LiFePO4 Batteries: Discussion Thread for Those Using Them as House Banks

[ebaugh]

I probably invalidated the AIC interrupt warranty on my ANL fuses since I took them out of the case. But could someone explain why AIC is even a factor for this fusing? I can see where a circuit breaker could fail due to internal issues, say getting welded on the inside, but how could my ANL fuses possibly fail to open at any current level? I suppose I could vaporize both a normal interconnect and a fuse simultaneously, but I just don't understand how it could possibly ever ever fail to open. Can anyone explain?

[ebaugh]

Originally Posted by hellosailor

eb-
The problem with your "bolt" theory is that so far no one has reported the presence of any "bolt" near that location.

I would suspect that every bolt and screw on that aircraft has been treated with a thread-lock and assembled with a torque tool of some kind. That doesn't make it impossible for one to come out and generate enough force puncture something, but it becomes extremely unlikely when there's no bolt observed in the area and no bolt location.
:-)

I maybe wrong here. Bolt was a generic term. But what caused the puncture shown in the X-ray? I suppose it could be an internal cell issue, but that doesn't explain the evidence of high current to ground.

[daddle]

[QUOTE="ebaugh;1182367"... I can see where a circuit breaker could fail due to internal issues, say getting welded on the inside, but how could my ANL fuses possibly fail to open at any current level? I suppose I could vaporize both a normal interconnect and a fuse simultaneously, but I just don't understand how it could possibly ever ever fail to open. Can anyone explain?[/QUOTE]

For DC, one failure mode is for an arc to form. The current continues while starting a fire. Switches and fuses are commonly rated much lower for DC than AC as the AC arc will likely extinguish during the AC cycle.

[ebaugh]

Originally Posted by bill good

Given the small difference in micro volts between cells in parallel in the normal cell operations, would it not be better to use very low valued fuses for the parallel & the high value for the series current limit? The parallel will keep balance between the parallel cells. I opt to solder a single fuse wire between the parallel cells with no trouble so far.
.

I did use 80A parallel fuses. In hindsight, I should have used smaller values, as you said, these connections mostly exist to keep the parallel cells in balance with one another. I would be surprised if they ever see more than a few amps. But for my cell type, the ANL form factor worked pretty well.

The series fuse configuration is a bigger question. I have a 300A fuse on the pack positive before it goes anywhere.

[s/v Jedi]

I hope it is common to use a class-T fuse in the positive lead of the bank. This used to be, and I think it still is, the best fuse to handle big currents on boats.

[ebaugh]

Originally Posted by diugo

Second, a tool accidentally grounding a terminal is likely enough that it is prudent to place the catastrophe fuse on the battery negative terminal---between the battery and ground. That way, if any cell terminal is shorted to ground, the fuse will blow. Third, fuses do have nonzero resistance, each of which will add to the total voltage drop under high current. So battery performance may be impacted by excessive fusing.

I'm not sure ABYC permits fusing on the negative connection? Anyone know?

Does the voltage drop cause significant inefficiency? The voltage drop itself is not a big deal alone since LFP is already 13+ volts.

[ebaugh]

Originally Posted by s/v Jedi

I hope it is common to use a class-T fuse in the positive lead of the bank. This used to be, and I think it still is, the best fuse to handle big currents on boats.

It is....

[diugo]

Originally Posted by Dsanduril

Blue Sea shows 6000AIC@32VDC, and this datasheet direct from the Cooper website has removed the distinction, but still lists the ANLs for up to 80VDC:

http://www.cooperindustries.com/cont...NL_Limiter.pdf

The BDFuses datasheet says it is from '04 (the Cooper one is not dated), so looks like Cooper changed the specs.

Thanks for the link. I noticed the 2004 date right off, and checked two other sites---one of them Cooper itself---and got the same 2004 datasheet.

Re a fuse on the positive line, that is certainly more common than finding it on the negative line. But putting it on the negative does have the advantage of protecting against errant tools.

[hellosailor]

ebaugh-
"But what caused the puncture shown in the X-ray? "
Since Boeing, the NTSA, the FAA, and Nippon Air haven't been able to figure that out, I don't mind at all saying I haven't got a clue either. Although, unlike all those august bodies, I am allowed to say "Gremlins." with a straight face.

daddle-
" how could my ANL fuses possibly fail to open at any current level?" I've always figured explosion disintegration was a fairly robust and reliable process, once you got the whatever it was to start exploding. Maybe the problem is that an ANL fuse subject to too much amperage will vaporize, turn into a conductive plasma, and then effectively remain a conductor courtesy of the plasma bridge? Maybe not for long, but for long enough to exceed some standard that says it has to STOP conducting within so many microseconds?
Or, maybe the problem is that when an ANL fuse blows, it fails to blow with sufficient force to dislodge any GREMLINS AGAIN that might be holding onto both haves and bridging the circuit?

