NexPower V3 hybrid battery unveil - Sodium-ion battery

It all depends on your expectations. From a mathematical point of view, this is definitely a non-linear graph. But if you look at other types of batteries, then this is quite a linear characteristic.

 

Certainly, in a graph like that, a person could say "this is almost linear", without sounding bonkers, if talking about the portion between Vexp and Vnom. That'd be, eyeballing it, the portion between about 1.28 and 1.22 V in the graph you attached to #99 (on the red trace).

The portion between Vexp and Vfull is a lot harder to call "almost linear" with a straight face, and to try to take the whole region Vfull to Vnom and call it "almost linear" would pretty much demand giving the word 'almost' hazardous-duty pay.

If I remember right, this little detour first arose from Fred_H's question in #94, which I think might be restated as a comparison of two ways to estimate the difference in usable capacity between two battery chemistries if the BMS keeps the working region between the same two voltage breakpoints.

It looks like Fred_H, following mudder, went with an early conversion from V to SoC for each chemistry, and then compared the resulting ΔSoCs.

I think that Fred_H, in #94, is asking why/how it improves the estimate any to use ΔV and a stated assumption that "the characteristic is linear", when the stated voltage ranges extend well up past Vexp and therefore the relationship may not fit the stated assumption very well.

It does seem to me the approach mudder and Fred_H used may depend less on the assumption of linearity.

 

Welcome to the world of analog signals. Such a signal is subsequently approximated by the BMS and forms a linear signal.

Simply taking 30% of the SOC is also not entirely correct. Look at the graph

 

If the BMS linearizes the input according to a known curve, then it is doing what it's supposed to, and that's not the same as us treating the curve as if it's a line.

If the shape of the SIB curve is that radically different from the shape of the NiMH curve, that'll be another fly in the ointment when trying to interpret what the NiMH-calibrated BMS thinks the SoC of the SIB is—another topic for another day.

Nonetheless, the basics of the mudder/Fred_H approach to estimating the usable capacity still seem sound: find the voltages corresponding to 40% and 80% SoC according to an NiMH curve, then find the SoCs corresponding to those voltages, using an SIB curve, then subtract to find the difference of those SoCs.

Maybe it'll come out something other than 30% if done carefully using the SIB graph above. (I haven't done it here just now.)