Determining Battery State of Charge (S-o-C)

[Bikes4two]

• I use a very simple handlebar display with my TSDZ2 (the VLCD6) which doesn't show actual battery voltage but the typical 4-bars 'fuel guage' system for showing S-o-C.
• It isn't even mildly accurate (and I didn't expect it to be) but the OSF Java configurator offers the ability to set the various voltage levels vs S-o-C
• So what might these settings be?
• Mr Google has given me a graph (below) for a typical curve and I can see from the non linearity of the battery cell discharge that determining the 3/4, 1/2, 3/4 points is going to be a very fine thing.
• Does anyone have experience of setting reasonably accurate S-o-C points or even a more granular S-o-C vs Voltage table/curve?

Many thanks, B4t

 

[Nealh]

Batteries very rarely have graphs as linear as that and it is very far from typical when a load is applied.

This LG mh1 is very different.


Samsung 29e


LG hg2

 

[Bikes4two]

Batteries very rarely have graphs as linear as that and it is very far from typical when a load is applied.

• Thanks for the reply @Nealh and I've book-marked the lygte-info.dk as it has some useful battery info.
• Who knows what batteries (and their discharge curve) are in my Chinese battery pack.
• I've had a little play with the graphs you provided and the Samsung 29e and LG mh1 have similar discharge curves and I shall adjust my VLCD6 configuration to reflect their extrapolated values as they are a better starting points than I have.
• In the longer term I could if really bored, connect up a temporary voltmeter and cycle around the local roads in a circuit, noting the voltage at say, every 5 miles until the BMS cuts off the battery and then work it back from there! I probably won't though

So the setting I'm going to try out are:
State of Charge (S-o-C)
75% - 3.65v
50% - 3.4v
25% - 3.25v

 

[Nealh]

For ebike use 3.2v would be 0 - 10% most BMS will cut at or before this value.
75% approx. 3.9v.
50% approx. 3.7v.
25% approx. 3.45v

 

[sjpt]

Don't forget that the voltage changes a large amount under load, often enough to drop 2 bars on our 5 bar battery life/voltage display. So even if you manage to have a more precise reading and to calibrate the voltage to percentage curve properly you will still need some extra level of interpretation of the result.

A cumulative current measure is the more reliable way. I think (not sure, others may correct me here) that our Bosch has this. However, it completely spoils it by converting remaining battery (which I suspect it knows reasonably accurately) to a range based on such an absurdly short riding history as to render it almost useless.