As solar charge controllers can be damaged if no battery is connected to their output, I would like to ask what measures can be taken if the connection between the battery and the solar charge controller is interrupted. This interruption can occur manually, for example if you switch off the battery’s main switch without remembering to disconnect the connection between the solar panels and the MPPT first, but it can also be caused by the BMS intervening. I don’t understand why MPPTs don’t switch themselves off when they no longer detect any current at the output. Victron’s SmartSolar MPPTs can be switched on and off remotely. So I don’t understand why the MPPTs cannot switch themselves off via this “switch” when no battery is connected any more. Technically, this shouldn’t be a problem at all.
As I have the Victron devices integrated into Home Assistant via MQTT, I was thinking of setting up an automation whereby, if the solar charge controller’s battery voltage drops to 0, it switches the solar charge controller off. Would a delay of perhaps 1–2 seconds still be dangerous for the solar charge controllers? What do you think of this solution? Is there a better way to protect the solar charge controllers?
Perhaps the misunderstanding stems from the fact that you think current can be felt
But jokes aside, an MPPT is a circuit that pumps energy from the PV input to the battery output. This is why such a circuit is also referred to as a charge pump. Even without a battery connected, there is still a voltage at the battery output, because the charge pump continues to operate and regulates it to the set charging voltage. Using current flow as an indicator only works to a limited extent, too, since no current may flow when the battery is full.
The problem with potential damage is that this charge pump has a certain regulation delay. If the battery suddenly disappears, for example because the BMS trips or the main switch is turned off, the charge pump would have to stop working immediately. However, this is technically impossible. There will therefore be a brief voltage spike at the battery output. Depending on how much energy was being pumped at that moment (the current PV output), this spike may be larger or smaller. If it becomes too large, components may be overloaded.
But it isn’t quite as bad as that, because the battery is usually not the only load in the system. A connected inverter also consumes the energy and dampens the voltage spike.
On the one hand, you should configure your system so that the BMS never trips, because it is a protection function and should never trip during normal operation. I also don’t consider a main switch to be a sensible solution; I would regard it merely as an emergency shut-off switch, to protect the system from damage in the event of a problem. If the MPPT were then to fail, that should be regarded as a consequence of the system fault.
@steffen-graap Thanks for your very vivid and very clear explanation and answer. So, there’s no point checking at the MPPT’s battery output whether there’s any voltage present in order to determine whether the battery is connected, since when it’s sunny and the MPPT is switched on, there will always be power there, whether or not a battery is connected. As you quite rightly pointed out, my 123SmartBMS only disconnects the battery in critical situations, so never during normal operation, and the main switch is there only for emergencies. Let’s just hope that the Victron devices and batteries do their “job strictly by the book” and that the BMS doesn’t disconnect them.
I’ve also had fuses blow due to thermal overload before. So far, no MPPT has ever given up the ghost because of it. However, that was in a 24 V system with a maximum module voltage of around 56 V.
So, unless you’ve completely pushed the permitted voltages in the system to their limits — for example, with a battery having a charging voltage of over 60 V — there should never be any problems in such a situation, even if the output voltage briefly rises by 5–10 V because the battery is missing.
Even if a switching regulator’s response speed is limited, it should still adjust itself within a few cycles! Otherwise, you couldn’t use these components for highly fluctuating loads at all! The capacitors just need to be large enough.
I also use the 123SmartBMS and am very happy with how it performs. It reduces the charging voltage in good time before a drifting cell reaches critical values. I have five MPPTs running in my system, and it has now been operating for more than five years. During that time, I’ve never had an unintentional shutdown. I haven’t installed a main switch in my home system, so it can’t become a problem either. I have installed one in my motorhome, though, but I’ve never used it.
