6 year old unused BYD packs 48V 50Ah, what to do?

PeterB

New member
Joined
Jan 4, 2024
Messages
7
Hello to all,

I recently got hold of 4 (new) BYD U3A1-50 battery packs for free.

These are 48Volt 50Ah (2,56kWh) packs with their build-in BMS in a 19" type of enclosure. (see pics).

They were produced in 2017 and have never been used, sitting idle in storage. The owner frequently checked the SOC status, by switching them on and measuring the battery voltage.

Now, January 2024, after 6 years, 3 out of 4 packs were still 51,5V appr., but recently one pack would not switch on anymore and gave an error about 'battery failure'.

I opened them and measured that 14 cells were around 3 Volt, 1 cell was 0.4V and 1 cell was 1.8V.

Reviving 'dead' Li-cells is always a tricky business I understood, because re-crystalisation might prevent the cell ever returning to a good state, and even turning them dangerous because they won't accept charge, but might blow up.

So, following advise, I started charging the pack with a modest 100mA.

The voltages of the lowest 2 cells rose pretty quickly.

I'm currently charging at 0,25 Amp. and after a couple of hours the lowest cell rose to 2.62V the other to 2.75V. (they're still charging now...)

Now I expect that these two cells will never be at the same state as their brothers, thereby making the whole pack bad/weak.

Since I'm running my house on a 24V off-grid solar system I'm a bit in doubt as of what to do.

I've been thinking about converting the packs to 24 Volts. As you can see in the picture the pack is made up of 2x8x2 cells. The numbering of the battery voltage sense wires makes you think the system is designed for either 24 or 48 Volt, as the sense wires are numbered 0 to 7 per set of 8 and numbering is not sequential from 0 to 15. Every set of 8 has also 2 temperature sensors.

Of course the BMS will be upset if you leave out halve of the cells o_O. So, first question is:

Question 1: does anyone know if these packs/BMS can be changed to 24 instead of 48V?

For the faulty pack I'm even thinking of removing the two weak cells and converting the pack into 8x3 giving 24V 75Ah. That would involve a total re-build of the pack, breaking old and making new bridges between cells.

Question 2: Does anyone have any experience with how best to proceed in these type of cases?

Well, that's a pretty long story, and probably not all details are covered, so ask me if you want/need to know more.

Thanks!


Peter BYD-front.jpgBYD-top.jpgBYD-to-front.jpg
 
Last edited:
Recap:
Low voltage cells (after sitting) can be compromised, but it's not as common (in my experience) as you seem to be thinking. Especially 1.8v is not that bad. 0.4v is worse but not nearly as bad as 0.0v. 1.9v and 0.4v cells are worth testing - I've had very hi success rate.

Detail:
My powerwall is made up of 12,000+ individual 18650 cells over several years. I've had many (500?) cells that that were very low voltage (1 cell was 0.4V and 1 cell was 1.8V) and recovered them by charging them and running them thru the regular test process (charge, discharge to find if capacity is reasonable, charge, let-sit for a few weeks to find self-dischargers, verify IR is reasonable). And.... there is no indication these cells are weaker than any other cells after many cycles - some are over 1,000 cycles.

So, I might (gently) suggest that you're overthinking the low voltage = week cell. If they recharge and don't self-discharge and don't get hot, then odds are they will serve well However, I'm always in favor of doing actual tests and collecting data....

- If you can, test each low voltage cell (or parallel group of cells) independently as described above to verify capacity, self-discharge, IR. If in reasonable range (90%+ original capacity, good IR, not hot, no self-discharge) then in my experience they will perform very well.

- If you can't do individual cell testing, then test the whole battery via a few cycles of reasonable discharge (you're planned operational low) and charge to you're planned operational hi and see if the cells stay in balance. If they act reasonably - then you have data to give you confidence and personally I would expect them to continue to act reasonably :)

Note: It's not a bad idea to test ALL you batteries - it's always good to have data on whether they work (reasonable capacity) as expected under charge and discharge thru operational voltage ranges and stay balanced.

As far as altering a BMS from 48v to 24v - you can try but it's hit-miss.
 
Last edited:
Recap:
Low voltage cells (after sitting) can be compromised, but it's not as common (in my experience) as you seem to be thinking. Especially 1.8v is not that bad. 0.4v is worse but not nearly as bad as 0.0v. 1.9v and 0.4v cells are worth testing - I've had very hi success rate.

Detail:
My powerwall is made up of 12,000+ individual 18650 cells over several years. I've had many (500?) cells that that were very low voltage (1 cell was 0.4V and 1 cell was 1.8V) and recovered them by charging them and running them thru the regular test process (charge, discharge to find if capacity is reasonable, charge, let-sit for a few weeks to find self-dischargers, verify IR is reasonable). And.... there is no indication these cells are weaker than any other cells after many cycles - some are over 1,000 cycles.

So, I might (gently) suggest that you're overthinking the low voltage = week cell. If they recharge and don't self-discharge and don't get hot, then odds are they will serve well However, I'm always in favor of doing actual tests and collecting data....

- If you can, test each low voltage cell (or parallel group of cells) independently as described above to verify capacity, self-discharge, IR. If in reasonable range (90%+ original capacity, good IR, not hot, no self-discharge) then in my experience they will perform very well.

