Why do we need overcurrent protection.

[Craig]

I was reading someone's thread and they were saying they needed a BMS with over current protection on both the charge and discharge side. I know this is a common request but WHY? . Obviously you need proper equipment such as not putting a 10Kw inverter on a 12v system which would draw 833 amps. But if you have correct sized charger and correct sized inverter combined with properly sized circuit breakers. What are the use cases for overcurrent protection?

I'm not saying don't do it but is this a real problem?

 

[gnubie]

Many years ago an ex-Telecom Australia (so post PMG, before Telstra branding) recounted a story where he was working in the battery room of an exchange and managed to drop onto and bridge a battery with a wrench. He had to kick it off the battery and it melted the sole of his shoe in the process.

A breaker will protect you down stream from the battery but it'd sure be nice to have some protection at the terminals too given how much current a lithium battery can crank out.

 

[Craig]

Many years ago an ex-Telecom Australia (so post PMG, before Telstra branding) recounted a story where he was working in the battery room of an exchange and managed to drop onto and bridge a battery with a wrench. He had to kick it off the battery and it melted the sole of his shoe ie process.

A breaker will protect you down stream from the battery but it'd sure be nice to have some protection at the terminals too given how much current a lithium battery can crank out.

Fully admitting that I do not know everything about BMS's. I do know a lot about how the Chargery works and assume in principal they are mostly the same. The over-current protection on the Chargery is downstream from the battery as well so arking the terminals would not set off the over-current protection. It may however set off any cells HVD or LVD alarm. I think this because I used a rather large resistor at one time to passively balance a high cell. when I did this the next cell in line registered high fairly quickly which caused a cell related HVD. It seems if one dropped a wrench across the terminals of some of the inner cells it would be similar to connecting a cell to a resistor as it would pull down the voltages of the affected cells.

Hope I'm not being argumentative just trying for a discussion.

 

[gnubie]

Sorry, I assumed an internal BMS where the only access to the cells is via terminals on the battery case, home brew or commercial.

 

[Maast]

Since the NEC considers overcurrent devices to be intended as wiring protection and NOT appliance protection I would say its a safety device if something crowbars the battery output to keep from cooking the wiring and starting a fire and that (valid IMO) thought process has transferred over to the DC side. Since mosfets when they go bad they 95% of the time are a dead short until something in the current patch melts I thoroughly agree. Considering all the bits of metal on/around inverters it's easy to conceive of a scenario something other than the inverter has caused a dead short.

 

[Dzl]

I was reading someone's thread and they were saying they needed a BMS with over current protection on both the charge and discharge side. I know this is a common request but WHY? . Obviously you need proper equipment such as not putting a 10Kw inverter on a 12v system which would draw 833 amps. But if you have correct sized charger and correct sized inverter combined with properly sized circuit breakers. What are the use cases for overcurrent protection?

I'm not saying don't do it but is this a real problem?

Assuming I understand your question (you are referring to BMS overcurrent protection with DIY lifepo4?):
This is something I have pondered on and off since I first began thinking about BMSes, I don't have a lot of insight and I do still have some unanswered questions, but I also have a theory.

I feel that its possible that people are misinterpreting 'overcurrent protection' listed as a feature on at least some of the FET based chinese BMSes.
I have a suspicion that unlike over and under voltage protection, overcurrent protection is a form of BMS self protection first and foremost, not a battery pack protection and not a replacement for real OCP (fuse or breaker on positive side).

I think that system design and possibly a system coordination capable device (like a Victron GX device or maybe Samlex Evo which can regulate overall charge rate) is the proper way to make sure the capabilities of your battery bank are not exceeded and/or regulate charge rate, and a fuse or breaker is the proper solution for catastrophic OCP. And that the original role of BMS OCP was to protect the BMS itself (and the system, by preventing BMS failure, but not to directly protect the batteries).

For a drop-in lifepo4 battery with a warranty, I do suspect that the BMS OCP is also meant to protect the battery (or more accurately protect the company from excessive warranty claims due to the limitations of the battery being exceeded by all the people buying a single 100Ah battery and a 3 or 4 kw 12v inverter).

 

[Craig]

Yes I was referring to our home brew packs.

 

[Craig]

Assuming I understand your question (you are referring to BMS overcurrent protection with DIY lifepo4?):
This is something I have pondered on and off since I first began thinking about BMSes, I don't have a lot of insight and I do still have some unanswered questions, but I also have a theory.

I feel that its possible that people are misinterpreting 'overcurrent protection' listed as a feature on at least some of the FET based chinese BMSes.
I have a suspicion that unlike over and under voltage protection, overcurrent protection is a form of BMS self protection first and foremost, not a battery pack protection and not a replacement for real OCP (fuse or breaker on positive side).

