Single vs. Multiple Class-T Fuses

[740GLE]

It’s not about being “first” is about being the most limiting ?

Heck a device could have 12awg connected to a 200a breaker/fuse, that 12awg will act like a fuse before a break/fuse operates.

 

[cs1234]

It’s not about being “first” is about being the most limiting ?

Heck a device could have 12awg connected to a 200a breaker/fuse, that 12awg will act like a fuse before a break/fuse operates.

I did say the other device would have to be rated under whatever the built in is. But, if the built in isn't spec'd for the circuit it's in, can we really trust it to operate at a certain spec?

 

[740GLE]

I wouldn’t trust it for OCP.

It’s built in and spec’d to be just a disconnect, it just happens to have AC OCP jiberesh writings on it that don’t apply to it’s current use.

Most other batteries don’t in include any disconnects (smaller 12v 100ah kind)

Seems like less confusion all around if there were no markings on the disconnect which would prevent a false sense of security.

 

[cs1234]

I wouldn’t trust it for OCP.

It’s built in and spec’d to be just a disconnect, it just happens to have AC OCP jiberesh writings on it that don’t apply to it’s current use.

Most other batteries don’t in include any disconnects (smaller 12v 100ah kind)

Seems like less confusion all around if there were no markings on the disconnect which would prevent a false sense of security.

I would rather have no disconnect / switch on the battery than one I'm not absolutely certain of.

Even a switch (no OCP) has to be rated for the voltage and current it's breaking in order to not damage the switch / arc. If you say it's rated for that, I'll take your word for it, as I don't own one of the batteries in question and am not currently in the market for one.

 

[740GLE]

The EG4 LP4 BMS is very sensitive to overcurrent...almost to sensitive.

Overcurrent as a in normal load? Or system faults?

Breaking 100a isnt quite the the same as trying to break +1000a

Ideally if the BMS shouldn’t be the first element to operate to break high current, those FETs can only take so much, or do these units use a contactor.

 

[cods4]

Overcurrent as a in normal load? Or system faults?

Breaking 100a isnt quite the the same as trying to break +1000a

Ideally if the BMS shouldn’t be the first element to operate to break high current, those FETs can only take so much, or do these units use a contactor.

In a genuine bolted fault, the BMS is quite possibly going to be toast anyway. Overcurrent protection such as a Class-T fuse could take as long as 100-200ms to operate, in which time a fairly substantial amount of energy would have dissipated into the FETs, turning them into popcorn.

But at least when the fuse trips, it is going to stop the BMS, conductors and possibly the cells from turning into a fiery mess.

 

[cods4]

Considering it's the first fault protection device in the circuit (being integrated in the battery), you would have to use something else rated for something under what the built in breaker is rated for. Otherwise it gets to be first at bat. Seems like a bad idea to even have it in the circuit.

Ideally yes, but that won't always be possible because you could be limiting the use of your battery.
In the case of a fault, the AC rated breaker is quite possibly going to try and trip, but fail and fuse shut, soon after which your class-T fuse (even if it is rated higher than the breaker) should trip before any damage is done to the cables or cells. Following the fault, the AC rated breaker will either have to be replaced or bypassed to get back online.

 

[cods4]

@Will Prowse Can I make a video request please?

I would really like to see some concrete information about what the prospective fault current of a LiFePO4 battery is. There is hardly anything I can find about it on the internet. And what I have found is often contradictory.
I understand it could be quite limited by the speed of the chemical reaction in the cells. I.e <10C, or could be as high as 20-30C.

This sort of information would be extremely helpful for us DIYers and could improve the safety of many systems out there.

I have some ideas on how you could do an experiment while keeping it relatively safe and avoiding a total battery failure.

 

[TorC]

I'm of the opinion that a breaker should not be used as a switch. NEC allows it (NEC 2011: 404.11 Circuit Breakers as Switches), and there are provisions for it in IEC, but I still don't like it and discourage the application of this unless you really don't have another choice. A breaker should be a breaker; if it doesn't suit that purpose, it gives a false sense of security.

Agreed on a breaker that's only useful for a switch. I wouldn't deliberately design one in.

I do have various devices where the breaker, sized and designed as an overcurrent fault device, is also the occasionally used service disconnect switch, with no other switch. I think this is what the NEC is referring to, and I don't see a problem with it for things like hot water heaters where no one turns them off except to service the unit.

 

[TorC]

@cods4 There isn't anything concrete, and installed fault current is apt to be highly dependent on local installation conditions. The cells themselves have a calculated dead short fault current on the order of 20kA. More cells in series does not, in and of itself, change this, as each time you increase the total IR, you proportionally increase the voltage.

