Lithium Cell Balancing at Nowty Towers - Resurrections

[nowty]

.
(I find the squared character as in “4mm²” looks better that a superscript 2 as in “4mm²”)

I didn't know how to do that, but I just copied your text and pasted it and it works.

㎡㎡㎡㎡㎡㎡

[AE-NMidlands]

.
(I find the squared character as in “4mm²” looks better that a superscript 2 as in “4mm²”)

I didn't know how to do that, but I just copied your text and pasted it and it works.

㎡㎡㎡㎡㎡㎡

apparently you are supposed to be able to type the ASCII code for the character in direct, (Alt 253) but I have never been able to get it to work...
https://theasciicode.com.ar/extended-as ... e-253.html

[nowty]

To make it easier to monitor my batteries, I installed a remote voltage monitor in my lounge. I had one for the old 24V system too, but I needed to change the voltmeter to work on the higher voltage.



And anything you connect to your battery bank you need fuses just in case, even with this low power voltmeter.

[nowty]

I continued to use my battery banks without balancers for at least a year more in its new configuration, but I found it became more difficult to manually re-balance such a large battery bank. I then experienced a low SOC discharge event and that caused a couple of cells to deeply discharge, so this forced me to have a rethink.

I bought these active cell balancers from China after I saw a review of them on youtube.



Front Side



Back Side



Connected up with fused links



Specification

• Operating voltage: 3.0V - 4.2V

• Working power drain 15mA

• Termination voltage: 2.9V (into undervoltage sleep state <1mA)

• Maximum Balance current: The Cell difference V / Balance current

0.1V / 1A
0.5V / 3.5A
1.0V / 5.5A

• Balance accuracy: < 5mv

• No external power supply is required, internal energy in each high cell is transferred to the lower adjacent cell in the pack to balance the entire battery pack.

Cascade Mode
So for my Growatt packs of 14 cells, they didn't make one for that size but I could use two 7/8S boards with one of the mid cells overlapped across both boards.

[nowty]

As I only had one balancing board to start off with, I set up a remote test point which I connected to half the cells.

It worked exactly as in the youtube video. I recorded each cell voltage before and after connecting, straight away there was some slight differences showing that the high cells were bleeding current and the low ones were having current added. The currents were too small to measure accurately with a clamp meter as my cell differences are quite small, typically 20mV to 30mV for adjacent cells.

But then I shorted each cell in turn with a 0.5 ohm resistor. This had the effect of dropping each cells voltage by about 0.5V which raised the charging current by about 3.5A which was very measurable with my clamp meter and It also matched the spec.

 
Before shorting the cell, voltage is 3.873V, current says -0.58A (which in reality is close to zero because the clamp meter does not zero very well).



After shorting the cell, voltage is 3.426V (drops by 0.45V), current says +3.02A (so its increased by about 3.6A)


 
I repeated this on every cell and I got a similar result. This resistor test really sold it for me, that's a reasonable amount of current to shift and the balancing board does not even get warm.

I left the balancer on all night, at the start there was a difference of 62mV between the highest and lowest cell. By the morning that max difference had dropped to 13mV at the test point and 20mv directly measured at the battery. The difference is due to the voltage drop in the 6 metre long cables from the test point. Some of these differences were positive and some were negative. This was consistent with the current flow direction with an outflow from the higher voltage cells and an inflow to the lower voltage ones.

Within another couple of days, all cells were within 5mV which is within the spec of the balancing board.

[Stinsy]

Just to add: the point of the balancing board isn’t to get all the batteries in the string exactly the same. The point is to prevent one cell drifting further and further from the others over a long period of time. This could cause undesirable consequences, eg if one cell is lower than the rest and the pack is discharged, the string voltage could be within spec but the one cell could be excessively discharged and become damaged. And the same for an over-voltage cell.

With lead-acid batteries this was a simpler problem to solve because lead-acid batteries are not damaged by moderate overcharging. When you overcharge the string you bring the low-voltage cell back up to match the other cells in the string (and the others dissipate the energy as heat). That said battery balancers were still used in some lead-acid installations.

[nowty]

Just to add: the point of the balancing board isn’t to get all the batteries in the string exactly the same. The point is to prevent one cell drifting further and further from the others over a long period of time. This could cause undesirable consequences, eg if one cell is lower than the rest and the pack is discharged, the string voltage could be within spec but the one cell could be excessively discharged and become damaged. And the same for an over-voltage cell.

