LFP battery - fuse test

[SunWatt]

I've read many posts about battery fuses but didn't find much on how to safely test if the fuse will work if there's a fault in a 48V system. So I thought it's best to ask here.

In your opinion, what would be the best way to test the battery fuse and what are the risks?

 

[LakeHouse]

I don't think a lot of people test their battery fuses, and for good reason; there's no non-destructive way to test them. If the fuse passes the test, it's destroyed (and so you still need to install an untested fuse), and if it fails the test, it's very probable that something else (battery, BMS, a switch, some wiring, anything nearby that happens to be flammable, etc.) would be destroyed or damaged.
My advice would be to buy a quality fuse from a reputable supplier.

 

[seneysolar]

My advice would be to buy a quality fuse from a reputable supplier.

100% this!

Makes me cringe when I see folks getting the super cheap breakers and fuses on amazon.

 

[SunWatt]

Thanks for chiming in.

Yes, buying a quality fuse from a reputable supplier is the safest bet. However, at the very least I'd like to verify and know for sure that a certain fuse type will work in case of a fault in the system.

Agreed. You've made a valid point that even if the test is successful, by placing a new fuse that is untested there is no guarantee it will work as the one used in the test. But at least I'll know this particular fuse is suitable for the application.

It would suck if the test would end up damaging the BMS or the battery cells, but the BMS should be able to disconnect the battery in case of a short circuit, right?

If anything, to begin with, I could at least test the fuse AIC with a single 3.2V LFP cell, and if it doesn't work at least I've limited the damage to a single LFP cell.

IIRC V/R=I, meaning 3.2/1.8miliohm = 2064A which is way above the fuse rating and should blow the fuse.

The part I'm having difficulty understanding currently is how big of a difference voltage makes in testing if the fuse will manage to break the circuit in a fault event.

LE: i.e. is it necessary to test the fuse with a 48V battery as well just to be certain the fuse will work?

 

[hwy17]

DC Breakers - Those could use a lot of testing. I don't trust them either way. You'll never see a breaker in an EV or grid scale ESS system.

Littelfuse, Blue Sea, or other name brand brand Class T - Trustworthy and proven, no need to test. But look up the trip curve and understand it, it's slower than you might think.

ANL and other non-Class T fuses - There are tests of these on youtube, and despite community concerns, I haven't seen them fail to break an arc.

 

[seneysolar]

A reputable fuse manufacturer will do the testing for you.

look for the published ratings and trust it. Eaton/Busmann knows what they are doing.

 

[LakeHouse]

If anything, to begin with, I could at least test the fuse AIC with a single 3.2V LFP cell, and if it doesn't work at least I've limited the damage to a single LFP cell.

I still wouldn't recommend it. But if you must: dress for welding. Welding mask for your head, and thick leather covering everything else.
And do it outdoors on a surface that isn't flammable.

 

[SunWatt]

Thanks for chiming in guys, truly appreciated.

I get a sense that in general there's not much interest in this topic here, so I wonder if there's anywhere else I could look for guidance on more technical topics like this? Some forums, groups, etc.

 

[seneysolar]

How do you plan on measuring this test?

Seems like anything other than touching 2 cables together while hooked to a battery and seeing if it blows would require some very precise and expensive equipment. Especially when you get into the 100A and above. Fuses and breakers all have a load/time curve rating anywhere from fraction of a second to over a minute.
I have seen guys using a power supply to test small breakers, expecting an instant trip at rated amperage. Most are disappointed that it takes much higher current trip it, but not factoring in the time it carries it. Plus turning the dial on a power supply while holding a stopwatch isnt going to suffice. My point is, if you are on an information quest, read and understand mfg ratings, select the style and size for the performance you expect.

 

[yabert]

IIRC V/R=I, meaning 3.2/1.8miliohm = 2064A which is way above the fuse rating and should blow the fuse.

The part I'm having difficulty understanding currently is how big of a difference voltage makes in testing if the fuse will manage to break the circuit in a fault event.

It's why you need fuse with rated current interrupting rating.
This thread will help you understand: https://diysolarforum.com/threads/eaton-amx-and-amh-fuses-alternative-to-class-t.60804/

 

[SparkyJJO]

I saw a video where someone inadvertently tested their Class T on a DIY 48V bank. BANG. It worked. Still messed up the dropped terminal but no other damage.

 

[SunWatt]

It's why you need fuse with rated current interrupting rating.
This thread will help you understand: https://diysolarforum.com/threads/eaton-amx-and-amh-fuses-alternative-to-class-t.60804/

Pure gold! Thanks for sharing.

 

[SunWatt]

How do you plan on measuring this test?

