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

[JmanC]

Originally Posted by CatNewBee

It is pretty simple.
Cells drift and get out of balance with each cycle.

Also the whole concept of BMS balancer with low currents is based on the idea, that you charge your battery to full, and this most of the cycles and not cycle at low soc. This allows small adjustments to one cell and approximation to the others over a few cycles, also keeps the cell in balance regardless of small drifts on each cycle. Avoiding the shoulders - effectively meaning not charging the battery to full - only causes larger drifts, that cannot be handled by the BMS in one or two cycles, but need either manual correction or many cycles with low currents, making the battery unusuable during this time.

There is a reason why the technical data points out, what the proper set points should be and how to charge the battery.

Since some report their banks have stayed pretty balanced after a good few years and hundreds of cycles. Why not simply charge a dropin to those balancing voltages say every 20-50cycles or once a month or something?
Wouldn't that be enough to keep that battery well balanced without going into the high voltages every cycle?

[john61ct]

Personally I do not currently plan to have anything to do with drop-ins at all.

A drop-in with cells of unknown quality, and a design without a BMS **and** no access to cells leads, would be the worst of all possible worlds.

IMO only cells from the top known-good makers should even be considered, my expectation is decades of use, many many thousands of cycles, IMO 2-3000 cycles is nothing, people happy with those numbers should indeed not concern themselves with the finer points of care.

Originally Posted by JmanC

So I take it that means my new dropin had bad matched cells or not a full balanced job?

If you have access to view the individual cell voltages, you should be able to ensure the balancing process continues until it gets done.

The less frequently this happens, the longer it can take, sometimes even many days when the balancing rate is too low.


> I agree the BMS dropin balancing is going to cause more longterm damage that any good it will do

Please don't put words in my mouth, I do not believe that.

I do think the way most BMSs do balancing is a very flawed design, and when they fail, or there are bad connections, etc yes they can damage cells.

I am in no way advocating "no BMS", but I view "BMS" as a collection protective functionalities, not a single gadget you buy and done.

_____
The only way to check for well matched cells - both Ah capacity and ESR, is with physical access to cell balance leads.

This should be done at commissioning time, so packs with excessive differences can be returned under warranty.

Originally Posted by JmanC

Why not simply charge a dropin to those balancing voltages say every 20-50cycles or once a month or something?

Wouldn't that be enough to keep that battery well balanced without going into the high voltages every cycle?

Maybe, with well-matched top quality cells, as long as they are in good shape.

But with many banks, for example those accepting poorly matched or lesser quality cells, working with older, worn packs, or even buying second-hand cells cheap

then balancing to completion may be required every week, or even every charge cycle.

In some cases, the start-earlier, higher rate balancing solutions may be needed, the standard BMS approach (stop-charge resistance bleeding at a very low rate, high V starting point) may simply not be up to the job, becomes completely impractical.

Then the other "more drastic" methods above may become necessary.

Some high-rate balancing gear can even be left active 24*7! Not advocating this nor any of the above, just laying out scenarios and options.

Many will say, cells requiring such heroic measures to be kept in balance are past EoL and should be scrapped and replaced. That is a valid approach, for those with decent resources, but others may prefer to keep going, and maybe get many years productive use out of them.

Of course, drop-ins don't give those choices, only when you have access to cell balance leads and the option of replacing the original BMS.

Originally Posted by JmanC

I always thought going into the shoulders simply ages the battery slowly overtime reducing capacity.

The factors that accelerate wear (capacity loss) are the same factors that exacerbate increasing imbalance.

The strategies that slow down capacity loss also delay imbalances, and **more significantly**, prevent inevitable capacity / resistance imbalances from emerging as a problem affecting performance.

[john61ct]

>> I absolutely do recommend going to higher SoC/voltage charge profiles, if so-called memory effects are observed.

>> This is not required for normal usage cycling however, should IMO be a maintenance protocol like equalizing FLA.

Originally Posted by Delfin

Well, I guess I should have been paying better attention to the evolution in your thinking, since not that many posts ago, you were recommending exactly the opposite.

Not at all, your misinterpretation.

For normal usage cycling, charge to well below the shoulders, setpoint voltage varies with the current rates, CV or CC-only, depending on the nature of the source whether human regulated vs automated etc.

