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

[ebaugh]

Originally Posted by T1 Terry

There is no history of lithium ferrous cells suffering an internal dead short, even if a cell in the 3v, 6v, 9v or 12 parallel pack was to go dead short, with a fuse in between the 6v and 9v parallel packs would mean the highest voltage at the shorted cell would that of 2 parallel packs in series. The pack with the dead short would already be pulled to a low voltage so the total voltage would be less than 6v, but possibly at very high amps.

Maybe the fuses would minimise potential damage to the other cells in the pack, but it adds so many failure points that would create cell capacity imbalance if a fuse failed due to fatigue, but wasn't noticed, that it may actually cause more problems than it would prevent.
T1 Terry

I won't debate if this is necessary or not, because I don't know. But it won't hurt, and for the most part detectable. I attached a picture of what I am considering, if the parallel fuses can be made small enough, any cell issue will be detected immediately. An overload on the bank side might trip a serial fuse and go undetected for awhile, but it won't damage the cells.

[Zil]

ebaugh; I can't get my head around how you lose 6 volts (9 volts) if a parallel fuse fails.

[ebaugh]

Originally Posted by Zil

ebaugh; I can't get my head around how you lose 6 volts (9 volts) if a parallel fuse fails.

Zil,

I don't lose 12V, or for that matter 6 or 9V. But if you were measuring the second cell voltage (the second column is a 4 cells paralleled) with a voltmeter between the 3V bubble and 6V bubble, it should read about 3V. OK, normally 3.2V since this is a LFP pack. But should any of the six parallel fuses blow, the voltmeter will read 0V. Detecting the failure of one of the 6 fuses. I call that the 6V level since that represents the second cell on a system designed to monitor cell voltage.

[NahanniV]

ebaugh:

I don't see how the voltage would go to zero either ?

Also, be aware that the Junsi monitor inputs are ground referenced; The controller calculates the differential, there is even a function that allows for the leads to be connected in any order and the controller will sort it out.

[ebaugh]

Originally Posted by NahanniV

ebaugh:

I don't see how the voltage would go to zero either ?

Also, be aware that the Junsi monitor inputs are ground referenced; The controller calculates the differential, there is even a function that allows for the leads to be connected in any order and the controller will sort it out.

Nathan and Zil,

You guys are right. I need to think more. It takes more than one fuse.... Hmmm.

[ebaugh]

OK. I thought I had it mostly solved. I was wrong. You need an ohm meter down the fuse line to detect an error. That would work. Not so easy to accomplish. I think the cell protection scheme works, but fuse failures are not easily detectable.

Or maybe I'm missing a different configuration? The voltmeter always had to be differential input, I don't have a Junsi, so it may or may not have that functionality. Differential means sensing voltage like a Fluke handheld DVM would, irrespective to ground, in this example for reading Cell 2,3,4 voltages.

We are leaving Panama tomorrow afternoon for San Andres Island arrival Sat morning. Maybe someone can figure it out? Starbucks gift certificate for anyone who can with one channel per cell differential voltmeter.

Bob

[hellosailor]

"fuse failures are not easily detectable."
You need fuses that blow more dramatically.<G>
Actually there are fuses with built-in idiot lights, just not in these types. What they do is put a cheap LED and dropping resistor in parallel with the fuse element, typically set to allow a 10-20mA current to light the LED when the fuse blows.
So, fuse blows? Red light next to it. Resistor still limits the current to 10-20mA which hopefully won't let anything else catch fire.
Bag of big cheap red LEDs, two cents each? And another nickel for the resistor?

Just saying, if you want a fast way to find blown fuses, there's an arguably reasonable way to do it.

[ebaugh]

Hellosailor...That's a good idea, but it won't let my battery manager detect the problem. I gotta go look in the battery box, I think? Bob

[hellosailor]

Depending on what your BMS is capable of detecting on, and what you're willing to spend, well sure, there's a way to turn an LED into a signal to the BMS. Or do like Boeing, and use a different battery type that emits gobs of easily visible smoke?

[Zil]

It would have to be an active short to light the LED. But the LED could be in a remote panel, wired from the bank. Also that is a possible signal to use for a trigger. Enjoy your voyage.

[hellosailor]

"It would have to be an active short to light the LED."
No, as long as the fuse has blown and there is ANY other connection being made, the LED will light. That could be an active short but it could also be any overcurrent condition. This is the crude but fairly effective commercial solution that's been on the market for a numbre of years now. There are better ways but nothing that can approach being as effective at such a low cost. Not perfect--just very cost effective.

[Gbtillman]

Interesting article, primarily the last few paraggraphs, on the Boeing battery testing and that only the LiFePO4 technology passes the testing.

http://www.foxbusiness.com/news/2013...tcmp=obnetwork

[roetter]

I want to help fiend replace his LA batts on a cat with LiFePo. I want to keep it simple for him and I am considering the per-packaged Balqon 12 V 90Ah times 6 batteries for him. Does anyone know what is inside the box I terms of BMS , specifically HVC and LVC?

The other option would be 4 times 700Ah clearance cells. But this would require that I have to spend two weeks with him installing and testing everything and I live on the other side of the continent.

Thanks for the help.

[hellosailor]

roetter, Balqon are a wholly-owned subsidiary of Winston, who have been reported to be in bankruptcy. Which means orders may not be filled (Balqon has apparently been sporadic about that) and there may be no company to warranty the batteries if the bankruptcy goes through. Just a couple of things to look into and perhaps beware of, for that source.

Their web site store indicates they "provide battery management solutions" which says to me that there are none in the battery itself.

[Zil]

HelloSailor; The fuses I referenced were the ones from positive to positive. I had to test using 9v and 1.5v dry cells. Yes the LED lights if the batteries are not balanced to volts.