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

[Checkswrecks]

Since this thread is about building house banks with LiFePo4 I'll throw in three things not discussed enough. These will be common sense to John, Newhal, CatNewbee and some of you others but worth repeating for somebody who may cruise through here after seeing "build a battery" YT videos.

The first is to not package prismatic cells tightly. While Prowse and others will package them up with black tape for their videos, the tape has a lot of give and it's just to make a video of a test. There are also videos made by people showing how they built pretty but tight wooden boxes or how they used actual clamps.

Besides wood not being a good idea for encasing a battery which can have a fire (duh) cells will swell, shrink, and move around a surprising amount with charge and discharge.

The second point is about the value of putting a simple piece of aluminum between cells. When prismatic cells get hot, the heat tends to be in the middle of the cell. That heats the adjacent cell and is the furthest distance to the edges for the heat to be transferred away. A piece of aluminum does two things. It diffuses the affect on the adjacent cell and it helps heat-sink the energy to where it can dissipate.

The third comes from the basic need to assume that a cell will fail in the worst way, even though we know LiFePo4 is pretty safe. Assume the charger fails and dumps a 5C rate, goes to a high voltage, or a weird box construction error causes everything from the charger to hit one cell. Think about where the cells will puke the cell guts to if a vent releases the cell contents.

I've seen a lot of failures since 2013 and know of one design which originally kept the vented electrolyte inside a tight battery case because of a regulation stating that the design could not spill (flammable) electrolyte beyond the case. Something like this is actually is pretty common in EV modules. The problem was that heat from an initial cell going into runaway would add to the radiant heat of that cell, enhancing propagation of the failure into adjacent cells it was stacked against.

That manufacturer put the battery case holding the cells into a slightly larger box with holes outside of the cell vent ports. This allowed the individual cells to be able to puke into the bigger box through holes in the inner case. Once the vented electrolyte was moved away from the adjacent cells their propagation problems ended.

[evm1024]

Originally Posted by Checkswrecks

Since this thread is about building house banks with LiFePo4 I'll throw in three things not discussed enough. SNIP %<

The third comes from the basic need to assume that a cell will fail in the worst way, even though we know LiFePo4 is pretty safe. Assume the charger fails and dumps a 5C rate, goes to a high voltage, or a weird box construction error causes everything from the charger to hit one cell. Think about where the cells will puke the cell guts to if a vent releases the cell contents.

SNIP %<

Let me be the first to say that it is unlikely to see a charger failure that will deliver 5C to any of the house banks that I am aware of.

In my own case I would need a 48 KW charger to do that.

Also, my shore power cannot supply 400 amps at 120 volts.

Even in the case of a 5C dump to a single 100 AH Battleborn battery I do not have enough available current. At best I might be able to come up with 300 amps.

And various circuit protections will come into play for a charger fault and open the circuit. Plus should there be a charger fault and the battery dumps high current into the charger I would expect the main house bank fuse to open. In around 01. to 0.2 seconds for a dead short and perhaps 1 to 5 seconds for a short limited to 500 amps. (based on ANL fuses)

[Cpt Pat]

Originally Posted by john61ct

This is obvious, from easy observation, just go back and read this thread, read Maine Sail's article and hundreds of posts from people with no axe to grind

> since no LFP bank supplied by any manufacturer, or assembled by any person with half a brain lacks circuitry to balance at voltages at, or above said shoulders

Yes top balancing needs to be done at high voltages.

If you are forced to use a stupid BMS that only balances at damagingly high voltages, you should not let it do your balancing. Just don't allow voltage to climb above the start-balance setpoint and you'll be OK.

Balancing can be done in many alternative ways, **if** it ever needs to be done at all.

But with quality well-matched cells in good condition, it does not need to be done frequently, or at all, for many years at a time

**if** you avoid going into the shoulders in your normal cycling.

Unfortunately, some people get hoodwinked into buying drop-ins that do not allow for proper care, and, along with the fact that the quality of the cells themselves is questionable, that will inevitably result in relatively poor longevity.

Yes, then people jump in, say who cares, don't worry about these persnickety details, long as we get X years we'll just replace them.

Which is fine I guess, but not what I'm about.

How about this:

Obviously it is unmatched capacity, differing ESR levels that are the primary cause of cell imbalance, especially as they wear from cycling and age.

Even if you do not believe that pushing end-charge up to high voltage/SoC exacerbates cell imbalances,

surely you can see that they do not **manifest** as a practical problem if you usually stay away from the shoulders?

