SOC Question


  • Hello I am still learning about the 4S REC BMS. My question this time is related to the SOC. I question the accuracy of the process used to determine the SOC. For instance, I charge my small 20 AH 4S pack to about 3.65volts per cell (I must say I am very impressed as to how well the BMS balances the pack). The Charge End parameter is set to  3.68. The SOC meter tops out at 100%. I then load the pack with various loads from 1-9 amps. the SOC drops over time vs the load. At some point the SOC shows 20% or less yet my cell voltages are about 3.18-3.2 volts. I know that there is more capacity left in the pack as I continue to load the pack until the voltage is about 2.9 volts per cell. At that time the SOC shows 1-3% left. As a cross check I am also using a Victron BMV 712 (parameters are set on the 712 for typical LiFePo4) to monitor SOC. When the SOC on the REC shows 20%, the Victron shows around 40% SOC. Can anyone give me some ideas on how to understand this situation?


  • You need to remember that the SOC is a calculated value based on several variables and if you have a zero calibration error it can get out of tune very quickly with as little as a 0.5% error.  Further, as you are using a very small test system you need to allow for the parasitic loads of the BMS and its relays.  Also, at 3.18 volts you are likely starting to fall off the \"knee\" of the curve and have very little capacity left in your pack.  You also need to understand that the two different SOC meters you are using may be using different math to calculate SOC.  SOC works best in applications like cars and golf carts, where the pack is charged to 100% every night, and the SOC can be programmatically reset to 100% daily.  In a mid range utilization application such as a boat, you can not expect your SOC to be very accurate for any amount of time.

  • Thanks for your reply. Yes I understand and think my small battery capacity can alter the effect of the algorithm as compared to what I will have on the boat. I do realize they measure SOC by two different methods but hoped they would compare a bit better.

Please login to reply this topic!