The BMS Was Supposed to Protect my Batteries?
- Thread starter Maine Sail
- Start date
Great video covering an important safety consideration that is easily overlooked. Your deep dive and recommendations are spot-on.
DC voltage limiter is needed on BMS inputs because their engineers don't understand the concept of more is better users?? or is it only RTFI as usual?
Until the contactor welds or the control circuit gets fried by the over-voltage event...
The whole point is that no BMS can protect against all forms of abuse. The responsibility is on the *system* designer to consider failure modes, effects, and mitigations to achieve an acceptable probability and severity of harm. In this case: if you're going to use a 100V solar array, you need something that can interrupt 100V - otherwise a single fault can cascade and bring down your whole system.
The topic here was a solar installation that with a controller failure could provide voltages higher than the BMS MOSFET specs. There is no failure mode from an external BMS that would be provoked here. The solar can't provide the >1,000A (sustained) required for the contactors to weld shut (and the voltage is irrelevant), and the control circuits are independent of voltage input to the point where they will survive long enough to shut the battery down before being damaged themselves.
I do agree that the system needs to be designed properly, but in this case, a 100V solar array is only two panels connected in series, most of the solar controllers for this size panel are rated for 150V, and nobody knows the ratings of the MOSFETs inside any drop in battery. It is not possible to design a system using reasonable solar and drop in batteries, other than to intentionally ham-string it hoping to keep everything working together if something fails.
An external BMS and contactors are about as robust a system as can be designed, with no compromises, and allows one to take advantage of all aspects of their electrical system.
I dislike MOSFET switches in critical power systems like a BMS. Even in an inverter, I prefer a line-frequency type over a high-frequency one because the necessary MOSFET's in them are less stressed. They are fine in solar controllers because those are rated for the voltages stated, and are usually embedded in giant heat sinks in open air. But I do not think MOSFET's should be protecting the battery. Nobody with these BMS's even knows if their MOSFETS are shorted or not until they tell the BMS to shut down and see if it does. Then when they turn it back on, they again have no idea if they are shorted now. You only find out a MOSFET has failed in your BMS when you find your battery ruined (as this case proves). Otherwise, it is all hope.
Mark
I do agree that the system needs to be designed properly, but in this case, a 100V solar array is only two panels connected in series, most of the solar controllers for this size panel are rated for 150V, and nobody knows the ratings of the MOSFETs inside any drop in battery. It is not possible to design a system using reasonable solar and drop in batteries, other than to intentionally ham-string it hoping to keep everything working together if something fails.
An external BMS and contactors are about as robust a system as can be designed, with no compromises, and allows one to take advantage of all aspects of their electrical system.
I dislike MOSFET switches in critical power systems like a BMS. Even in an inverter, I prefer a line-frequency type over a high-frequency one because the necessary MOSFET's in them are less stressed. They are fine in solar controllers because those are rated for the voltages stated, and are usually embedded in giant heat sinks in open air. But I do not think MOSFET's should be protecting the battery. Nobody with these BMS's even knows if their MOSFETS are shorted or not until they tell the BMS to shut down and see if it does. Then when they turn it back on, they again have no idea if they are shorted now. You only find out a MOSFET has failed in your BMS when you find your battery ruined (as this case proves). Otherwise, it is all hope.
Mark
It is 100v. Solar panels put out high voltage low amp current. When panels are placed in series the voltages are added, so a 300w panel may have an output of about 50v at 6 amps, place 2 in series and the combined output is 100v at 6 amps. Higher voltages are used so that smaller gauge wire can be used. Panels are placed in series to get the higher voltage and to have one controller control multiple panels instead of one controller for each panel.
We could, but our computers don't run very well on whale oil.
None of this makes much sense, until it does. Then you wonder what the big deal was, it's so simple.
On the lines of what colemj is pointing out, if you take a solar panel configuration with a 100 volt open circuit voltage and connect it to a 12 volt battery, it does not apply 100 V to the battery. The solar panel output is current limited and the voltage instantly drops to the battery voltage and dumps whatever current the solar panel puts out into the battery but this will never exceed the short circuit current of the solar panel.
If the converter failed as a short circuit, this could have been that there was no longer any charge regulation and the battery heated because it was overcharged. Should the BMS have prevented over charging?
If the converter failed as a short circuit, this could have been that there was no longer any charge regulation and the battery heated because it was overcharged. Should the BMS have prevented over charging?
With a lead acid battery the internal resistance of the battery is high enough to reduce the voltage a bit. Applying 120v to a 12v battery will cause the battery to rapidly over heat and expell hydrogen gas which is explosive. This will happen at fairly low voltages, around 15 to 16vdc.
However, a LFP battery has very low internal resistance and will not cause the voltage to drop significantly. The BMS can mitigate the high voltage and shut off charging. In the case Rod explains, the controller failed and the full 100v went to the battery's BMS, the MOFSETs couldn't handle that much voltage and failed allowing the battery to experience an extreme over voltage condition. The MOFSETs were rated at 60v (if I remember correctly, it's been a week since I watched the video), the 100v fried them.
That's only true until the battery reaches full charge and the BMS tries to interrupt the current. When the switch opens, the voltage rises to 100V instantly.
An excellent point. A true high reliability system needs to incorporate mechanisms to *detect* when a fault occurs, otherwise redundancy is defeated.