If SoH testing shows no "memory effect", then there's never any need to go higher. Assuming no balancing that requires doing so.

So far I have not seen any "memory effect"; perhaps the above variations in the charge-termination setpoints prevents it. Note also that my capacity benchmark protocol **requires** going to much higher than usual voltage/SoC, a great example of the classic Heisenberg / observer effect - testing to determine the state of the system, in itself alters that state.

But if capacity **did** decline and that effect were suspected as the cause, then start going to higher voltages, maybe hold Absorb to lower endAmps, vary current levels etc. whatever works to overcome the effect and restore capacity, details discussed in those specific threads.

[john61ct]

>> 3.65Vpc charge termination, along with other specs laid out by cell manufacturer are **absolute maximum** ratings, the limits to which the cells can be subjected for short times, without causing immediate irreparable damage.

Originally Posted by Delfin

So, in your thinking, if a cell is exposed to 3.7 v, it is immediately and irreparably damaged.

Again misinterpretation, I have never even thought such a thing. Just silly to interpret the above that way?

If a power supply is rated at 1000W, then it should not be used anywhere near that power level continuously, if you want good longevity.

Same with physical specs, say an adhesive's tensile strength and elongation ratings, need to factor in a margin if high reliability is critical.

The point is to avoid overly stressing components.

I think fundamentally, the desire to get to a theoretical 100% SoC, as opposed to say 96%, is a subconscious holdover from decades od following the care protocols required for lead banks.

[john61ct]

Originally Posted by Delfin

manufacturers recommendations for charging voltages were based on their desire to shorten the lifespan of their products

Again, nope, just your interpretation.

From above

Originally Posted by john61ct

recognize the cell manufacturers are huge billion-dollar companies, and all their focus is on their main customers, first the Chinese military, and secondly EV manufacturers. These have their own PhDs in battery science and DC power, and priorities are high-C-rate discharge for propulsion and range / per-use cycle capacity - expected lifespans are relatively short in those use cases, and optimizing for longevity is just not high on their list of priorities.

The purchase of LFP cells directly by consumers for use as "House bank" storage in a mobile context and for solar storage is at most .0001% of the cell manufacturers' sales, not even on their radar in economic terms, even as a niche sub-market.

It is for these same reasons that industry and academic research is not targeted toward maximizing longevity, but oriented around those EV & other high C-rate use cases, so **very** different from our much gentler House bank cycling.

And thus no testing / different specs are developed by industry for the House bank use case

These extreme high C-rate (both charge and discharge )+ deep DoD% use cases, being 99.9999% of their market, prevent lifespans longer than the rated 2-3000 cycles, piddly factors like storage voltage or adjusting the stop-charge point will have **near zero impact** on longevity.

So why would anyone do expensive formal studies on them for the shallow-cycling, gentler currents of mobile House banks, at .0001% of the use cases?

That is by far the main reason I've always stated for our reliance on the less rigorous amateur / hobbyist test results, plus vendors like T1 Terry and Maine Sail.

______
The auto industry could easily have developed utilitarian cars and parts supply chains, so that their customers could have the option of buying vehicles that routinely last 50-100 years.

The free market gives them, as rational players, no incentive to do so. Just as there is no incentive to double or triple lifespan.

Not talking any conspiracy here, just an understanding of basic economics.

And this is just one relatively minor contribution to this issue, but I'm sure you'll be trotting out that "tinfoil" straw man again in another few months.

____
Edit - see LTO chemistry, can go 10,000 cycles, also extremely safe, performs much better in the cold, can go down to zero volts with no damage, crazy high charge and discharge*C-rates, up to 10C without stress, fully charged in 6-7 mins!

Been around a long time, but very little uptake. Why?

Because their lower energy density, even lower than LFP, makes them unsuitable for EV / propulsion where weight is **the** critical issue.

The mobile House bank, even Solar banks for S&B homes are just not at all significant markets for the international players needing to invest billions in production facilities.

Completely uninterested in developing robust sales/distribution channels to retail or direct to consumers, all the oxygen sucked up by their huge-volume institutional customers.

[Q Xopa]

Originally Posted by john61ct

Again, nope, just your interpretation.

From above

These extreme high C-rate (both charge and discharge )+ deep DoD% use cases, being 99.9999% of their market, prevent lifespans longer than the rated 2-3000 cycles, piddly factors like storage voltage or adjusting the stop-charge point will have **near zero impact** on longevity.