As Dieter rightly pointed out, it also depends very much on the system and how much headroom the components have to withstand the overvoltage. This depends, on the one hand, on the selected battery voltage and, on the other, on the charging voltage the MPPT is designed for. MPPTs that can only charge 12 V and 24 V systems are more likely to have problems than those that can also charge 48 V systems. If the battery voltage is, for example, only 12 V, you have plenty of headroom before anything gets damaged.
The problem is that the MPPT can only push energy in one direction (PV → battery). On the battery side, that energy has to be consumed/absorbed. If that doesn’t happen, the voltage rises. In normal operation, energy is constantly being drawn, so the MPPT really only needs to make sure that it pushes in enough energy quickly enough. It doesn’t need to worry about taking energy back out.
You can almost compare it to a car driving up a hill. All you really need to do is press the accelerator. When you let off, the car slows down by itself. You never need the brakes. But if the hill suddenly flattened out, the car might have a problem if it couldn’t brake.
ideally, the MPPT contains a step-down converter with two MOSFETs controlled via pulse-width modulation and a low-pass filter. In that case, the output voltage mainly depends on the duty cycle, but it can never result in any significant overvoltage. It’s different when diodes are used, but then the losses are higher!
and i think victron uses the better method, so the system only needs to make minor adjustments when the load changes. however, it’s also possible that the second FET is only driven as a diode, but even then there shouldn’t be any danger because of the large capacitors.
however, i haven’t analysed this in that much detail yet!
no MPP has to be traversed from the left-hand side
a load is present
no significant current was flowing when the load was switched off
The PV input is followed by a MOSFET installed “the wrong way round”, whose reverse diode is intended to block current flowing from the battery to the PV system. It is switched on during normal operation so that the voltage drop across the Rds(on) is reduced to almost zero. If there is no load, the output voltage always rises to the level of the PV voltage as long as the input MOSFET is still switching on — even if the pulse width is small. The output is galvanically connected to the input through the MOSFET.
The output voltage can reach an even higher level, independently of the control loop, if the load is switched off while a higher current is flowing. The storage inductor then discharges its energy. If this energy cannot be absorbed by a component (the output capacitor), it continues to raise the voltage until something breaks down.
Let’s assume an MPPT 250/100 on a 12V system with a PV voltage of 220V and an output current of 100A. The BMS disconnects the charge because a Cell_overvoltage occurs.
To understand the effects, you first need to understand where this high output current comes from in the first place. Assuming 15V at the output, this would be a power of 1,500W. Solar modules at a module voltage of 220V can easily provide that. For example, by connecting five 440Wp modules in series. However, these would provide only 6.8A. This is because there is an input charging capacitor, which is charged by the modules. From this capacitor, the path leads through a switching transistor and a storage inductor directly to the battery connection, which also has a capacitor. The current is constantly switched on and off by the switching transistor. When the transistor switches on, both the module string and the charging capacitor supply current to the storage inductor. The current flowing in this process is considerably higher than the module current, because the capacitor could supply almost infinitely much current for a short time. However, the inductance of the storage inductor prevents an extremely rapid rise in current. While the transistor is conducting, the current rises in a sawtooth pattern by only a few percentage points. Until the transistor switches on, current flows only because of the energy stored magnetically in the storage inductor. After the transistor switches off in the preceding phase, it has fallen in a sawtooth pattern by the same amount. This current flow is enabled by the freewheeling diode located at the connection between the switching transistor and the storage inductor, which leads to ground. The on/off ratio of the switching transistor is approximately (module voltage − battery voltage)/battery voltage, i.e. about 1:6. The current is therefore supplied by the storage inductor for 6 time units and by the module charging capacitor for 1 time unit.