Exactly, I had one event prior to me fitting the balancers where I let the whole battery bank drop to circa 10% SOC. Two adjacent cells on one battery string had drifted over time and were at a lower voltage, conversely, the other 12 cells in the string were at a slightly higher voltage to match the string voltage of all the other strings which were connected in parallel.

Below shows the discharge curve for my batteries. You can see that at the bottom end, the voltage falls off a cliff. These two slightly lower cells, effectively were at maybe sub 5% and the whole battery pack was still discharging. This caused those two cells voltage to fall off the cliff and collapse further. In the end my Sunny Island battery inverter decided to shut down with a battery fault error as it detected a far too high voltage reduction for a relatively small amount of discharge and therefore decided something was not right even without running with a lithium cell BMS.


Conversely the same thing happens if you have too high a cell whose voltage rises rapidly at the top end of the charging cycle.

If you end up with a cell too high, it gets stressed and will prematurely age. It can also experience thermal runaway if its still being actively charged. The energy stops being converted into chemical energy and is converted into heat. If the cell cannot dissipate all the heat, it gets hotter until the electrolyte boils and then the metal can bursts with the excess pressure.

At the bottom end, a low cell may go into reversal and end up with a negative charge applied to it. If there is enough negative charge then copper bleeds out and migrates through the electrolyte and can cause a short circuit in the cell. A similar thing can happen if you charge at sub zero temperatures.

[nowty]

Later I received a second balancing board so I could cascade the two in order to balance the whole of a single battery bank of 5 batteries in parallel.



The battery balancers were fitted to the centre battery (battery 3 shown in yellow) of the 5 batteries in a parallel string, but my parallel fused cell interconnectors should ripple out the balancing to all 5 batteries, at least in theory. On one hand these balancers only use small amounts of current to balance up but then on the other hand, the drift without balancers is so slow I reckoned it would do the job.



And it did, they have been running now on both battery banks for around two years and all cells are always within around 5mv of each other. My local test point makes it easy to do a quick check but I do a full audit on the actual cells every 6 months or so.



My last battery cell voltage audit in March of this year, the yellow cells are the ones directly connected to the battery balancer.


I have also checked them after both a long heavy charge and a heavy discharge and they still remain in balance throughout.


And for Joeboy, etc, who may want to build up a supplementary 15 cell Lifepo4 to match their pylontech battery bank, there is a new version of the board which will do it nicely using a single board for around £50 on ebay or alliexpress.

[nowty]

To recap this thread,

• I built up a large 10 x 4.8kW battery bank using re-used Growatt batteries.

• I used them in a SMA Sunny Island battery inverter.

• I split them into two smaller banks with isolators and fuses.

• I used some simple active balancer boards to keep all the cells automatically balanced.

What I did not mention (to keep it simple) was that I also have another battery bank consisting of 8 x US2000 2.4kWh Pylontech batteries.



This is connected as a third battery bank in parallel with the other two.

The Pylontechs are LFP, 15 cells of nominal 3.2v. The max charging voltage is 54v, although I believe the batteries are full at around 53v and they go into a standby state.
The Growatts are LMCO, 14 cells of nominal 3.65v. The max charging voltage is 58v.
Therefore, as a compromise I have been charging all the battery banks at 54v which keeps the Pylontechs from alarming, but I am losing about 10kWh of capacity from the top end of the Growatt battery banks.

I have a relatively cheap idea to increase the voltage of the Pylontech battery bank, so it matches the voltage of the Growatt battery banks.

I am going to add a single LFP cell in series with the Pylontech’s, therefore making them 16 cells of LFP and will increase the charging voltage to around 56v or 57v.

10kWh of Pylontechs are circa £3k+ now so I bought these for £330, arriving here Monday. Hoping to unlock the same amount of capacity from my existing Growatt batteries for 10 times less dosh.

[Tinbum]

Not sure this is a good idea, but still thinking about it.

I see the new batteries are 120Ah but your Pylontechs are 50Ah. Are you proposing not to fully charge them and / or how are you going to keep them in balance with the pylontech cells?

Edit- i think i would have been inclined to remove one growatt cell.