Anyone who has built a 16S lithium iron phosphate pack around a Seplos BMS has stared at the cell list at some point, wondering whether 3.338 V next to 3.341 V is a problem. Usually it is not. The trick is to read the spread at the right moment, and to follow it over weeks rather than minutes.
Why the spread depends on the state of charge
LiFePO4 cells have a very flat voltage curve between roughly 20 % and 90 % charge. Two cells holding noticeably different amounts of energy can show almost the same voltage on that plateau. At both ends of the curve the voltage changes quickly, so small capacity differences become visible as larger voltage gaps.
| Moment | Typical healthy spread | What to make of it |
|---|---|---|
| Resting, mid charge | A few millivolts | Tells you little about balance: the plateau hides differences |
| Charging, approaching full | Can widen to several tens of millivolts | Expected; the BMS balancer starts working here |
| End of a long full charge | Back down to a small value | The real test of a well balanced pack |
| Deep discharge | Widens again | The weakest cell shows itself first |
How balancing works on a Seplos BMS
The Seplos boards balance passively: when a cell is above the others near the top of the charge, the BMS bleeds a small current from it through a resistor. That current is modest, so balancing a pack that has drifted takes time spent at high charge, not one quick visit to 100 %. If your inverter stops charging as soon as the pack reaches its target voltage, the balancer barely gets a chance to work.
- Allow the pack to stay near full for a while on sunny days, or schedule a longer absorption from time to time.
- Check that the balancing threshold set in the BMS matches the charge voltage your inverter actually reaches.
- Never change BMS protection settings without understanding them: the box only reads them, and so should a quick fix.
Patterns that deserve attention
A single odd reading is noise. These patterns, seen over several days, are signals:
- The same cell is always highest at the end of charge. It reaches full first and trips the protection early, which shortens every charge. Longer balancing time usually helps.
- The same cell is always lowest at the end of discharge. It may have lower capacity than its neighbours, and it sets the usable capacity of the whole pack.
- Spread at full charge that grows month after month. Check busbar torque and cell connections first: a loose terminal looks exactly like a weak cell.
- A cell temperature clearly above the others. Often a connection heating up. Worth checking with the pack switched off.
Follow it with history, not snapshots
The difference between a pack that is fine and one that is drifting only shows over time. Seplos Monitor reads every cell from the RS485 port, shows the spread live on its screen and on your Wi-Fi, and, with the free Cloud, keeps up to three years of curves. Look at the spread at the end of each full charge for a month: if the line is flat, your pack is healthy; if it slopes up, you know which cell to look at, long before it costs you capacity.
Because the box is strictly read-only, it never touches the balancing or protection settings you chose for your build.