I think that system design and possibly a system coordination capable device (like a Victron GX device or maybe Samlex Evo which can regulate overall charge rate) is the proper way to make sure the capabilities of your battery bank are not exceeded and/or regulate charge rate, and a fuse or breaker is the proper solution for catastrophic OCP. And that the original role of BMS OCP was to protect the BMS itself (and the system, by preventing BMS failure, but not to directly protect the batteries).

For a drop-in lifepo4 battery with a warranty, I do suspect that the BMS OCP is also meant to protect the battery (or more accurately protect the company from excessive warranty claims due to the limitations of the battery being exceeded by all the people buying a single 100Ah battery and a 3 or 4 kw 12v inverter).

I think you make a good point about it being something different than we think it is.

 

[Maast]

Just reread the OP and realized that I missed the "BMS" over current protection. I agree with Dzl that it's a BMS self-protection thing.

 

[Ampster]

Since mosfets when they go bad they 95% of the time are a dead short until something in the current patch melts

Reundancy has value, particularly in safety systems. I often wondered why I needed a fuse if I had the BMS that could cut off voltage. As @Maast reminded us, if the FETs fail in a closed state in a BMS we need the fuse or breaker as our last resort.

 

[GXMnow]

You can think of it a few different ways. Over Current Protection is there to reduce the chance of things failing when something goes wrong.
A fuse is an absolute must as they are the most reliable device to keep things from catching on fire. But you have to actually replace it to get it back running again, so I think of the fuse as the last resort protection. Next in line is having a resetable circuit breaker. This should be a lower current than the fuse so it should trip and give you a chance of clearing the fault and turning it back on. But a breaker might not be fast enough in there is a major high current short. The breaker does not need to be replaced, but you usually still need to go and turn it back on. Then there is the BMS. Depending on how it is programmed, the BMS can safely turn the power off to protect the system, and many can auto reset and turn back on. It is not required, but it can add convenience. Since a good one is actually measuring the current and can have a defined current limit, and a timer to allow some extra surge current for a specific time. This programmable nature can give early protection and prevent the other devices from tripping when it is a minor overload. That one time the microwave, A/C compressor, and the refrigerator all happened to start at the exact same time. The power get's interrupted, no wires melted, the fuse didn't blow, and after a 2 minute delay, it turns the power back on, and the restart delays are hopefully different enough that everything can start up again without damage. It may not be required for a safe system, but it sure beats having to go replace a fuse out in the solar shed.

 

[Dzl]

Related question, is it actually safe, reliable, and/or advisable to use FET's for overcurrent protection. I'm thinking medium to long term, and I don't have the necessary electronical knowledge to answer the question myself, but I have a perception of FET's being somewhat fragile/sensitive compared to other OCP options.

 

[toms]

Overcurrent is to protect the batteries.
If you have a 100ah battery that is rated at 1C discharge, you don’t want to exceed 100amps current. The BMS will prevent this.
If you allow 200amp current, the battery will provide this at a shortened lifespan.
As you said, if you size your loads and charge sources correctly - you don’t need over current protection.

 

[Craig]

You can think of it a few different ways. Over Current Protection is there to reduce the chance of things failing when something goes wrong.
A fuse is an absolute must as they are the most reliable device to keep things from catching on fire. But you have to actually replace it to get it back running again, so I think of the fuse as the last resort protection. Next in line is having a resetable circuit breaker. This should be a lower current than the fuse so it should trip and give you a chance of clearing the fault and turning it back on. But a breaker might not be fast enough in there is a major high current short. The breaker does not need to be replaced, but you usually still need to go and turn it back on. Then there is the BMS. Depending on how it is programmed, the BMS can safely turn the power off to protect the system, and many can auto reset and turn back on. It is not required, but it can add convenience. Since a good one is actually measuring the current and can have a defined current limit, and a timer to allow some extra surge current for a specific time. This programmable nature can give early protection and prevent the other devices from tripping when it is a minor overload. That one time the microwave, A/C compressor, and the refrigerator all happened to start at the exact same time. The power get's interrupted, no wires melted, the fuse didn't blow, and after a 2 minute delay, it turns the power back on, and the restart delays are hopefully different enough that everything can start up again without damage. It may not be required for a safe system, but it sure beats having to go replace a fuse out in the solar shed.

Just for the sake of discussion if your BMS has a delay wont the fuse or breaker trip before the delay times out? Again caveat is all equipment is sized properly

 

[GXMnow]

It depends on a lot of things. My battery is fused at 250 amps, My BMS will cut off after 3 seconds above 200 amps. breakers are actually quite slow unless the fault current is double the rating. Fuses have several different speeds. Even a fast blow fuse can still take more than a second to trip at 50% over current.