I am not arguing against such a solution. But Rodd's point still applies: you need to select / design said BMS to ensure it can withstand the highest voltage that may be present in your system - and interrupt the highest current while you're at it. I don't know about "no" compromises - coil current is a big one - but I do tend to trust mechanical disconnects a bit more than solid state.
Also, to anyone thinking of trying to use one to protect a battery with an internal BMS, note that the contactor alone is unlikely to be fast enough to prevent the BMS from being damaged by the 100V transient.
That's an overly broad and misleading statement. A MOSFET's stress level is a function of how the circuit was designed, not how fast it's switching. Thermal, environmental, and mechanical stress are all factors as well as electrical. The drive inverter in my EV has been flawlessly switching tens and hundreds of kW at 350V and >20,000 cycles/second for 13 years and 4000 operating hours while subject to all the extremes that come with an automotive environment in New England.
However, reliable solid state load switches (like a BMS) are one of the hardest design challenges in power electronics. Inrush currents, capacitive loads, and short circuits can generate enormous thermal stresses and most power MOSFETs are particularly bad at handling them. I would therefore be wary of trusting any BMS to do much heavy lifting - be it a budget Chinese battery or a supposedly premium US brand - without objective evidence of its capability.
Lets use an example of a 400 watt solar panel that is somehow 100 volts open circuit. Maximum current out of this 400 watt panel with a 100 volts open circuit might be about 5 amps (maximum power point voltage of 80 volts and 5 amps) The short circuit amp of this panel might also be about 5 amps.
When you apply that solar panel directly to a battery,. it can only put 5 amps into the battery. No more than this. What happens to the battery voltage if you put in 5 amps. Maybe it goes up .1 volts, maybe .3 volts.
Remember that the solar panel 100 volts is open circuit with no load or zero current. This is way different than a 100 volt power supply that can attempt to put out 50 amps at 100 volts.
In our case above, when the 400 watt solar panel example was put directly on the battery, the battery voltage might only increase by a few hundred millivolts.
This is how solar panels work.. They are basically a current source with an open circuit voltage limit.
FYI, google what the internal impedance of a 100 amp hour LFP battery is. I got 2 to 4 milliohm.
At 4 milliohm, that 400 watt panel with 5 amp short circuit current only raised the battery voltage by .02 volts. That will not cause a problem.
However, a 400 watt solar panel connected to a battery with no regulation (ie, the failed MPPT controller looked like a short between in and out) will cause the battery to over heat and damage itself if its left on for long enough. Most likley the battery cooked because there was no regulation on the charge long after the battery was full. The mystery is what is the BMS supposed to do if the battery is being overcharged because regulation was damaged.
You cant think of a solar panel as a fixed voltage source since its not even close to that.
I disagree with this, but I wasn't even thinking of switching frequency in terms of a unit's robustness. I was thinking of the large iron transformer in a line-frequency unit that can absorb a lot of stuff that can damage MOSFETs. Surges, motor LRC, power factor mismatches, etc. The tiny transformers in high frequency units are little help there at all.
But switching frequency absolutely has an effect on MOSFET because switching creates thermal and electrical stress on them. A MOSFET switching at high frequency is less robust than one switching at low frequency.
Mark
This is correct as far as it goes. However, what you are missing is that not everyone has a single panel (100V 400W as in your example). Lots of people have multiple panels with multiple controllers, and even other charge sources running at the same time.
While your example panel with a failed charge controller doesn't do any damage initially, as soon as the other panels and/or charge sources bring the battery to full charge, then that damaged part of the system rapidly comes into play. A LFP battery at 14.6V can be brought to 100V with 5A pretty quickly.
Mark
Yes.. The point earlier is that a solar panel might have an open circuit (no load ) voltage of 100 volts but it will not put out more current than its short circuit spec which is usually close to the maximum power point current. A solar panel looks like a constant current source over much of its operating voltage range and when connected to something like a battery that is below the open circuit voltage, the panel just injects near is short circuit current into the battery and this generally will only result in an initial voltage rise of the panel current times the internal impedance of the battery. Lets even go to an 800 watt panel (it was on an RV) with an open circuit voltage of 100 volts and a short circuit current of 10 amps. 4 milliohm battery impedance results in the initial battery voltage increase of a small 40 mv. Not damaging.
But now the charge controller is damaged so that 10 amps just keeps getting put into the battery with no regulation so the battery voltage rises because its getting charged. Pretty common knowledge that you can damage a battery putting an unregulated solar panel on it.
But as high current pointed out, after some time of unregulated charging the BMS will trip and disconnect the battery from the panel. Now the panel has no load and the voltage jumps up to the open circuit spec of the the solar panel.
Here is a lithium battery spec from Victon 7. Technical data
Maybe I missed it but I do not see anywhere in this spec a mention of the voltage the battery input can take when the BMS is tripped. If your trying to design a system taking this into account, you would need to know this number. Without this number, you cant design a system. You can speculate that its some common breakdown voltage of junctions of a device such as a fet but a zillion variables could make that a bad speculation.
Victron didnt give a number for max battery input voltage when the BMS is tripped.. If this is important, why isnt it specified. If you damage a charge controller so its no longer regulating and leave it connected to a battery, your likely to damage that battery for a wide range of panel open circuit voltages.