Seems like anything other than touching 2 cables together while hooked to a battery and seeing if it blows would require some very precise and expensive equipment. Especially when you get into the 100A and above. Fuses and breakers all have a load/time curve rating anywhere from fraction of a second to over a minute.
I have seen guys using a power supply to test small breakers, expecting an instant trip at rated amperage. Most are disappointed that it takes much higher current trip it, but not factoring in the time it carries it. Plus turning the dial on a power supply while holding a stopwatch isnt going to suffice. My point is, if you are on an information quest, read and understand mfg ratings, select the style and size for the performance you expect.

For the time being, I don't plan to do any measurements. The sole goal is to verify that the battery fuse can indeed open the circuit in a fault event.

The DIY battery that I'm currently working on is small in size, by design made just big enough to meet the power requirements I'm after, no more. There's more to this project but the bottom line is, I first need to nail the wiring and OCP devices.

What I'm looking for is a fuse in the range of 30-60A with more than enough AIC for a 48V LFP-based system.

I do understand there's more to a fuse than its current rating. Hence the reason why I'm trying to learn as much about it as possible.

 

[MattiFin]

Thanks for chiming in.

Yes, buying a quality fuse from a reputable supplier is the safest bet. However, at the very least I'd like to verify and know for sure that a certain fuse type will work in case of a fault in the system.

Agreed. You've made a valid point that even if the test is successful, by placing a new fuse that is untested there is no guarantee it will work as the one used in the test. But at least I'll know this particular fuse is suitable for the application.

It would suck if the test would end up damaging the BMS or the battery cells, but the BMS should be able to disconnect the battery in case of a short circuit, right?

If anything, to begin with, I could at least test the fuse AIC with a single 3.2V LFP cell, and if it doesn't work at least I've limited the damage to a single LFP cell.

IIRC V/R=I, meaning 3.2/1.8miliohm = 2064A which is way above the fuse rating and should blow the fuse.

The part I'm having difficulty understanding currently is how big of a difference voltage makes in testing if the fuse will manage to break the circuit in a fault event.

LE: i.e. is it necessary to test the fuse with a 48V battery as well just to be certain the fuse will work?

Trouble starts above 20 volts or so. Typical arc voltage in welding is about that.

Testing the interrupt capability at 3 volts or 48 volts gives totally different result.

”Homebrew” test setup would be huge 48v (or more) battery bank, oscilloscope and shunt resistor for current measurement. Huge contactor or pneumatically operated switch with contact piece made of 4lbs chunk of copper. For battery bank you would want something less finicky than lithium. Car starter batteries for example, ie 60 trcuk batteries, 6 in series and 10 in parallel.

I’d venture to questimate that most 48v BMS’s wont survive solid short. Overload, yes but direct short at output is a diffferent beast to handle.

 

[SparkyJJO]

Someone here had a BMS blow apart due to a short I think. Hey, it self fused!

 

[MattiFin]

Styropyro would have the battery bank to do the testing

 

[seneysolar]

Styropyro would have the battery bank to do the testing

 

[ArcFault]

A reputable fuse manufacturer will do the testing for you.
Look for the published ratings and trust it. Eaton/Busmann knows what they are doing.

Exactly. Manufacturers establish time-current (TC) curves for each circuit interrupting device and in the case of fuses perform destructive lot testing in the manufacturing process. The purchaser may also require additional testing and a "Certificate of Compliance" if the end use is a critical application. Some end users may have the expensive calibrated equipment needed to perform testing but in the case of fuses it is a destructive test and at best you can only establish that the representative sample from a fuse lot meets the manufacturer's TC curve.

Ungrounded DC systems should have fusing on both the positive and negative conductors. A single ground of either conductor to ground (potential of leakage current) is of concern but does not create a fault current. Having a method to detect leakage current is important in ungrounded systems. The major concern is either a direct short between positive and negative conductors or a simultaneous short to ground of both. In both instances it would require both fuses failing to open. Off the the top of my head I can't tell you risk of a fuse failing to open but assuming a conservative 1 in 1,000 (1.0 x E-3 probability) it would be a 1 in a million (1.0 x E-3 * 1.0 x E-3 = 1.0 x E-6) chance of failure of both fuses. Not exactly something to spend time, money, and effort for testing.

Electrical coordination is the single most critical aspect of circuit design and unfortunately is often overlooked in amateur designs. The link below gives a decent overview of TC curves / selective coordination. Prior to desktop computers the manufacturer would provide TC curves on mylar or velum of standardized size and scale. We would lay these in layers on a light box and trace the curves by hand.

https://www.mavtechglobal.com/pdf/white-papers/Time-Current-Curves-Whitepaper-2019.pdf