If the current is switched off on the output side, the energy in the storage inductor discharges into the capacitor located at the output. With an assumed output capacitor of 100 microfarads, its voltage increases by approximately 20V per microsecond. This may not destroy the MPPT, but it triggers a safety system with several protective stages for the battery, its BMS and the connected electronics. The MPPT receives no information, or no timely information, about the shutdown, and the developer had to assume that connected components would be at risk in such a situation. All the protective stages fail because they cannot prevent the storage inductor from discharging. There is only one exception: a short circuit at the PV input. In newer 250/100 units, this is carried out using a latching relay and prevents any further energy from being supplied. Nevertheless, connected components may already have been damaged. Timescales in the nanosecond range are sufficient for this.
you’ve got that slightly wrong. With two MOSFETs in a push-pull output stage, the converter operates in both directions. If the output voltage is above the target voltage, energy is transferred from the output back to the input. The amount of power transferred is irrelevant here. So it cannot result in an overvoltage either. The voltage may rise slightly because the control system adjusts the ratio to compensate for losses and resistances, but normally the output voltage always has the same ratio to the input voltage as the duty cycle!
It’s a different matter if the inductor is driven into saturation! Its inductance then decreases!
The small non-isolated Orion converters normally work like this. At least I can confirm it for the very small 24/12 5A one I have. If the output voltage rises above the nominal voltage, current flows mercilessly in the wrong direction, which is why I had to connect a diode behind it!
However, if the PV voltage rises by 10%, the output voltage would also rise by 10% if the control system did not intervene!
I think they probably use a MOSFET instead of the diode, because that results in higher power losses. Unfortunately, only Victron knows exactly what the circuit looks like and how the component is controlled!
And first you have to upload something or subscribe to something to view the schematic. It’s annoying!
Exactly, this concerns the reverse-current diode. It has either been replaced by a MOSFET or bypassed. Otherwise, a 50 A controller would get considerably hotter and need a fan.
I want to develop a protective mechanism so that the overvoltage is diverted.
The fact remains that the battery’s BMS can disconnect at any time. There’s nothing that can be done about that, since you don’t know which battery will be installed in the motorhome.
I’ve already tested a voltage-limiting circuit set to 16 V.
It takes the SmartSolar 850 ms to regulate back to the final charging voltage.
Can anyone confirm this?
That’s an awfully long time.
That can be done.
I’ve already tested it.
But it also shows how sluggish the voltage regulation is when returning to the charge termination voltage.
The lower you clamp the voltage, the longer it takes.
When clamped at 18 V, it takes 500 ms. When clamped at 16 V, 850 ms.
I’ve built a minimum voltage clamp.
15.9 V => 1 mA
16.6 V => 100 mA
16.9 V => 1000 mA
The voltage regulator cannot bring the voltage down; that isn’t possible with this type of regulator. It can only transfer energy from the PV side to the battery side, not the other way round.
You should simply connect a dummy load; the voltage will then drop more quickly.
It’s an MPP tracker. There is no separate current control.
Your circuit operates the MosFet in the analogue region. This makes the current–voltage transition gradual. The voltage is still dependent on the current and could become too high. An inverter could therefore go into overvoltage shutdown. This is particularly the case when lead-acid batteries are operated at low temperatures and larger load steps occur. To design such a circuit, you first have to determine which situations can arise. This also includes the possibility that the current may be higher than the value set on the controller. With the 75/15, it can briefly exceed 15 A. A LiFePO4 battery can go into cell overvoltage protection not only at a high SOC at the end of the charging process, but also at a lower SOC and with the full charging current. To test this, an auxiliary circuit is required that simulates a BMS shutdown using a timer, evaluation circuit or manual trigger: A Mosfet interrupts the negative lead of the LiFePO4 battery, which is not fully charged. It is driven by a Schmitt trigger. A low-pass filter is connected to its input. Upstream of the low-pass filter is the trigger signal that triggers the oscilloscope. Because of the low-pass filter, the Schmitt trigger responds with a delay and switches the MosFet off after a delay. This makes it possible to see a short section with normal voltage in the oscillogram, followed by the event. If the battery is charged via a shunt, a sudden increase in current can be detected and used to trigger an event. This would simulate an MPPT that outputs an overcurrent when an MPP is exceeded, causing a BMS shutdown due to cell overvoltage.
