Different capacity packs in parallel, discharge ratios

[Alfalfameister]

Please explain something to me:

I have two packs in a battery bank. Pack A: 16S 105Ah, Pack B: 2P16S (280Ah cells) 560Ah. Each has their own Xiaoxiang BMS (aka Overkill). Both are connected to a busbar, to make the battery bank a nominal 665Ah bank.

So, 105Ah and 560Ah packs - 16% and 84% ratio.

Last night (until 05:30 the next day), I was discharging them a steady 1180W (as read by one Current Transformer - may not be accurate, but it's close). Took readings of each BMS. Total voltage was/is within 0.1v of each other (ie, like 52.89 and 52.81 or something like that), which could just be BMS's readings of each pack. I didn't take a voltage reading of the battery bank as a whole.

One would think that the discharge ratios would be 16% and 84%. But here are the readings I took when I had the chance (have to manually look, no logging, which would have been a LOT better):

Readings from EACH BMS:

About 17:00
7.5A and 15.5A (33%/67%)

18:00
5.8A and 17.2A (25%/75%)

21:30
1.3A and 21.7A (6%/94%)

03:30 (got up to pee, might as well look)
5.1A and 18.2A (22%/78%)

05:00
6A and 17.4A (26%/74%)

06:00
4.9A and 18.6A (21%/79%)

Turned on water kettle to make coffee:
17.4A and 46.4A (27%/73%)

07:00
0A and 9A (0%/100%)

Voltages are still within 0.1v (as they should be - they're probably identical, but each BMS reads it differently, and there might be losses from battery terminals all the way to busbars anyway).

SO... I am THEORIZING that they AVERAGE about 16%/84% anyway (if you check the ratio at 21:30, it's 6%/94%... and I didn't get to check at 21:00 or 23:00). They should still hit the LVD and HVD at the same time (because they are VOLTAGES), but SoC would not be identical throughout (but they would be close is my guess).

At first, I was worried that at the beginning (30%/70%), the 105Ah pack, if that ratio had remained, would definitely be discharged first. But I guess they even out anyway. Kinda like there's a stretchable string (in terms of SoC, not the voltage, which remains nearly identical) that when it gets to a certain point, they pull each other towards an equilibrium.

I don't really have a question, just thought I'd share, and maybe someone can explain why what happens, happens.

 

[snoobler]

Nice work.

The two batteries have significantly different internal resistances as well, which impacts voltage drop. It's also good to double check your connections are tight and in order. The flatness of the discharge curve can make for some unpredictability of which bears what load.

It's worth directly checking the individual batteries for their voltage.

Since they're voltage-tied, one can't deplete significantly before the other.

 

[Alfalfameister]

The two batteries have significantly different internal resistances as well, which impacts voltage drop.

And, there you go - a very succinct explanation of why that is, and explains it in considerably less words than I would have.

Since they're voltage-tied, one can't deplete significantly before the other.

Good to know.

As of right now, in my part of the world where the sun is currently at high noon, the 105Ah is charging at 11 amps and the 560Ah at only 23 amps (I have other loads in the house being used). I suspect that as the 105Ah gets more full later, the ratio discrepancy will get bigger, and more juice will go to the bigger pack.

EDIT: 30 minutes later, the 105Ah is charging at 5+ Amps (down from 11A), and the 560Ah is 23A

 

[Craig]

Picture your two batteries as pyramids. Now turn one upside down so the point is down. You have not changed their capacities . Now imagine them full of water. If you drain the water of both 1 inch the upside down battery will obviously drain out much more water than the upright one. Batteries are not necessarily like pyramids but the metaphorically speaking the "shape" of the storage area is different between almost every cell. That is why at the extreme top and bottom ends they drain/fill at different rates. I know this is kind of convoluted but hope it makes some sense

 

[snoobler]

Dude. Really? You had to use another "pointy" reference?

Sheesh.

 

[Craig]

Dude. Really? You had to use another "pointy" reference?

Sheesh.

LOL sorry I didn't even see that one coming.

 

[GXMnow]

This is basically what I expect to happen. On a few threads, I have said you can parallel different batteries, but don't expect to get more peak current. The capacity will add and they will share the current, but you can't expect them to always share evenly. This is only running about 24 amps, and at one point you have one bank taking 94% of the load. This is a large imbalance with very different banks, but even if they were very similar packs, you could see the current sharing sway up and back like that. My two identical strings slip about 10% up and back at moderate currents. At 40 amps, they moved as far as 19 and 21 amps at one point, but it did balance back out and even slid the other way a bit on another cycle. If any one string can handle all the current, it will never be a problem. If you need more current than a single string can handle, then you need to make sure they are as well matched as possible and still over build.

 

[Alfalfameister]

If any one string can handle all the current, it will never be a problem.

Good to know! Currently, max I can draw is about 60A (66A is max of inverter, and I set it to 90%).

Future inverter I'm going to set up to replace current one, though it has 190A capability, I'll set max to 100A only (just about 1C of the 105Ah pack), but it will RARELY even see this. It will hover on the 25A discharge region for the most part, with occasional forays into the 60+A region (ie, make coffee). Anything in excess, I get from the grid (ie, if I use 100A with my current setup, it's 40A from grid and 60A from inverter; with future inverter (Deye 8kW aka Sol-Ark), I'll set to 100A max draw, but again, it will rarely ever even see that).

So, should the 560Ah pack fail, the 105Ah can still handle it. If the 105Ah pack fails, the 560Ah should handle it as well.

I kinda added another pack to the bank not for more current, but for more capacity.

I use it for grid sharing at night (from late afternoon to 5:30am the next day, no matter the loads, it releases a continuous 23+A or a little under 1200W, unless loads become less than that, in which case, it only releases whatever the loads are).

 

[Alfalfameister]

[double post]