diy solar

diy solar

Can BMS damage MPPT Charger?

fafrd

Solar Wizard
Joined
Aug 11, 2020
Messages
4,188
I’ve read a few places that MPPT charge controllers can be damaged when a BMS goes into rapid shutdown (to protect the battery) and am trying to understand the issue:

During nighttime (solar off) if the BMS goes into Low Voltage Disconnect, the MPPT was not charging and so it should be fine, correct?

During daytime while MPPT is charging, if the battery reaches High Voltage Disconnect, the charge controller should be in Float so little/no current should be flowing so again, MPPT should be fine, correct?

During daytime while the MPPT Charger is charging, in general it shouldn’t be possible to trigger Low Voltage Disconnect while MPPT is charging and most inverters can be set to shut down based on SOC or low voltage before LVD is triggered, so it seems like the MPPT has little to worry about in that case as well, right?

So is it only in the case of a failing battery and one cell quickly rising up to trigger OverVoltage Disconnect before the overall battery is close to being charged and while significant charge currents are flowing out of the MPPT that damage can occur?

I’m struggling to understand the scenarios where BMS can damage MPPT, the mechanism whereby an MPPT charge controller can be damaged, and the easiest/cheapest design alternatives to protect against such damage.
 
I think the question you are asking is whether it is okay to disconnect the battery(s) from an SCC while panels are still connected.
This horse has been beaten plenty.

Any pointers (or search terms to use) would be appreciated.
 
 
As I'm not sure if all MPPT, so I will say most MPPT CC can be damaged if solar is connected before battery. Also if battery disconnected first. Mr Sandals is referring to PWM testing. PWM can take it. If you are troubled, see if your custom program on the CC can be set to disconnect the solar side on high voltage. Or use a voltage sensing relay to disconnect the panels before the BMS high voltage setting.
 
MPPT charge controller is a Buck converter:


It works by periodically closing a switch allowing current to flow from PV input (with capacitors), similar to PWM. But instead of shorting the input to battery like PWM, it applies PV voltage to an inductor. Current in an inductor can't change immediately, it ramps up or down over time (like inertia of a heavy weight.)

That keeps a steady current flowing into the battery. If you suddenly disconnect the battery, the current tries to keep flowing anyway and generates a high voltage which could damage something. That's how your ignition system produces a high voltage spark. In mechanical systems, that moving weight slams into something (like hammer stopping on a nail.)

There are transient voltage suppressors intended to protect circuits. A suitably sized one placed across the MPPT's battery terminals might protect it in the case of lithium batteries with disconnection by BMS. The TVS should have a voltage higher than battery would ever charge to, but low enough to protect the MPPT. A transistor circuit driving a suitable resistive load could also do the same, e.g. buy a shunt regulator and set it suitably, maybe a 60V regulator for a 48V battery. Use both, and the TVS can eat a short spike until the shunt regulator kicks in.
 
I highly doubt an mppt charger would be damaged by connecting panels only. All they are is a buck converter with some extra control logic for mppt and extra components to prevent the battery backfeeding the panels.
 
I highly doubt an mppt charger would be damaged by connecting panels only. All they are is a buck converter with some extra control logic for mppt and extra components to prevent the battery backfeeding the panels.

A couple of ways it could happen, depending on how conservatively designed/spec'd the charge controller and how careful user is about Voc:

1) Input capacitors see full PV voltage. But transistors only see Vpv - Vbat. If Vbat = 0V, transistors hold off more voltage.
2) MOSFET Vgs(max) is something like +/-20V. If MPPT hasn't seen Vbat and established suitable Vgs before PV is applied, the circuit could be pulled to a level that blows the MOSFET gate.

It depends on how the MPPT was designed and what negative testing was performed.
 
I was specifically told by Morningstar that MPPT can be damaged by disconnecting battery before the panels. Now, maybe it has to be while high charging is active, but damage is damage.
 
I was specifically told by Morningstar that MPPT can be damaged by disconnecting battery before the panels. Now, maybe it has to be while high charging is active, but damage is damage.
Read this thread.
 
MPPT charge controller is a Buck converter:


It works by periodically closing a switch allowing current to flow from PV input (with capacitors), similar to PWM. But instead of shorting the input to battery like PWM, it applies PV voltage to an inductor. Current in an inductor can't change immediately, it ramps up or down over time (like inertia of a heavy weight.)

That keeps a steady current flowing into the battery. If you suddenly disconnect the battery, the current tries to keep flowing anyway and generates a high voltage which could damage something. That's how your ignition system produces a high voltage spark. In mechanical systems, that moving weight slams into something (like hammer stopping on a nail.)

There are transient voltage suppressors intended to protect circuits. A suitably sized one placed across the MPPT's battery terminals might protect it in the case of lithium batteries with disconnection by BMS. The TVS should have a voltage higher than battery would ever charge to, but low enough to protect the MPPT. A transistor circuit driving a suitable resistive load could also do the same, e.g. buy a shunt regulator and set it suitably, maybe a 60V regulator for a 48V battery. Use both, and the TVS can eat a short spike until the shunt regulator kicks in.

