Calibrating SOC

[FFMan]

Fingers crossed the EV will be road legal in a few days. So thoughts turn to calibrating the SOC calcs using data from the shunt.

I've started the build with a 360v 9kwh pack from a BMW.

What voltage per cell do people routinely charge their packs to ?

What voltage per cell would you consider a safe discharge point ? And for this i should use the lowest cell reading from the BMS ?

To calibrate my soc i guess it is determined empirically, so a full charge, and drive it until i reach my minimum level and record how many amps where consumed ?

In terms of the driving it until the battery is at minimum, as it approaches the lower voltages, i assume i stop to get an accurate voltage reading ? Should I wait a bit to let it settle or might it bounce up and affect the use in the real world where stopping could be less likely ?

In terms of the shunt data, it returns an instant current reading as well as a count. Which figure do people use and how do you maintain a total. I assumed the count is the number of amps (?) since the last reading was sent, so I am accumulating this, but I'm not sure this is correct.

any advice much appreciated.

[arber333]

Fingers crossed the EV will be road legal in a few days. So thoughts turn to calibrating the SOC calcs using data from the shunt.

I've started the build with a 360v 9kwh pack from a BMW.

What voltage per cell do people routinely charge their packs to ?

I generaly go to 4.08V per cell. NOT more than 4.1!

What voltage per cell would you consider a safe discharge point ? And for this i should use the lowest cell reading from the BMS ?

I would not drive below 3.3V per cell static and/or 3V per cell dynamic. Today cells can survive below 2.8V but the problem is the heat generated with higher internal resistance which would quickly raise temperature of the cell and cause the death of it or even fire!

To calibrate my soc i guess it is determined empirically, so a full charge, and drive it until i reach my minimum level and record how many amps where consumed ?

In terms of the driving it until the battery is at minimum, as it approaches the lower voltages, i assume i stop to get an accurate voltage reading ? Should I wait a bit to let it settle or might it bounce up and affect the use in the real world where stopping could be less likely ?

In terms of the shunt data, it returns an instant current reading as well as a count. Which figure do people use and how do you maintain a total. I assumed the count is the number of amps (?) since the last reading was sent, so I am accumulating this, but I'm not sure this is correct.

any advice much appreciated.

I am slowly working with that. SOC should be counted via Ah consumed with some margin for dtemp and peukert effect, which is still present with Li chemistry... You would get some sort of linearization from one knee to the other.
I would take lower knee and transition to 0% with some 15% before true empty. That would give you some Km as a reserve.
At the higher knee i would transition to 100% at 4.05V per cell for complete battery voltage.
Simply jump to 100% every time battery stays at 4.05V for more than 30s. Would that be enough time?

I know that would mean you could jump from 85% directly to 100% with fast charging but that shouldnt be too bad. You need to consider from 25% you could go directly down to 0% too.

If you implement that you would be recalibrating your SOC level every time you charge your cells.
On a side note you would have a conservative estimate for SOC when previously empty, thus safer...

[Bigpie]

Not in any scientific way, I charge until a cell hits 4.15V, that pretty much settles at 4.08V. I'll typically not go below 3.2V day to day, though have had to push to 3.0V and even slightly under to make it to be able to charge on a couple of occasions.
I just took 20% off my packs stated Ah capacity and use that as my 'usable' Ah, use the shunt to count the consumed Ah and use that to calculated the %.

[muehlpower]

Even if you don't use an Orion BMS, there are hints on how to do the balancing. Also how the SOC is determined. Here are the instructions:
https://www.orionbms.com/manuals/pdf/or ... manual.pdf

[FFMan]

Even if you don't use an Orion BMS, there are hints on how to do the balancing. Also how the SOC is determined. Here are the instructions:
https://www.orionbms.com/manuals/pdf/or ... manual.pdf

thanks that is useful