This is where the skeptics chime in and swear gremlins live in the clouds and only attack aricraft, never boats. Oh, sure, like anyone would believe they don't attack boats. <VBG>

[diugo]

Originally Posted by Maine Sail

FWIW I just popped a number of MRBF and ANL fuses, on purpose, using my Li bank..

None failed to pop and none disintegrated. I even popped a 300A MRBF using 15' of 8GA wire and my 400Ah Li bank... Just curious and wanted to see how these fuses handled the Li current into a dead short.

This experiment is highly flawed. The fuse has resistance of its own, say five millohms. Add this to the battery's two millohms, and the maximum possible current is 13.3/.007 = 1900A, according to Ohm's Law.

How much resistance does 15' of #8AWG add? Another 12 millohms by my reckoning. 13.3/.019 = only 700A. Not really a good test of the A.I.R.

The fuse interrupt rating becomes more of factor as the bank voltage approaches the fuse's voltage rating---since doubling the voltage doubles the amperage.

One more point. A fuse is required to handle the interrupt current without exploding or catching fire, but also failing in a mechanical way that might weld its terminals together. It is also required to operate in a specific, usually very short period of time. It does so to protect connected equipment---most notably, the insulation on cabling. You can imagine that with thousands of amperes flowing right next to you, every millisecond counts when you're made of plastic.

[hellosailor]

"This experiment is highly flawed. The fuse has resistance of"
I appreciate the point but question the numbers. Couple of years ago I was told in no uncertain terms that I could not use the common DC breakers as a battery primary breaker, because they were only rated for some 3000A and even a single Group31 battery could put out more than that in a crowbar short.

So assuming a breaker and a fuse have similar resistances, and a Group31 lead acid battery has no more resistance than Maine's Li battery...Are those assumptions rashly flawed? Or is there something more still to be considered here?

And then again, if the purpose of Maine's experiment was to see if the ANL fuses are practical and will blow...the math is all nice but he did prove they will work. The math contradicts the manufacturers' recommendations. What's wrong with that picture?

[diugo]

Originally Posted by ebaugh

Does the [fuse's] voltage drop cause significant inefficiency? The voltage drop itself is not a big deal alone since LFP is already 13+ volts.

My battery has an internal resistance of about 8 millohms, which creates a 0.5-volt sag all by itself at 60A. However, my inverter sees a whopping 0.7V total sag as a result of the large 200A class T fuse and about 4' of #4AWG cabling. This suggests the fuse adds 3 millohms---quite significant relative to the other resistances.

[diugo]

Originally Posted by hellosailor

...The math is all nice but he did prove they will work. The math contradicts the manufacturers' recommendations. What's wrong with that picture?

Ugh. All he proved was that the fuse would blow at 1900A and 700A respectively---well below the claimed 6000A rating.

[daddle]

Re the rating in breakers: I would not be so sure the breakers are expected to survive the 3000 amps. Just that the circuit opens. I have had plenty of breakers cease to properly function after an extreme overload. Perhaps that is the reason for the recommendation.

[diugo]

Originally Posted by hellosailor

"Couple of years ago I was told in no uncertain terms that I could not use the common DC breakers as a battery primary breaker, because they were only rated for some 3000A and even a single Group31 battery could put out more than that in a crowbar short.

So assuming a breaker and a fuse have similar resistances, and a Group31 lead acid battery has no more resistance than Maine's Li battery...Are those assumptions rashly flawed?

Not rashly flawed, just a bit shortsighted. A particular group31 AGM battery has a resistance of 2.2 millohms according to Odyssey Battery - AGM product line . So its maximum possible current is therefore 12.6/.0022 = 5727A which is obviously more than 3000.

But is that really the only resistance in the circuit? Even in a crowbar short, the crowbar itself has resistance. Moreover, there is relatively high resistance at the poor contacts, and it's not inconceivable that even a breaker rated at 3000A could successfully trip before the battery exploded.

I suppose one could argue that in order to be "conservative", the resistance of all other components in a breaker-protected circuit should be assumed to be zero. In reality, though, electricity doesn't work that way.