@OverkillSolar Good to hear from you. Glad to have your input. My assumption of generic was that it appears (externally) to be similar to so many Chinese BMS units I have reviewed. I was thinking Overkill was just a generic reseller of those so my comments were a bit harsh. No offense was intended to you directly whatsoever!
The Chinese units that I have reviewed in the past were very rudimentary, didn't trip until extremes, didn't have much settings. Just "last-ditch effort" scenarios. I tend to try to avoid those BUT on my side I tend to try to operate in tighter tolerances and do the best I can to study the chemistry, whitepapers, field tests, lab tests etc to optimize. IMHO what happens at the end of every charge cycle is compounded and has an effect over time.
With respect to what I wrote above, it was less generic and more about the specific cells he is working with here. They are grey market and tend to drift worse than others because of the IR and other issues IMHO.
The 16s 100a BMS (available starting next week) has 20 Pairs of FETs, each rated 120a per the datasheet:
https://datasheet.lcsc.com/szlcsc/1912111437_CRMICRO-CRSS042N10N_C410926.pdf
I have not seen any of these fail closed under normal loads. (speaking about all the models) I tested a sample of this 16s unit by directly short circuiting it, and it was able to break the current twice before failing on the third attempt. It still broke the circuit, but the smoke came out.
I have tested the 4s BMS with a 300a load with no failures. Thermal management is the primary limitation.
Good to know. I worry about FET's failing closed, it has happened in the past but properly sized (by the user buying the right capacity BMS) it definitely sounds like it should be fine based on your comments above. Some BMS's use temp compensation with comms to the inverter to back down current or load shed as the temp rises.
I'm not sure about codes for DC systems, but a quick search found that parallel conductors can be used in AC systems per the NEC as long as they are properly installed:
In the 2017 NEC, parallel conductors inside an auxiliary gutter must be grouped together to prevent current imbalance in the paralleled conductors due to inductive reactance.
www.electricallicenserenewal.com
I'm glad you found that. I would recommend a re-read of the opening statement:
"Parallel conductor installations are covered in NEC 310.10(H) and are permitted for each phase, polarity, neutral, or grounded conductor in sizes 1/0 AWG and larger."
Anything smaller and if you lose one conductor, you are likely to have the insulation in the other one melt before any breakers trip, if they do. The thermal limits of the larger cables allow them to handle higher surge currents and the likelihood of one failing of that size is reduced significantly over say, an 8 awg.
I have to point out that the cells can not do this on their own, this only makes sense if there is a large active balancer connected.
Anyway, if this is happening, you need to top balance the cells, or the cells are damaged and mismatched.
I may be misunderstanding your comment here but if I'm understanding correctly though, here are my personal thoughts:
I do agree that a large difference is definitely a result of either bad cells or improper balancing. However, there is always still a small amount of imbalance at the top of the charge cycle with these cells or so I have found. And they continue to drift which gets worse at higher C rates and the farther in the knee a person goes (Ohm's law & fundamentals of the chemistry).
At the top of every cycle, there will always be a "first cell" that reaches the target voltage by 0.002 or greater probably, regardless of top-balancing. Here is a comment from the REC SI manual:
"
The communication between the REC BMS and the SMA Sunny Island charger is established through the CAN bus.
All the parameters that control the charging/discharging behavior are calculated by the BMS and transmitted to
the Sunny Island unit in every measurement cycle.
The charging current is controlled by the Maximum charging current parameter. It’s calculated as Charging Coefficient('C','H','A',’C') x Battery capacity.
The parameter has an upper limit which is defined as Maximum Charging Current per Device ('M','A','X',’C') x Number of Devices ('S','I','S',’N').
When any cell reaches the voltage interval between Balance Voltages Start and Balance Voltage End, the charging current starts to ramp down to 1 A x Number of Devices until the last cell rises to the End of Charge Voltage. At that point the Maximum charging current is reset to 0 A and the charger is disabled also via the BMS I/O interface. End of Charge, SOC hysteresis and End of charging cell voltage hysteresis prevent unwanted switching. SOC is reset to 100 %, Power LED turns ON 100 %. "
This is cell-level charging. In my humble opinion it's best to make sure you have every cell gently finish each charge cycle without oscillating over the target voltage so that is what I tend to advocate when a person can afford it.
Other BMS units actually modulate the charge voltage sent to the inverter/charge controller in every measurement cycle (~100ms) to more precisely control the end of charge current. For some applications, this provides even better current control than manipulating a charge current target value.
BYD battery systems simply shut down if you do not have communication. Tesvolt as well, along with other commercial and high-end residential systems. LG too as I recall.
I don't know why you think this would wear out the cells. If one cell goes overvoltage before the pack, this means they are not balanced. This is the entire reason why the cells need to be top balanced.
In a properly functioning system, the charging sources will limit themselves to stop charging at the correct voltage level (3.500v per cell, IMO), and the cells will all be below the overvoltage cutoff.
The wear effect is linear (I'm sure you aware of this but I'll mention it for others in case they are curious). The farther up the knee you take a cell for any given length of time, the more it's lifespan is reduced. Even commercial providers are finding this out after having put cells in the field and operated them for a few years farther into the knees. So for longevity of banks with cells that tend to drift, IMHO one needs to be aggressive about tapering the current as the first cell reaches the target voltage in every cycle.
Can a person just use a tail current and absorption phase? Yes, I think so but for those interested I would recommend a test with this and extensive cell voltage logging (every 100MS). Then graph it and see if (using a top balanced pack) any given cell oscillated past the target during the absorption phase. I think the results will be interesting.
I posted a photo above of a bank of cells being pulled/pushed out of balance. To get a concept of those, that is an LG factory-built and balanced pack. It was installed and operated for at least 1000 cycles (a bit of a guess but should be accurate based on the capacity loss) with a commercial BMS with 1000s of other modules. Each module had a BMS slave with passive balancing. And yet, without any individual cells being touched from the commercial installation, David Poz saw an imbalance and he felt he needed to cycle the BMS at the top a few times to bring them back into balance (no active balancing). Following that he had more usable capacity. As he cycles those cells, if no balancing at the end of or during every cycle is applied, they will once again drift out of balance.
The Overkill Solar BMS that I'm using (two of them) tracks the number of times the BMS has had to enable the high voltage disconnect for both the cell and battery levels. So far, those numbers are big fat zeroes.
My cells were well balanced (top balancing) to start with and continue to stay balanced. The BMS is doing some balancing. My charging devices are doing a good job of not providing more charge than necessary. I've put in the time to make sure all these devices are synchronized in their setup. I'm not charging or discharging at more than .5C. I don't see any problems.
Good to know! How many cycles in are you?
So again, no offense intended with my comments. I tend to be harsh in steering people away from the low-end BMS's that they may see are low cost now but could cost them money 2K cycles later. One of those "every cycle counts" scenarios. But I can see that is not the case with yours!