Guru
Joined: 13/10/2014
Location: AustraliaPosts: 2232
| Posted: 01:40am 19 Jan 2024 |
Hi Aaron, great to hear: Apology for another long post.
As you likely know, lot of talk about when to start Active balance, IMHO there is no one rule for that, it depends on the Cell mix condition, typical charge current and the selected charge voltage for the system.
I found that cells in all three 48V packs are fine with a 3.3v start, not the 3.4V or more as some "experts" recommend, in my case due to the lower full SOC and charge voltage.
I also need to protect the batteries from over current charge, apart from adding an extra protection layer when running them in parallel, I've had occasions under sudden drop in the night time and early mornings inverter load usage, and with batteries at a lower SOC, the charge current was close to 100A into the 130AH pack, and there is still 20A into the load, and not all of the solar was connected.
I now have three banks in parallel, 2 x 170Ah and 1 x 130Ah. I'm just building the fourth bank of 170Ah cells. Once the fourth bank goes in it should never trip my 60A charge setting on the 130A bank.
Four banks allow me to use an 80A (or LESS) fuse in each bank, I can draw a 670A surge peak DC input for the big workshop machinery loads and the fuses are fine.
At the moment, with only the three banks in parallel, my remaining SOC prior to Solar charges starting is 86% for each 170A pack, and 65% for the 130A pack.
In the above result, I'm only charging to 3.346 per cell and then cutting the charge, no tail current at the moment so I might be lucky to be even at 85 to 90% SOC. The Cell voltages quickly drop to 3.34v per Cell unloaded. Hardly any Active balance, if at all.
With this usage and the intermittent nature of most of the big daily loads and with four 48V battery banks, my maximum charge current should be limited to around 30A to 40A per battery. The three banks track each perfectly during the charge cycle and SOC calc right up to the charge cutoff voltage.
They have recharged from last night and morning usage by 9.45am. All are indicating 93% SOC and sitting at exactly the same Cell voltages. This is confirmed by the JUNCTEK Current shunts and Charge monitors, however the initial SOC set values could have been incorrect. These JUNCTEK are wired between each Blue BMS Bat Neg lead and the Cell pack Negative terminal, the voltage Sense lead is connected to Cell Pack Positive. These are within a few mv accuracy of each other and the three BMS units.
Now, I'm fully aware that this is not the correct way to cut charge, however with these settings I'm not going to get much above 85 to 90% SOC, and as my discharge SOC is modest to say the least, these Cells should be quite happy, especially in their AIR conditioned, dry and fully lined room. Just need a top balance every 6 months or so.
BTW the Active balancer has completely stopped any cell getting more than 30mv in front as they approach the point of voltage cut off, and this allows the Voltage trip point method to work, for now!
I was also looking at the waveforms on DC cables from Inverter to BMS with a 370 MHz BW DSO. A lot of the high frequency noise is from the Solar chargers at higher currents. I already have HF bypass caps and toriod suppression, so I decided to try a 60,000uF CAP bank across the DC cables between the AC Inverter and the Cells and BMS, this made a huge difference to charge cutoff voltage stability and obviously residual low frequency noise.
I did not want this high value of capacitance close to the Inverter FETS, there is already 40,000uf directly across them, more than enough to dump a few thousand Amperes and launch the inverter FETS into orbit with an over current fault condition.
These new caps are at the end of 1.2 meters of 0G cables coming from the AC Inverter, and mounted very close to the BMS, battery packs and Solar chargers. The only thing is, they also needs an Auto precharge circuit. These are so simple and easy to implement and I would never trust myself with a manual switch, with this value of cap and the low internal R of these units, it's just a matter of when something is destroyed, not if.
BTW That is not in response to your setup and precharge as they are different, it's just what 640AH of LFP and 0G cable feeding these caps can, and will do.
The current wave setting is just a sensitivity setting for small current (noise) display, so it doesn't hunt around at zero, I think that is also in the Manual somewhere, the fact that they needed this confirms my suspicion of limited HF noise filtering to the ADC current input of the ARM Micro.
FYI: The old 12V drop-in LFP battery sets used 100A DALY BMS and I also flogged those, seriously so, and not a single failure, and there were four 12V BMS units connected in "series" between 56V in each of the three parallel 48V banks.
You may be right on the cycle calculations, I'm just so busy getting all this redone that I haven't had a chance to even look at it.
BTW I'm sick of this humid weather and rain, spent the last few weeks in the Air Con workshop. There is a narrow path that I manage hack out each morning, through what seems like 2m high grass to reach the workshop, it's almost overgrown again by late afternoon, and back to a jungle by morning, rinse and repeat
.
Edited 2024-01-19 13:44 by KeepIS