I appreciate these inputs and the suggestions.

And to be clear, this is the specific case I am asking about. ‘Connecting solar panels before battery’ is a different scenario and one I am not concerned about’. It is the scenario that BMS abruptly and unexpectedly disconnects the battery in the middle of a high-current charge cycle that I am asking about.

Your explanation has helped me to understand the potential concern - inductor-induced voltage spikes.

So my first follow-on question is woukdn’t this be easy to measure with an oscilloscope? Hook it up to capture transient above a threshold like 60V then disconnect battery in the middle of a charge cycle.

Has anyone ‘looked for’ such battery-disconnect-induced voltage voltage spikes?

My second question is that the protection fix/circuitry you have outlined is so trivial, wouldn’t any half-decent MPPT charge controller have that built in?

And my third question is, since most low-end MPPT charge controllers limit battery voltage to about 50% of MPPT voltage, the Buck Converters seem to be limited in voltage conversion range - does this provide any upper-bound on the magnitude of a potential inductor-induced voltage spike as you’ve outlined?

What I mean is, if you have an MPPT controller rated for up to 48V battery but you are only using it to charge a 24V battery, does that inherently provide some additional ‘margin’ as far as a voltage spike from charging at 24V being less likely to damage transistors designed to charge batteries up to 48V?
 
If the MPPT had no capacitance on its output, to the first order, voltage spike when disconnecting battery would be infinite volts (not some ratio of PV to battery voltage). Inductor value and capacitance value define how high voltage will go for a given current.
Practically speaking, there will be parasitic capacitance and losses that limit it, so voltage headroom will provide some margin.

Good chance MPPT started back in lead-acid days, with no BMS disconnecting it.
Advice working on cars is, 'do not disconnect battery to "test" alternator; that will kill it.'
The designer would have to think of this scenario, design to address it, and test it (with margin) to deliver a reliable product.
As you are aware, the leading aircraft manufacturer in the world failed to do that for a "manned platform."

I would hook up a TVS before testing with an oscilloscope. My oscilloscope is worth more than my charge controller.
Actually, I was just testing my "new" TVS box yesterday. Tried to use a HiPot, but the large capacitance (and possibly indicator lamps) tripped it around 20 Vrms. So I instead used the capacitor & diode from a microwave to create 600 VDC and threw a switch to connect the TVS. All phases clamped right at 200V according to my scope.
 
Read this thread.
Yes, I read that post. Read all of it. Not just the part where they could not destroy a PWM CC. I only know what Morninstar told me about my MPPT only being used for charging, not used as power supply to a load. I want to see Will do the series of tests on MPPT. I use the custom set high voltage disconnect of the MPPT. It switches the solar panels off when my battery is 13.8 volt.
 
The only time I can see a bms shutting down during high charging is due to thermal shutdowns. High or low temps could cause battery disconnect at most any charging level. This was my concern which lead me to this thread. That being said, my cc has been disconnected during charging on occasion. I have an old rover cc which has a bad terminal, so I replaced it in my main system. Now it runs my shop system, and I’ve moved the charge wires from one battery to another without issues, or tripped over them? occasionally. It does not seem to have been damaged in any way. That being said, I do wonder if I should get a cc that features a diversion load switch.
 
I looked into this with my Renogy Rover, because I wanted to install a lifepo4. The Rover is specifically designed for lithium. I can’t find it now, but I did find a thread that covered this. It said, essentially, that you should not connect panels to cc without battery, because that can create issues due to fluctuations in voltage, However, Disconnecting a battery from the charge controller does not cause the same issues. Now, as I say, this cc is designed for ‘lithium’, and so I suppose they took this into account when building it. I unplug my batteries frequently. I had a situation where all my batteries were dying, and I was having to ‘hot swap’ several through my two CCs to charge them all separately. Now I have a new 100ah lifepo4, so, hopefully, that is all behind me. I’m gonna get a 280ah prismatic bank ASAP, as well.
 
I looked into this with my Renogy Rover, because I wanted to install a lifepo4. The Rover is specifically designed for lithium. I can’t find it now, but I did find a thread that covered this. It said, essentially, that you should not connect panels to cc without battery, because that can create issues due to fluctuations in voltage, However, Disconnecting a battery from the charge controller does not cause the same issues. Now, as I say, this cc is designed for ‘lithium’, and so I suppose they took this into account when building it. I unplug my batteries frequently. I had a situation where all my batteries were dying, and I was having to ‘hot swap’ several through my two CCs to charge them all separately. Now I have a new 100ah lifepo4, so, hopefully, that is all behind me. I’m gonna get a 280ah prismatic bank ASAP, as well.
To clarify, the reason to not connect panels before battery is that the inconsistent voltage coming from the panels can cause the cc to incorrectly identify the system voltage. It could decide that it is connected to a 24v system, when the battery is 12v or vice versa. Once the cc is powered, and has properly identified system voltage, this is no longer a concern.
 
Back
Top