Guru
Joined: 13/10/2014
Location: AustraliaPosts: 2232
| Posted: 06:01am 11 Jan 2024 |
During the course of looking for serial data command info for the DALY BMS, I was reading about the problems being reported by users of the DALY BMS, and apparently these so called failures are one of the reasons some changed to another brand of BMS.
Even some of the DOA complaints looked like possible user error. There are complaints of incorrect cell balance voltages, and claims that when measured with their trusty meter, the meter said it was all good, but looking at the typical steps taken to find the problem, not once did I see the obvious question, or the correct way to test and measure for this apparent error. Swapping the BMS would likely fix the error by default, but there is no way to know.
Now I agree, the software is still a bit amateurish, so many models to be catered for with the same program - dumb. There is no way to really set it up for the correct charging mode for LiFePO4, but this is the same for almost all the affordable BMS units out there, so I'm not going to change because of reported user errors of which some appear to be dubious.
I see a lot of users making mistakes early on with settings and construction, and the crap excuse for a manual may be to blame in some cases.
Anyway, I decided to experiment with making the BMS do something that a BMS is not meant to, but humor me, it revealed a hidden problem and possible failure mode.
I configured the BMS to switch off the charge MOS above a lower specified voltage, this voltage should be around 90% of charge, by programming the Solar chargers to run at 1 volt higher, it should cause the batteries to be "SEMI" isolated from the DC supply system, unless pulled down by a big load on the Solar charges or Solar output drops off.
The Four Solar chargers I have would, and they can with ease, power the inverter, I have plenty of Solar.
So once the battery is at roughly 90% and the Charge MOS switches off, there would be minimal small voltage cycling of the battery, which is bad for LFP.
It did work, and dumping a 2kW load onto the system already powering the Home, workshop with Shed AriCon, does not cause a sag in the voltage from the Chargers, so the Battery MOS does not reconnect.
However, one must realize that there remains the possibility a momentary high charge current if the charge MOS reconnects with the Solar charges not loaded down and 1V higher than the Batteries - which had been disconnected for a few hours. Like I said just a crude "what if" test.
It highlighted a potential problem, and I would assume for most BMS devices using physical external communication connections cables to anything that shares a common battery Negative connection.
Think of four BMS devices, each with a cabled COMS link back to a single display and sensor device, and this device would likely have a connection to Inverter Neg / system ground. Even without that connection, each BMS is connected to the system Neg Rail and one has an overcharge MOS switched off, the four are plugged into the same monitor device, so all BMS are connected to a common ground via coms leads and BMS negative leads.
The Common mode BMS with only Negative battery switching is going to partly isolate the Negative terminal of its battery on charge over-volt-trip. But the coms lead shares the common Negative ground of the BMS (battery NEG terminal). If the Charge MOS is off, the coms lead now becomes a Negative connection between BMS Battery Negative connection and the BMS connection to the System (load) negative rail, bypassing the MOS. What could possibly go wrong!
Anyway, I'm going to program a micro to correctly disconnect the Positive lead of the batteries when the correct charge voltage and charge tail current conditions are intelligently met - and reconnect when the SOC has dropped to a predetermined lower level.
I have an idea for a fail safe way to pull current instantly from the battery bank under high transient load demands, and it would allow the use of high voltage high current electric Vehicle power solenoids instead of FETS if desired, I've been using these solenoids in the Inverter - fabulous devices and low power when enabled.
I'm sure if I read this back it will sound confusing - so I won't
.
Edited 2024-01-11 17:38 by KeepIS