And do you really believe that pushing past the shoulders reduces lifetime cycles?

That in itself means worse imbalance problems arriving sooner. . .

After two years without any cell balancing circuitry, I got out my Fluke VOM, my clip leads, and my 2.7 ohm resistor to do my annual manual cell top balancing. Same as last year, there was nothing to do. Cell voltages at 100% SOC and 0.03C charge current: 3.49, 3.51, 3.53, and 3.48 volts. I'll try again next year.

Except for an occasional charge to 100% SOC every 30 cycles or so, I run between 20% and 80% SOC to stay off the knees, and operate the bank at sub-fractional charge and discharge rates (charge < 0.2C, discharge < 0.05C). My bank does about 100 full cycles per year.

Since I don't have a sealed drop-in battery, I can confirm each cell is healthy. With drop-in batteries that prohibit individual cell access (without sawing the case open), I would feel committed to having cell balancers, and cell balancers would be just something else to go wrong. If one shorts (as MOSFETs are known to do) - it becomes a "cell destroyer." In normal circumstances, the balancers are vampires - slowly sucking the cell charges away. I'm not a big fan of the drop-in approach. Especially in higher "C" rate usage where the surrounding drop-in case prevents heat dissipation.

[john61ct]

Thanks Cap'n!

______
These came up in another thread, thought I'd post in a more LFP relevant location.

CTEK and Tecmate/Optimate from the automotive world, make great high-tech sealed chargers which don't tend to get used in a marine context much, likely because (afaik) they don't make high-current units.

But they are well designed, high build quality, very precise and last forever.

______
The Optimate TM-291 charger is an excellent quality and surprisingly useful unit, designed for those tiny LFP starter batts, so low amps.

Usually I'd never use a 14+V charger, but as a "finish" charger on even a large 4S bank it's superb,

ends up **balancing** the four cells / groups just perfectly **without** any BMS or balance lead connections, just the pack-level power pair.

Crazy I know, and no idea how it does it, functionally just a "black box",

apparently to do with how it pulses, rather than just sitting at a static absorb voltage, oscillating between 14V and a bit higher, maybe 14.5? The lower-volt cells at lower resistance accept more energy per pulse or something.

In any case it works!

------
Their TM-500 is a 2A DC-DC, can be used for part of that "hybrid LFP / lead" discussed in a couple dedicated threads

Similarly but ten times bigger at 20A, CTEK has improved their lead-designed B2B line

now D250SE supports "LFP profiles" but unfortunately does not allow user-custom setpoint adjustability,

does include a basic SC though.

[Cpt Pat]

I see a big elephant in the room with drop-ins: heat dissipation. Putting cells inside a sealed, unventilated, plastic box seems to invite LFP's greatest enemy: heat. Unless the manufacturer is using some "special" plastic with high thermal conductivity: I don't see a battery box -- I see an oven.

That may not be an issue at fractional C rates, but as rates approach or go beyond 1C, I'd start to worry about how to dissipate heat.

And if a single cell should ever overheat for any reason, sealing all the cells in an unventilated plastic box seems to greatly increase the risk of a cascading thermal failure of all the cells. Sealing the cells in plastic thermal insulation seems like a very bad idea to me.

I'd like to know the thermal conductivity of the battery case, in degrees C per watt battery surface-to-ambient.

[newhaul]

Originally Posted by Cpt Pat

I see a big elephant in the room with drop-ins: heat dissipation. Putting cells inside a sealed, unventilated, plastic box seems to invite LFP's greatest enemy: heat. Unless the manufacturer is using some "special" plastic with high thermal conductivity: I don't see a battery box -- I see an oven.

you do however realise that that 140 watts of heat is nothing considering for example the Battleborn 100ah battery . It weighs 14,000 grams. (31pounds)

[john61ct]

Any elevated temps will lose lifecycles, but if otherwise cared for well likely not a high percentage, and off a long-distant back end.

The inaccessibility of the BMS, both settings & data, and no per-cell connections seems more serious to me.

But so far they seem to be doing the right thing with warranty service claims,

so maybe the questionable cell sourcing is not such an issue,

if the company lasts past ten years from purchase date may well be a decent buy for those

who are happy with limited C-rates, and would rather not get into the details.

[Cpt Pat]

Originally Posted by newhaul

you do however realise that that 140 watts of heat is nothing considering for example the Battleborn 100ah battery . It weighs 14,000 grams. (31pounds)

The amount of mass is important, as you point out. Given those numbers, and assuming no heat dissipation at all, it would take 100 seconds to raise the internal battery case temperature 1 degree C, and 7,500 seconds (125 minutes) to raise the temperature from 25 C to 100 C.

