frequent Contributor
  • **
  • Posts: 838
  • Country: fr
depends on your 'device'
at home, you do what you want, it's your responsibility
some specific  (industrial, medical, automotive...) may impose battery protection by design, tested when validating the device in labs
to be clear, no protection means first assembly line tech coming Monday head full of Sunday bear will explode the battery. I saw this in one facility, i explained the hw designer battery protection didn't cost that much for industrial hw, he implemented it. of course i don't say this make 100% sure battery usage in the real life, but if done correctly expands battery life and safety. more intelligent people than me analysed this and this was the final conclusion, every battery with some power inside should be at least have a resettable fuse. 18650 is no little battery, everyone advise to handle with care li batts
generic q, generic a
 

Online David Hess

  • Super Contributor
  • ***
  • Posts: 16510
  • Country: us
  • DavidH
The device will be installed on the side of a building near a car park. Do you think it's safe to discharge the battery from 3.6v downwards in hibernation mode with no protection?

It is safe in the sense of not being dangerous but it will severely damage the battery.
 

Offline BILLPOD

  • Regular Contributor
  • *
  • Posts: 224
  • Country: us
Good Morning Brycha34,  just get an 18650 battery that has protection circuitry built in.   :-/O
 
The following users thanked this post: skepticelectroneophyte


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf