Incorrect SOC indicator on my Cycle Analyst v3

[Fastolfe]

I have a 24V 24Ah LiFePo4 battery pack in my bike. I configured the right chemistry ("LiFe"), the right number of cells (8) and the right capacity (24 Ah) in the CA. The open voltage it displays between 90% and 10% charge seems correct (26.something V). Yet the SOC indicator only shows full when the battery is fully charged, and drops to empty after only 2 or 3 Ah - not even close to the true capacity of the battery pack. Everything else works fine, it's just that the fuel gauge thingy is pretty much useless.

Anybody knows what I might be doing wrong here?

 

[teklektik]

I have a 24V 24Ah LiFePo4 battery pack in my bike.
...
the SOC indicator only shows full when the battery is fully charged, and drops to empty after only 2 or 3 Ah - not even close to the true capacity of the battery pack.
...
Anybody knows what I might be doing wrong here?

Nothing - this may be a CA issue (for LiFePo4 only).

LiFePo4 is handled somewhat uniquely in the V3 because of its flat discharge curve - it is very difficult to determine SoC from the open circuit voltage. The CA normally combines expended AH and open circuit voltage to arrive at the SoC, but in the case of LiFePo4 it has classically weighed the Ah component very much more heavily owing to the unreliability of the voltage measurement for that chemistry.

A release or so back we made a small adjustment to the V3 LiFePo4 SoC algorithm to pay more attention to the open circuit voltage. I noticed some wonky SoC behavior but it was unclear if other factors on the test bike with LiFePo4 were driving it. However, this SoC indicator issue has appeared more pronounced on recent rides over the holidays in much colder weather (saggier battery voltage).

Anyhow, this matter is on the agenda for discussion this week.
So - I think for now you should fall back to Ah as the most reliable indicator of SoC.

 

[Fastolfe]

I see. Well that's alright, it's just a small cosmetic issue anyway really: As you say, it's easy enough to simply read the Ah and guestimate the remaining charge. Thanks for your explanation!

 

[teklektik]

The very flat LiFePo4 discharge curve introduced unforeseen problems with the mod mentioned above which was actually intended to improve accuracy - apologies for any difficulties this caused. The SoC calculation for LiFePo4 has been reverted to the original design to more heavily weigh Ah expenditure.

This corrects the SoC indicator problem reported by the OP -- affects LiFePo4 only. The revision will appear in new firmware to be released shortly.