As a total system, the exact busbar connection resistance, wire gauge to the fault, how long said wire is, and all of that has a significant effect on actual fault current. In short, AFAIK, we don't know, but we have a reasonable range:

Greater than ANL or MEGA fuse rating, but less than Class T fuse rating.

 

[740GLE]

@Will Prowse Can I make a video request please?

I would really like to see some concrete information about what the prospective fault current of a LiFePO4 battery is. There is hardly anything I can find about it on the internet. And what I have found is often contradictory.
I understand it could be quite limited by the speed of the chemical reaction in the cells. I.e <10C, or could be as high as 20-30C.

This sort of information would be extremely helpful for us DIYers and could improve the safety of many systems out there.

I have some ideas on how you could do an experiment while keeping it relatively safe and avoiding a total battery failure.

Take some 4/0 and short it across a 100ah cell with a solid knife switch (with a rope
attached to it) see if the conductor will glow and melts before the cell expands.

It’ll be your version of a rector SCRAM ?

 

[John Frum]

More cells in series does not, in and of itself, change this, as each time you increase the total IR, you proportionally increase the voltage.

It does actually.
Its the interconnects.
For a 40mm2 busbar each interconnect is ~.00017 ohms.
That assumes good joinery.
It could be much higher.

 

[TorC]

It does actually.
Its the interconnects.
For a 40mm2 busbar each interconnect is ~.00017 ohms

Technically true, but for our purposes here it's false. Yes, the resistance goes up, but each .17mOhm is matched by 3.2V from the cell: .17mOhm busbar +.19mOhm cell IR = .36mOhm/3.2V. Doesn't matter whether you've got one cell or one hundred cells. The fault current through that is the same. Total power in a fault, however, goes up.

 

[Will Prowse]

@Will Prowse Can I make a video request please?

I would really like to see some concrete information about what the prospective fault current of a LiFePO4 battery is. There is hardly anything I can find about it on the internet. And what I have found is often contradictory.
I understand it could be quite limited by the speed of the chemical reaction in the cells. I.e <10C, or could be as high as 20-30C.

This sort of information would be extremely helpful for us DIYers and could improve the safety of many systems out there.

I have some ideas on how you could do an experiment while keeping it relatively safe and avoiding a total battery failure.

It's dependent on size of cell, internal cell temperature, terminal, cell type and a few other factors. And it's massive. I would stick with t class fuses regardless. Just makes sense.

Usually across the BMS is the highest amount of resistance on a battery. Easy to measure. This limits current in dead short event. If you have a shorted FET, that single fet will take all the current and resist massive amounts of current from flowing.

 

[John Frum]

Technically true, but for our purposes here it's false. Yes, the resistance goes up, but each .17mOhm is matched by 3.2V from the cell: .17mOhm busbar +.19mOhm cell IR = .36mOhm/3.2V. Doesn't matter whether you've got one cell or one hundred cells. The fault current through that is the same. Total power in a fault, however, goes up.

Have a look at the math.
Rename the text file to .ods
If you spot a mistake let me know.

 

[cods4]

@cods4 There isn't anything concrete, and installed fault current is apt to be highly dependent on local installation conditions. The cells themselves have a calculated dead short fault current on the order of 20kA. More cells in series does not, in and of itself, change this, as each time you increase the total IR, you proportionally increase the voltage.

As a total system, the exact busbar connection resistance, wire gauge to the fault, how long said wire is, and all of that has a significant effect on actual fault current. In short, AFAIK, we don't know, but we have a reasonable range:

Greater than ANL or MEGA fuse rating, but less than Class T fuse rating.

Theoretically it could be 20kA. But as I said, the chemistry of the cell supposedly limits this by quite a lot. It could be far far lower, which is also important to know as you need enough current from your cells to trip your selected fuses in a suitable time.

Take some 4/0 and short it across a 100ah cell with a solid knife switch (with a rope
attached to it) see if the conductor will glow and melts before the cell expands.

I was thinking a 1000A slow blow fuse and a DIY shunt, with an oscilloscope to monitor the voltage across the shunt, and the voltage across the cells. This would tell us if the cell voltage drops right off before the current gets very high.

 

[cods4]

It's dependent on size of cell, internal cell temperature, terminal, cell type and a few other factors. And it's massive. I would stick with t class fuses regardless. Just makes sense.

Usually across the BMS is the highest amount of resistance on a battery. Easy to measure. This limits current in dead short event. If you have a shorted FET, that single fet will take all the current and resist massive amounts of current from flowing.

Agreed, there will be some variability, but it could give us an idea. I.e if it's north of 10kA for a 300Ah cell then it's going to be more than 3kA for a 100A cell etc.

But if it's <3kA for a 300aH cell then that would make for some very different design decisions.