Of course, the thermal resistance of the case isn't infinite, as it would have to be in the above example. But intuitively, I'd guess thermal transfer of cell case-to-ambient would be very poor.

Even in my sub-fractional C application with my prismatic cell pack exposed directly to air, I see signs of accelerated aging, in the form of slightly reduced capacity in the cells in the middle of the pack; which are exposed to air on only their three outer sides (the cell bottoms are not exposed). It's too small to worry about, but I did rotate the cells in the pack a year ago as a test, and now the cells that were on the ends of the pack that were moved to the middle show slightly less capacity compared to the cells on the ends.

If I were to build a large pack operating at high C rates, I'd immerse the cells in a liquid (maybe transformer oil) and provision a circulation pump and heat exchanger. Which is the opposite thermal approach to what is implemented with drop-in plastic-encased batteries.

[john61ct]

Apparently mineral oil is used for long-term storage setups.

At least for plastic-encased, aluminum may have different requirements

[newhaul]

Originally Posted by Cpt Pat

The amount of mass is important, as you point out. Given those numbers, and assuming no heat dissipation at all, it would take 100 seconds to raise the internal battery case temperature 1 degree C, and 7,500 seconds (125 minutes) to raise the temperature from 25 C to 100 C.

Of course, the thermal resistance of the case isn't infinite, as it would have to be in the above example. But intuitively, I'd guess thermal transfer of cell case-to-ambient would be very poor.

Even in my sub-fractional C application with my prismatic cell pack exposed directly to air, I see signs of accelerated aging, in the form of slightly reduced capacity in the cells in the middle of the pack; which are exposed to air on only their three outer sides (the cell bottoms are not exposed). It's too small to worry about, but I did rotate the cells in the pack a year ago as a test, and now the cells that were on the ends of the pack that were moved to the middle show slightly less capacity compared to the cells on the ends.

If I were to build a large pack operating at high C rates, I'd immerse the cells in a liquid (maybe transformer oil) and provision a circulation pump and heat exchanger. Which is the opposite thermal approach to what is implemented with drop-in plastic-encased batteries.

if you are seeing capacity losses you should have posted that in the memory effects thread .
But you did post the opposite saying you have seen no losses . Which is it ?

http://www.cruisersforum.com/forums/...ml#post3042952

Hard to follow and set my bank in the best possible to mitigate issues if I get conflicting information.

[Cpt Pat]

Originally Posted by newhaul

if you are seeing capacity losses you should have posted that in the memory effects thread .
But you did post the opposite saying you have seen no losses . Which is it ?

http://www.cruisersforum.com/forums/...ml#post3042952

Hard to follow and set my bank in the best possible to mitigate issues if I get conflicting information.

I am seeing what may be a very small capacity loss in half my cells (2) due to what may be accelerated heat-induced aging - indicated by higher aging of cells in the center of my pack versus at the outer ends of the pack, caused by what I suppose is differential heat dissipation at the pack ends versus the pack center. I have experimented to test that hypothesis by swapping the cells positions to place the center-mounted cells at the ends, and observed the same effect following the cell pack positions after repositioning the cells. This observation may prove my hypothesis. But I stress, in my case, the effect is very small. One could argue it is statistical noise. What I do know for sure is that heat accelerates cell aging. That is well proven. And that cells having limited heat dissipation capacity will experience more heat.

That is very different than memory effects, which I am not seeing and which would presumably affect all the cells equally. I had hoped I had communicated clearly in my post above that I was referring to thermal effects, not memory effects. I don't know how to be more clear. Any conflict with memory-effects is imagined, because I have never experienced nor described memory-effect capacity degradation with my pack.

The effect I am seeing is very small, but the heat generated is also very small at a maximum charge rate of equal to or less than 0.2C. If I were charging those cells at 0.5C while they were sealed in a thermally-insulating plastic case -- I strongly suspect the entire pack would undergo significant thermal aging. My point is: placing cells inside what essentially comprises an oven, as is done with drop-in batteries inside plastic cases is not good!

Heat is bad. LFP batteries generate heat. Dissipating that heat is good. Drop-in battery replacement outer cases are sealed plastic. Therefore those cases are bad. Is that simple enough? Memory-effects are a separate topic.