- If you can't do individual cell testing, then test the whole battery via a few cycles of reasonable discharge (you're planned operational low) and charge to you're planned operational hi and see if the cells stay in balance. If they act reasonably - then you have data to give you confidence and personally I would expect them to continue to act reasonably :)

Note: It's not a bad idea to test ALL you batteries - it's always good to have data on whether they work (reasonable capacity) as expected under charge and discharge thru operational voltage ranges and stay balanced.

As far as altering a BMS from 48v to 24v - you can try but it's hit-miss.
Hi and thanks for the info & tips.

I indeed presumed that those low cells would be damaged. But indeed as we say in Dutch "to measure is to know", so I will test them all. Good to hear there are reasonable chances they are ok.

I don't have special charging equipment, so I charge them with my adjustable bench power supply which I set at the desired end voltage (3.55V cell voltage is a good safe value for LiFePO4?). What is, in general the end/stop current 0.01 or 0.02C ?

Max charge current is limited to about 4 Amp.
For discharging I probably start with 4 car head lights.

Do these BMS systems give a SOC based on voltage or do they count charge (Coulombs going in/out)?

... still plenty to find out here :)

to be continued.
 
Hi and thanks for the info & tips.

I indeed presumed that those low cells would be damaged. But indeed as we say in Dutch "to measure is to know", so I will test them all. Good to hear there are reasonable chances they are ok.

I don't have special charging equipment, so I charge them with my adjustable bench power supply which I set at the desired end voltage (3.55V cell voltage is a good safe value for LiFePO4?).
3.65v is a typical max charge voltage - but 3.55v is not bad. Keep in mind that LifePo4 cells have a max charge (3.65v is typical) but immediately drop sharply for the first part of the discharge curve to ~3.45. So if you charge to 3.65 or 3.55 is not really that important to test capacity - however - be consistent so you have apples to apples data.

1704485966997.png



What is, in general the end/stop current 0.01 or 0.02C ?
My bench supply has CC and CV (constant voltage). If you're does as well, it will go to 0 as lithium-ion doesn't have constant discharge like lead acid unless it's a bad cell (e.g. self-discharging cell). It's OK to stop if you want but this requires manual intervention and can cause inconsistent results.

Note1: LifePo4 chargers are not that expensive.

Max charge current is limited to about 4 Amp.
Fine - its just a matter of time - e.g. more amps you get faster results.

For discharging I probably start with 4 car head lights.
Head lights are OK but I would ask what's you're plan for these batteries - some kind of inverter I would imagine. Suggest buying an inverter and using that for load tests - maybe a little easier to plug things into an inverter to adjust load so you don't take all day for 1 test and to simulate the loads you're planning to run to ensure the batteries perform as you expect.

Do these BMS systems give a SOC based on voltage or do they count charge (Coulombs going in/out)?
Don't have any info on this.
 
Last edited:
On aliexpress you can find a lot of testers.
Make sure you buy the ones with a cut off v
My number one are the zb's testers.
You can hook up several of them to discharge/capacity test one cell.
But dont exeed the C rate.

For charging it can be done with a 4A bench supply, indeed time would be a isseu.
With charging also keep the C rate in mind.

With lifepo4 chem, after 3.55 not much is gained, in the long run the cells would be happier with a max of 3.50v
Those cells must be in compressed state, if one is bulging, you could try to form it back, but damage will occur, you will lose capacity.

1656550744478.png

LFP Voltage Chart.jpg


A bit from experience: the cells dont like 100% or 0% but 99.5% to 5%, i keep mine between 99% and 5%
When forming a battery all your cells in that battery must be as close as possible regarding IR and most important SOH(remaing capacity)


Like most chemistries lifepo4 below minimum v cells can recover, but as you will find out just like lead acid, li ion or NiMh the SOH will be low.
Most will be a self-discharger or have a high IR.
Some will recover after a few cycles; some are just to tired.
 
On aliexpress you can find a lot of testers.
Make sure you buy the ones with a cut off v
My number one are the zb's testers.
You can hook up several of them to discharge/capacity test one cell.
But dont exeed the C rate.

For charging it can be done with a 4A bench supply, indeed time would be a isseu.
With charging also keep the C rate in mind.

With lifepo4 chem, after 3.55 not much is gained, in the long run the cells would be happier with a max of 3.50v
Those cells must be in compressed state, if one is bulging, you could try to form it back, but damage will occur, you will lose capacity.

View attachment 31241
View attachment 31242

A bit from experience: the cells dont like 100% or 0% but 99.5% to 5%, i keep mine between 99% and 5%
When forming a battery all your cells in that battery must be as close as possible regarding IR and most important SOH(remaing capacity)


Like most chemistries lifepo4 below minimum v cells can recover, but as you will find out just like lead acid, li ion or NiMh the SOH will be low.
Most will be a self-discharger or have a high IR.
Some will recover after a few cycles; some are just to tired.
Thanks for the graphs, nice to have all values/sets next to each other.

I'm currently making a dedicated charger with which I can charge with up to 10 Amp. with adjustable max. volt (charger for 1 cell only!) and cut-off current. This way I can charge all cells separately up to the same SOC before starting a discharge test with a 48V 800W inverter, which should arrive shortly.
 
Back
Top