So why would anyone do expensive formal studies on them for the shallow-cycling, gentler currents of mobile House banks, at .0001% of the use cases?

That is by far the main reason I've always stated for our reliance on the less rigorous amateur / hobbyist test results, plus vendors like T1 Terry and Maine Sail.

______
The auto industry could easily have developed utilitarian cars and parts supply chains, so that their customers could have the option of buying vehicles that routinely last 50-100 years.

The free market gives them, as rational players, no incentive to do so. Just as there is no incentive to double or triple lifespan.

Not talking any conspiracy here, just an understanding of basic economics.

And this is just one relatively minor contribution to this issue, but I'm sure you'll be trotting out that "tinfoil" straw man again in another few months.

____
Edit - see LTO chemistry, can go 10,000 cycles, also extremely safe, performs much better in the cold, can go down to zero volts with no damage, crazy high charge and discharge*C-rates, up to 10C without stress, fully charged in 6-7 mins!

Been around a long time, but very little uptake. Why?

Because their lower energy density, even lower than LFP, makes them unsuitable for EV / propulsion where weight is **the** critical issue.

The mobile House bank, even Solar banks for S&B homes are just not at all significant markets for the international players needing to invest billions in production facilities.

Completely uninterested in developing robust sales/distribution channels to retail or direct to consumers, all the oxygen sucked up by their huge-volume institutional customers.

Uncanny, so misinterpreted.

[tanglewood]

Imbalance will happen over time due to variations in cells. It's inevitable.


Just because you operate "away from the knees" doesn't mean you are avoiding or preventing imbalance. It will still develop, but just be less detectable.


The only time imbalance really matters is as you approach top of charge, or bottom of charge. Then cell that are out of balance will reach higher or lower voltages sooner than all the others, and there starts to be a risk of driving that out of balance cell into an undesirable voltage range. The further away from the edges you stay, the more imbalance you system can tolerate without driving any one cell out of range. Otherwise it really doesn't make much different at all.


As for "drop-ins", I think it's unfair to say that a drop in has cells of unknown quality. How do you know what cells are used in that drop in, or any of the non-drop-in batteries available? Do you know who's cells are used in Victron's, Lithionic's, and MasterVolt's products? And do you know if they have changed cell vendors?


I also think it's incorrect to say that drops ins don't have a BMS. They typically do. It's just self contained.

[Maine Sail]

Originally Posted by john61ct

You actually helped me figure out shipping on multiple occasions for several heavy / bulky technical items, including a big bank, and I sent several customers your way as prospects for bare cells.

Confused who you are talking to here?

[newhaul]

Originally Posted by Maine Sail

Confused who you are talking to here?

that is my question as well

[john61ct]

Originally Posted by tanglewood

I also think it's incorrect to say that drops ins don't have a BMS. They typically do. It's just self contained.

Yes, never seen one without, CNB was I think suggesting that.

>> As for "drop-ins", I think it's unfair to say that a drop in has cells of unknown quality. How do you know what cells are used in that drop in, or any of the non-drop-in batteries available? Do you know who's cells are used in Victron's, Lithionic's, and MasterVolt's products? And do you know if they have changed cell vendors?

This is my list of known-good quality cell makers

Winston/Thundersky/Voltronix, CALB, GBS, Sinopoly and A123 (now Lithium Werks / Valence / Super B)

Lithionics definitely states which of these they source from, might be Winston? Even if they didn't, I put a lot of trust in Bruce from my dealings and personal contact.

Victron apparently uses K2, which yes would be (a bit) suspect to me

but again, I have a lot of trust in the company's desire to preserve their reputation long-term. Same with Mastervolt, but don't have much experience with them.

Trojan, not as much, and BTW they definitely use K2 as the OEM for their new Trillium line, too early to say there.

So yes, be very skeptical is a wise default, until "proven" otherwise over time.

IMO these packaged systems are not as good value as buying quality bare cells, as long as you're willing to get the gear and put in the time and effort required to learn care for them well.

Not a suitable approach for the average boater though for sure.

