18650->12V battery for UPS: solving the repetitive disconnect

Thread Starter

hroldan

Joined Oct 12, 2016
18
Hi everyone,

I have 3 UPS units, 2 of them with 12v batteries built using 18650 batteries + BMS working perfectly, I use these units for my computer equipment and have saved me from blackouts and voltage spikes, etc.

However, I have a more recent model with fancy functions, like, battery level, frequency indicator, voltage and LOAD, and this unit disconnects continuously. I know what this is, it's because the 12v battery built using a BMS and 18650 cells keeps receiving energy (charging) while also allowing it's voltage to be detected, BUT... when the full charge is reached, the BMS automatically disconnects the battery to stop receiving charge. This causes the UPS to detect no battery, triggers an audible alarm, and besides being annoying... if a blackout occurs during this short moment, the UPS would not perform it's duty, just because it senses the battery is gone.​

I know this short "disconnect" takes places on all sorts of BMS for 18650, and it happens really fast, but this UPS is too clever for this. I only have issues with this UPS model. Sure, it works perfectly with lead acid batteries and I don't want to get rid of this unit.

  1. Any ideas or suggestions?
  2. What do you think about this: placing a decent size capacitor (or 2) in parallel, this way, the BMS will disconnect as usual but the capacitor will be enough to keep reporting voltage back while the BMS comes back. I must say 99.99% of times I have an idea like this I try it before asking questions, but this time is different considering the situation, I want to be careful and all ears before trying, because I'm pretty sure this setup might involve things I have absolutely no idea I should be asking first.

Thanks in advance.
 

Art Mezins

Joined May 26, 2019
11
Something I keep saying on these "questions" -- how about a schematic, even a rough one? List parts used (e.g. manufacturer & model #). Are the 18650's welded together with nickel bands? Do the 18650's have "internal protection" that's being triggered? I know that raw gel cell Lead-acid batteries don't have any "protection" other than their intrinsic properties. No scope pictures of what's going on, just "simple descriptions" doesn't cut it for engineering anything. You're not asking a local garage mechanic for help, we're (mostly) engineers and we like DETAILS! And don't be afraid to screw up, we all do and that's how we learn. So where's the schematic or diagram, model numbers, etc?
 

nsaspook

Joined Aug 27, 2009
13,420
Your're right, a dumb UPS is better here.

It's common problem with charging systems primary designed for fancy multi-stage LA batteries charging and monitoring. My LiFePO4 battery bank BMS would cause a short inverter warn alarm (the solar charge controller adjusts it back to normal range quickly) for battery over-volt during a fully charged disconnect in full sun. The FM80 charge controller can be adjusted to prevent that BMS alarm.

Usually with the 'smart' chargers with signal connects (USB/Serial) to a PC there are commands that can adjust charging parameters. So, it might be possible to dumb the UPS down a bit if it uses the right communications protocol.
http://www.apcupsd.com/manual/#apc-smart-protocol
http://www.apcupsd.com/manual/#resetting-the-ups-battery-constant
 

eetech00

Joined Jun 8, 2013
4,004
Hi everyone,

I have 3 UPS units, 2 of them with 12v batteries built using 18650 batteries + BMS working perfectly, I use these units for my computer equipment and have saved me from blackouts and voltage spikes, etc.

However, I have a more recent model with fancy functions, like, battery level, frequency indicator, voltage and LOAD, and this unit disconnects continuously. I know what this is, it's because the 12v battery built using a BMS and 18650 cells keeps receiving energy (charging) while also allowing it's voltage to be detected, BUT... when the full charge is reached, the BMS automatically disconnects the battery to stop receiving charge. This causes the UPS to detect no battery, triggers an audible alarm, and besides being annoying... if a blackout occurs during this short moment, the UPS would not perform it's duty, just because it senses the battery is gone.​

I know this short "disconnect" takes places on all sorts of BMS for 18650, and it happens really fast, but this UPS is too clever for this. I only have issues with this UPS model. Sure, it works perfectly with lead acid batteries and I don't want to get rid of this unit.

  1. Any ideas or suggestions?
  2. What do you think about this: placing a decent size capacitor (or 2) in parallel, this way, the BMS will disconnect as usual but the capacitor will be enough to keep reporting voltage back while the BMS comes back. I must say 99.99% of times I have an idea like this I try it before asking questions, but this time is different considering the situation, I want to be careful and all ears before trying, because I'm pretty sure this setup might involve things I have absolutely no idea I should be asking first.

Thanks in advance.
If its not maintaining AC output when utility service is lost then its faulty.
UPS stands for "Uninterruptible Power Supply". So dump the UPS.
 

nsaspook

Joined Aug 27, 2009
13,420
If its not maintaining AC output when utility service is lost then its faulty.
UPS stands for "Uninterruptible Power Supply". So dump the UPS.
It thinks the battery is dead (a fault condition) because the charge current goes to almost zero when the BMS does a fully charged disconnect (the battery monitor expects to see a small float maintain current after a full recharge cycle). The UPS is fine, it's the BMS protecting (disconnecting) the Li type battery that's fooling it.
1697846947804.png
The orange line slope (240 -> 512) is the calculated amount of energy in the battery and the green line is the solar energy going into the battery to recharge it. The LiFePO4 battery internal BMS cuts off the charge just a little bit after my tracking calculations say it's full.
 
Last edited:

eetech00

Joined Jun 8, 2013
4,004
It thinks the battery is dead (a fault condition) because the charge current goes to almost zero when the BMS does a fully charged disconnect (the battery monitor expects to see a small float maintain current after a full recharge cycle). The UPS is fine, it's the BMS protecting (disconnecting) the Li type battery that's fooling it.
Most worthy UPS's have a "runtime test" that determines the amount of float charge required for the battery while it is under load, so the battery really is never actually disconnected and is operating the load. This test also prevents overcharging the battery.
 

nsaspook

Joined Aug 27, 2009
13,420
Most worthy UPS's have a "runtime test" that determines the amount of float charge required for the battery while it is under load, so the battery really is never actually disconnected and is operating the load. This test also prevents overcharging the battery.
I've written some of those runtime test routines, They are designed and calibrated for DUMB Lead-Acid batteries that don't disconnect internally from being charged (they will still output power usually). A typical Lithium battery BMS will totally disconnect the charging circuit when it thinks the Lithium battery is full to protect from over-charge and a possible fire. When the DUMB Lead-Acid routine sees the no-charging current profile in any part of the charge cycle, the most common thing is to flag that as an error.

The double conversion, always running on battery, utility power only keeps battery charged type UPS are better at handling the BMS disconnect issue.
 
Last edited:

bassbindevil

Joined Jan 23, 2014
836
You should really have a charger that is designed for lithium-ion batteries. The "BMS" module does not manage charging, it just protects from catastrophic overcharging. LiFePO4 chemistry is a better match to a lead-acid charger.
 
Top