12V LiFePO4 with integrated BMS.
SmartSolar 75/15
PV panel up to 200W. The same behaviour always occurs at the battery connection.
If, while the battery is charging, the BMS “switches off” the battery — i.e. disconnects the cells inside from the terminals — this causes an overvoltage that rises to almost the PV open-circuit voltage.
Is the SmartSolar faulty?
Or is the MPPT control loop simply too slow?
What voltage did you set the battery to?
14.40 V.
After the overshoot, it is regulated back to 14.4 V.
Try removing the 14.2 V LiFePO4 profile from the presets.
It’s exactly the same with my cheap drop-in batteries and the MultiPlus.
The BMS disconnects the battery cells, and the MPPT or MultiPlus has nowhere to send the energy.
I seem to remember there have already been several posts about this, along with a discussion about the correct charging voltage.
If your battery switches off the charging port while charging, your cells probably aren’t properly balanced. Can you see the individual cell voltages? How high are they just before it switches off?
Hello Hannes,
they’re very basic LiFePO4 batteries without any communication, available in various configurations.
Using the battery manufacturer’s values of 14.4 and 14.6 V, I can’t get them to work properly with the MultiPlus. They do work with 14.2 V, though 
I’ve also left the battery connected to a bench power supply for quite a while, but switching it on and off didn’t make any difference, even after a long time.
You’ll find plenty of discussions on this forum about how high the cell voltage should be allowed to go at most.
But that is actually a state you should avoid at all costs. The MPPT’s behaviour in this situation is “normal”. Disconnecting the battery is not good at all, especially when the MPPT is operating.
This is a fault condition (or also a normal operating condition) that can occur when the BMS disconnects the battery while it is charging.
Apparently, there have already been several cases in the motorhome sector where, during winter storage, the BMS switched off the charging port because the battery was too cold. As a result, the SmartSolar output effectively no longer had any load, which led to overvoltages.
The few devices that were switched off, such as the heater, were then damaged by the overvoltage.
There are very few cases. The exact boundary conditions are not known.
I am trying to reproduce the fault and have also observed overvoltages depending on the SmartSolar’s operating state.
CH1: Output voltage
CH2: Input voltage
CH3: Input current
What I have found so far is that when the SR is operating in MPPT mode and there is a load drop from 100% to 0%, the buck converter continues operating actively for another 5 to 15 ms, charging the output capacitor and causing the voltage to rise to 20–25 V. It then decays within 30 seconds.
If the SR is operating under output-current limiting, it throttles back immediately when the load drops.
To me, this suggests that the MPPT control loop is significantly slower than the current- or voltage-limiting sections.
However, I do not understand why the voltage-control loop does not take priority over the MPPT.
The buck converter obviously can.
Try setting it to 3.370 V/cell, or 3.45 V max., and then check the process??
When the motorhome is stored over winter, there are normally no loads connected, so there is no voltage spike when the system switches off due to frost. Besides, this usually happens at night or when there is no sunlight. So under normal circumstances, it doesn’t pose any danger.
The fault occurs during charging; in that situation, the battery is the load that is disconnected by the BMS.
In summer, it can occur precisely when the BMS disconnects the battery because of an individual cell overvoltage.
Think about what happens when the operating point of the MPP tracker is located at a second MPP with a lower voltage due to shading. Then the irradiance increases. The operating point shifts to a higher voltage in order to reduce the power. But what happens when the shading moves away? The power increases, and to reduce it, the operating point has to shift further toward the higher voltage. In this case, however, the power is not reduced but increased significantly. As long as the operating point has not passed the MPP, the power cannot fall, and an overcurrent occurs. Since there is no battery, this manifests itself as an overvoltage. I suspect that no inverter with its capacitors is connected either, which would absorb the current and allow only a slight overvoltage.
When charging the connected batteries, there is an IU characteristic curve, after all. So there is both current limiting and voltage limiting (the end-of-charge voltage).
These are control loops as well.
Regardless of what happens at the input, surely the output voltage should enter voltage-limiting mode.
The BMS should actually communicate with the charge controller before it has to shut down, which is why I don’t really see the BMS shutting down as a normal operating condition. Still, it’s an interesting experiment, even if this is getting a bit too scientific for me here. 
It does! But that can only be achieved by shifting the operating point. Even switching off abruptly would cause the energy stored in the charging inductor to be released
In the motorhome sector, the “normal” state is that the solar charge controller does not communicate with the battery BMS.
All sorts of manufacturers are combined with one another.
I can see from the input current that, after the load is switched off, energy is actively being transferred from the input to the output. This takes several milliseconds—in this case, 12.8 ms.
It can’t be just the energy in the energy-storage inductor, because the switching frequency is 30 kHz. That doesn’t explain the milliseconds.
Fault condition – yes
Normal operation – no
If the BMS has to interrupt charging repeatedly, something is wrong.
The simplest thing to try would be to reduce the charging voltage, as already mentioned.
You could also try significantly reducing the charging current so that the BMS has more time to balance the cells.
Ideally, you can use an app to check the battery and view the individual cell voltages.
It is certainly a normal situation that will occur more frequently as pre-purchased LiFePO batteries installed in existing systems become more widespread.
For example, a motorhome sits unused all winter. The cells are slightly unbalanced, and there is not enough time during the first charge. The BMS switches off.
Regardless of how often this situation occurs, it is not good if an overvoltage then occurs and damages other devices.
Right. Let’s get back to the SmartSolar’s control behaviour.
Why does this overvoltage not occur with a shore power charger?
Why does this overvoltage occur only with the SmartSolar?
And why only when the SR was in MPP mode before the load was disconnected? (millisecond range)
When the SR is limiting the output current, the overshoot is many times smaller (µs range)
Because a shore-power charger may not experience a sudden rise in input voltage.
That said, in practice the whole situation looks completely different from your tests.
In a motorhome, all the loads are probably only very rarely switched off at once. So the output capacitor cannot charge up as much, or it will be discharged again more quickly.
If the charging port switches off due to an imbalance, the charging current may still be as high as 2 A. So that’s not a problem for the solar controller either.
The same applies in freezing conditions. How much charging current is actually produced then? In my case, it rarely exceeds 2 A. And for that to happen, the controller would really have to switch off in full sunshine. How likely is that?