 

[BiduleOhm]

To answer the original question: BMS OCP is mainly to protect the BMS itself as fuses are too slow to protect the mosfets (by many orders of magnitude). The other reason is that if you design it right it can be so fast that you don't blow an expensive fuse and you vastly minimise damage where the short happens (less molten metal flying, less fire risks, etc...) as the current doesn't have time to rise too much (at those times frames you need to account for inductance of everything; mainly wires and cells).

Related question, is it actually safe, reliable, and/or advisable to use FET's for overcurrent protection. I'm thinking medium to long term, and I don't have the necessary electronical knowledge to answer the question myself, but I have a perception of FET's being somewhat fragile/sensitive compared to other OCP options.

It's actually more and more used in the aviation industry because you can have a clear warning when something goes wrong (usual breakers need to be checked one by one visually), there's no moving/mechanical parts (reliability, life span, weight, ...), you can use them as a switch (mechanical breakers don't like that too much), ...

So no, mosfets aren't fragile/sensitive if you design things properly. The most sensitive thing would be the gate but if it's protected with zener/TVS there's no problem at all. Next thing would be exceeding the drain current but if things are properly sized then that will not happen either

Just for the sake of discussion if your BMS has a delay wont the fuse or breaker trip before the delay times out? Again caveat is all equipment is sized properly

Usually there's two OCP: one allowing a mild overload for seconds/minutes (mainly to avoid mosfets heating to much), and one fast who acts like a fuse.

//

NB: with that said mosfets often fail shorted when they fail so being a high advocate of fail safe systems I highly recommend to have a fuse on your battery even if you have an electronic fuse (aka BMS OCP) downstream. They can use electronic fuses instead of classic fuses in the aviation industry because they are triply (or more) redundant systems and they have been tested extensively to very high standards (i.e. not like a cheap chinese BMS...)

 

[Zil]

If I had a short circuit with the LFP bank. I would rather replace a fuse than the well over stressed BMS.

 

[Cal]

Those BMS that have internal disconnects may have OCD to save their disconnect switch. Chargery on the other hand has no OCD. This system relies on fuses as a disconnect. Nothing wrong with that. Disconnect speed is not an issue.

I was wondering how the over current is detected. An ADC measuring shunt current would be too slow. It would be fairly simple to use an Op Amp to amplify the shunt voltage and then use a comparator to detect a specific voltage level. What would be an acceptable over current level? 150% of rated shunt?

 

[BiduleOhm]

I was wondering how the over current is detected. An ADC measuring shunt current would be too slow. It would be fairly simple to use an Op Amp to amplify the shunt voltage and then use a comparator to detect a specific voltage level. What would be an acceptable over current level? 150% of rated shunt?

Yep, far too slow (even with very nice mosfets you need to be single digit µs fast...). The op-amp/comparator is exactly the solution I chose on my design as it's the simplest and fastest solution I could find.

I used a bit over 400 A (430-ish A IIRC) for the hardware limit as the software limit will be 400 A for a 300 A continuous rating and a 500 A shunt. I'm actually contemplating upping the limit to near 500 to avoid nuisance trips as 400 is maybe a bit too conservative.

 

[Maast]

Even though I'm an electronics toad, I've found myself preferring the thermal breakers w/switch as string-level battery protection. The thermal nature means it can take a few seconds of overload without tripping and having the switch on it makes it easy to disconnect a string to do maintenance.
They can take such a HIGH overload for several seconds so it's very important to size them correctly or even a little smaller than you'd think.
I like the self-resetting thermal breakers too for certain applications.

 

[Cal]

I used a bit over 400 A (430-ish A IIRC) for the hardware limit as the software limit will be 400 A for a 300 A continuous rating and a 500 A shunt. I'm actually contemplating upping the limit to near 500 to avoid nuisance trips as 400 is maybe a bit too conservative.

I don't think you want the OCD too tight. Will the disconnect signal get latched? Or get reset after a certain time period? Got to watch out you don't have a high current oscillator. Perhaps add a disconnect counter?

 

[BiduleOhm]

Yeah that's why I'm thinking about setting it higher; pfff I can't stop making changes... ?

Yep, the hardware protections are latched by AND gates configured as special S-R latches and need a push of a button to be reset. I chose a manual reset because, excepted for the short-circuit one, they're all redundant to software ones so they should only trip if there's a problem with the battery and the software has a problem too.

The software ones will be latched in software and can be reset however I want as it's in software. Probably something like 1 or 2 retries after a 20 or 30 sec delay and then if it's still faulted I wait for a user reset to avoid cycling indefinitely. I can also have different strategies depending on the fault type (like no retry for the most important ones and infinite retries for the less important ones for example).