When Samsung or Apple have battery issues, battery vendor reproduces the failure condition and Samsung or Apple update their charging/monitoring configuration. The problem goes away without replacing millions of batteries.
Here GS Yuasa continues to claim there is no failure scenario. So neither Thales nor Boeing can offer a fix to charging/monitoring system, other than putting it is a box.
I'm not sure where you or Wlederling are going with discussions about the charger, since as far as I know the charging system was found to be working properly in both of the incident aircraft as well as in additional production chargers by Boeing, the NTSB, and Underwriters Laboratories. While the incident batteries were too damaged to conclusively identify the cause, internal short circuit appeared to be most likely, and they did observe manufacturing problems at GS Yuasa that could lead to internal shorts that are known to be capable of causing thermal runaway.
I don't know about Apple, but in the case of the Samsung Note 7, it was not a charging issue. They had a battery manufacturing defect. Actually, if I remember right, they had two different manufacturing defects, either of which could cause the runaways, and they were able to identify those in part by dismantling a large number of intact cells.
i.e. don't try to cram 100% charge into the battery when you have limited knowledge of the actual charge state.
i.e. 75% +-3% known charge state and then blindly send 25% equivalent of juice in you have 50% chance of overcharging.
( the basics of the charger algorithm were published during the NTSB proceedings.)
As far as I can find, the flight data recorder monitored an indirect voltage on the lines somewhere downstream of the battery, but the battery management unit monitored it directly. The BMU data was used only locally by the BMU, not transmitted to the FDR, except for a fault indicator if problems were detected. The BMU actually monitored each cell individually, could balance the charge of each cell, and would stop all charging via a relay built into the battery box if any cell exceeded a set limit.
The charge limit was 4.025 V per cell (Docket SA-536, Airworthiness Factual Report Exhibit 17-J, page 29). Normal full charge for lithium cobalt oxide cells is 4.2V, but GS Yuasa appears to have specified the pack to have a nominal full charge of 4.025V per cell. From what I know of lithium ion batteries, I presume this was to keep the battery in a conservative operating range, not due to some novel chemistry they might have been using. In other applications, 4.025V would be around 80-90% SoC. Overcharge testing also was done to 36V (4.5V/cell). This will degrade the battery but should not result in a fire at that point.
I'm still looking through the Boston incident docket and not seeing the claims you're making about the charging algorithm.
That would be an even worse charging algorithm than you're suggesting, because you can't determine state of charge for a partially charged lithium ion battery accurately enough to handle charging that way. Coulomb counting starting at a known state of charge comes close, and is often used for meters, but not for charging. Full charge voltage and full discharge voltage are the known points. The UL report describes the standard lithium-ion charge algorithm - maintain a constant current of 46A until the voltage reaches 32.2V for the pack (4.025V/cell), then switch to constant voltage, at which point the charge current will naturally taper off as the battery approaches full charge.
* Edit - I wonder if some other party might have made some commentary on charging that was misunderstood. The constant current to constant voltage transition usually occurs somewhere around 75% SoC, depending on the charge rate. However, from that point, a fixed amount of energy is not passed to the battery. Voltage is held constant and termination occurs once the charge current has declined to a certain level.
Last edited by iamlucky13
on Tue Dec 05, 2017 8:52 pm, edited 1 time in total.