Personally I use Gg class HRC fuses as they are much cheaper and easier to get here in New Zealand. They have similar interrupt capacity to Class T fuses, but have a slower blow time, which is one of the reasons I want to make sure the fault current isn't too low.

Good point about the BMS. In many cases this would act as the fuse as it will either trip or fail open in a fault. But it is obviously not designed for this so can't really be considered as an over-current protective device.

 

[TorC]

Have a look at the math.
Rename the text file to .ods
If you spot a mistake let me know.

OK, so going from four cells in series to 32, which is rather higher than most would use, drops fault current about 11%. Looks like a pretty minimal variation to me. Yes, from cell alone to including busbars makes a significant difference, but not enough to change fuse classes. Fuses are rated to interrupt on current up to their rated voltage. Yes, there is a curve with voltage, but it's small, and not one you can safely rely upon.

 

[John Frum]

OK, so going from four cells in series to 32, which is rather higher than most would use, drops fault current about 11%. Looks like a pretty minimal variation to me. Yes, from cell alone to including busbars makes a significant difference, but not enough to change fuse classes. Fuses are rated to interrupt on current up to their rated voltage. Yes, there is a curve with voltage, but it's small, and not one you can safely rely upon.

My point was, the dead short amperage goes down when you go from 0 to 3 interconnects as in a single cell to a battery.
If you consider some ancillaries like a BMS, an MRBF fuse very likely has sufficient breaking capacity for a simple 4s battery.

 

[Shimmy]

Personally I think of the Class-T fuse as being short circuit protection

I'm of the opinion that a breaker should not be used as a switch. NEC allows it (NEC 2011: 404.11 Circuit Breakers as Switches), and there are provisions for it in IEC, but I still don't like it and discourage the application of this unless you really don't have another choice. A breaker should be a breaker; if it doesn't suit that purpose, it gives a false sense of security.

One benefit of the circuit breaker (as functionally a switch) is that the connection cables have an OCPD now. I don't think the BMS qualifies as a overcurrent current protective device.

 

[740GLE]

Don’t think anyone designs a system counting on a BMS as a OCPD.

Also if the AC rated breaker has no DC rating, can you really claim it as an OCPD protecting the cells or the connection cables?

For all we know the BMS will fry before that AC rated breaker operates.

I think a test of these units would be good, 300-1000 amp external fuses and dead short, see what it would take to operate that AC breaker.

 

[Mithril]

In terms of placement here is my 2c:
All battery banks should have A T-class or similar placed as close to to the busbar or wherever positive is connected. Having an unfused positive wire going from a battery bank is risking a short due to a damaged cable and is a bit like putting a AC breaker by the outlet and not in the breaker box.
Fusing individual batteries within a bank is likely a good idea, required AIC rating left as exercise for the reader.
Fusing and/or breakers should be at every place where wiring is going down in size (respective to the batteries, or main bus of the system), fuses/breakers should be no larger than allowed for the gauge of wire, but can be smaller. Since these are "after" the T-class or similar fuses they shouldn't need the same level of AIC, so long as they can handle whatever wouldn't blow an upstream fuse.
If you also want t-class for each AIO/Inverter thats fine, but they do NOT replace the need for fuses per battery bank.

 

[AHTrimble]

In terms of placement here is my 2c:
All battery banks should have A T-class or similar placed as close to to the busbar or wherever positive is connected. Having an unfused positive wire going from a battery bank is risking a short due to a damaged cable and is a bit like putting a AC breaker by the outlet and not in the breaker box.
Fusing individual batteries within a bank is likely a good idea, required AIC rating left as exercise for the reader.
Fusing and/or breakers should be at every place where wiring is going down in size (respective to the batteries, or main bus of the system), fuses/breakers should be no larger than allowed for the gauge of wire, but can be smaller. Since these are "after" the T-class or similar fuses they shouldn't need the same level of AIC, so long as they can handle whatever wouldn't blow an upstream fuse.
If you also want t-class for each AIO/Inverter thats fine, but they do NOT replace the need for fuses per battery bank.

Excellent!
And thank you for going back to the original question/post.

 

[Zapper77]

Don’t think anyone designs a system counting on a BMS as a OCPD.

Also if the AC rated breaker has no DC rating, can you really claim it as an OCPD protecting the cells or the connection cables?

For all we know the BMS will fry before that AC rated breaker operates.

I think a test of these units would be good, 300-1000 amp external fuses and dead short, see what it would take to operate that AC breaker.

Be cheaper to remove the breaker and test it. No need to destroy the battery or bms.

 

[cs1234]

Be cheaper to remove the fuse and test it. No need to destroy the battery or bms.

Spoiler: Pointless meme.