(Deep sigh.) Therefore, my post does not belong in the memory effects thread. There is no "thermal effects thread," though I would argue there should be. So here is my post, in the more general "...house banks" thread. And the discussion was about drop-in batteries (Battle Born), to which my posts directly pertain.

----

How much more would it cost to construct the drop-in battery container case (to make them physically the same dimensions as the AGM batteries they are intended to replace) of a material, like aluminum, that has much better thermal conductivity?

[Cpt Pat]

(Sorry folks. I may be violating one of my own rules.)

No.. you violated one of CFs. Deleted attachment. (weavis for the mods)

[evm1024]

Originally Posted by Cpt Pat

I see a big elephant in the room with drop-ins: heat dissipation. Putting cells inside a sealed, unventilated, plastic box seems to invite LFP's greatest enemy: heat. Unless the manufacturer is using some "special" plastic with high thermal conductivity: I don't see a battery box -- I see an oven.

That may not be an issue at fractional C rates, but as rates approach or go beyond 1C, I'd start to worry about how to dissipate heat.

And if a single cell should ever overheat for any reason, sealing all the cells in an unventilated plastic box seems to greatly increase the risk of a cascading thermal failure of all the cells. Sealing the cells in plastic thermal insulation seems like a very bad idea to me.

I'd like to know the thermal conductivity of the battery case, in degrees C per watt battery surface-to-ambient.

I do not consider heat to be LFP's greatest enemy. Heat is everything's enemy (or excessive heath rather). For example:

Most humans will suffer hyperthermia after 10 minutes in extremely humid, 140-degree-Fahrenheit (60-degrees-Celsius) heat. Death by heat stroke is not far along.

The 140 watt figure introduced by newhaul is excessive. Remember that at 98% charge efficiency there is only 2% available for heat generation. On a 100 AH battleborn charging at 13.3 volts and 100 amps (1C) we have 13.3 x 100 x 0.02 = 26 Watts.

As we try to understand all this stuff we should keep in mind that theory is just that and book learning is just that. I'm not putting anyone down I'm just saying that when I charge my 700 AH bank at max charge current that I can generate I do not see the battery terminals on the cell rise very much in temp using an IR pyrometer. Theory and reality clash.

Moving on to Cpt Pat's experience. How was the temp increase measured? I'm more curious than anything else.

Also, how was the individual cell capacity measured? Typically cell chemistry have a higher reactivity as the temp goes up. This results in a higher capacity from the warmer cells. This appears to be backwards from what you were describing.

And, heat induced capacity loss appears to result in permanent loss. So, this also appears to be contradictory to what you have observed.

I wonder what is going on? It goes back to how the individual cell capacity was measured. I guess you could measure a lower cell voltage at low SOC and equate that to lower capacity. Is that true?

What bank size and who makes your cells. I recall that you have stated that but do not remember off the top of my head.

[evm1024]

Let me correct myself - it was Cpt Pat that introduced the 140 watt figure.

Originally Posted by Cpt Pat

100 amp charge rate at 14 volts = 1400 watts. Assuming 99% charge efficiency (which may be generous), the heat being dissipated is 140 watts. With no thermal conductivity, that power will raise 140 grams 1 degree C per second above ambient.

The 2 comments here are first off there is a math error and second the charge voltage for the majority of the charge cycle would be less than 14 volts. Not low enough to make a big difference but lower.

For my cells the typical charge voltage is around 13.5 volts average (3.375 vpc) for the majority of the charge cycle.

But let's use 14 volts for the calculation:

14 volts x 100 amps = 1400 watts

A 99% charge efficiency is a 100% - 99% = 1% loss (1% / 100 = 0.01)

Thus 1400 watts x 0.01 loss = 14 watts (available to generate heat)

[tanglewood]

Originally Posted by Cpt Pat

I see a big elephant in the room with drop-ins: heat dissipation. Putting cells inside a sealed, unventilated, plastic box seems to invite LFP's greatest enemy: heat. Unless the manufacturer is using some "special" plastic with high thermal conductivity: I don't see a battery box -- I see an oven.

That may not be an issue at fractional C rates, but as rates approach or go beyond 1C, I'd start to worry about how to dissipate heat.

And if a single cell should ever overheat for any reason, sealing all the cells in an unventilated plastic box seems to greatly increase the risk of a cascading thermal failure of all the cells. Sealing the cells in plastic thermal insulation seems like a very bad idea to me.

I'd like to know the thermal conductivity of the battery case, in degrees C per watt battery surface-to-ambient.

This is probably part of why drop in batteries typically have pretty restrictive current limits.