[Maine Sail]

Originally Posted by donradcliffe

The 200 ahr drop in that I am working on now (Their Rev H) has a BMS with a parasitic load of 3 amps FOR EACH BATTERY, and that was down from their original Rev which had 5 amps. Theses batteries would quickly die with no load, and the whole RV would go dark in 2-3 days. They had multiple charge sources (inverter/charger, solar, and big alternator, which were not set up with the proper charge voltages, and the HVC would again make the RV go dark.

We had a customer swing though here with the EcoTek batteries who was having difficulty. The basic design was okay, metal box, decent cell compression/restraint, dual bus - charge and load, heavy duty busbars, CALB SE200AHA cells and contactors as opposed to FET's. The physical location of the batteries & voltage regulator on the RV was pretty ridiculous but interior space was already maximized. We don't really do RV's, but he plead a good case. We found the parasitic draw of the BMS was low, very low, when the batteries were switched off. But when turned on, and he had two banks, the four contactors, two in each battery, were pretty greedy.

We found the implementation/installation, initial charging voltages, and absorption duration, had been, according to him, apparently murdering quite a few batteries. He had pages of documentation on failures of these batteries in RV's such as his.

His BMS disconnecting, I suspect, was also responsible for three alternator failures (the BMS did not have the ability to cut off the regulator prior to a disconnect). This owner had been through three alternators and was on his second replacement bank but the manufacturer of the RV, according to him, fought him tooth and nail.

In the short time we had to work on the RV, we tweaked the cheap solar controller & MC-614 and suggested he talk to the RV maker about inverter/charger (looked like the same manufacturer that Sterling Power uses) as he claimed it was holding a 4 hour absorption on each charge cycle and had been set to 14.6V. We left it on a setting which I believe was 14.1V and 13.4V.

I later received an update from him saying that the new charging recommendations, from the actual battery manufacturer, for the MC-614 were near exactly how we had programmed it here. They have been working fine ever since. We never did hear where he wound up for an inverter/charger but the first one had literally melted itself trying to charge the bank.

The battery manufacturer today recommends bV of 14.0V, b1c of 0.6, Av of 13.8V, A1c of 0.6 and FV of 13.6V. These are the basic regulator parameters for the MC-614 and CALB SE cells.

[anthem00]

Originally Posted by Maine Sail

His BMS disconnecting, I suspect, was also responsible for three alternator failures (the BMS did not have the ability to cut off the regulator prior to a disconnect). This owner had been through three alternators and was on his second replacement bank but the manufacturer of the RV, according to him, fought him tooth and nail.

Mainsail - thanks for the information. I know everyone talks about cutting the regulator before cutting the bank load. 'm curious - I spoke with Balmar awhile bank about this, and I remember succintly asking the the question of how long before B+ or power (brown or red) cut off to the regulator before its safe to cut battery load from the alternator circuit. They told me a minimum of 6 seconds, and I should be good at 10 before the diodes/magnets would be safe from a disconnect. I havent seen this time delay mentioned but wanted to throw that out there. . Maybe its "safe enough" to cut and then break instaneously, but balmar is saying they want to see at least a 6 second delay before the break. I've programmed that in to mine, but just wondering whether others are properly accounting for that time delay.

[nebster]

The ecotrek setup was a real disaster in many ways. Part of the problem with it is that there were many implementations, including of the battery pack internals.

The earliest packs had super high parasitic draw, probably because of big, inefficient single-coil contactors.

All the packs have the problem that the dual-bus design is fundamentally at odds with a two-way device, namely an inverter-charger.

But the worst thing of all was the mistake of putting multiple dual-bus batteries in parallel and allowing/encouraging the user to switch on or off the batteries individually. Once any one of those batteries gets out of sync with the others in terms of SOC, then the paralleled-dual-contactor design falls over: a battery going high opens its charge bus relay, but leaves its discharge relay connected. That discharge relay is paralleled with all the other discharge relays... but on the other happy batteries, their discharge relays and charge relays are still both closed. So the charger can continue to charge the battery whose BMS has tried to shut it down, by feeding back through the other batteries' busses.

A similar problem exists on discharge.

And, of course, there were a litany of other amateur-hour mistakes made on these high-end motorhomes. Solar panels wired in series and shading each other on the roof, cheap chargers misconfigured, and so on.

On many models, a lead-acid battery was added in parallel with the lithium batteries to try to put a stop to the BMS-based destruction of the alternators (because the regulator was not implemented/programmed correctly, as MaineSail describes). That lead acid buffer battery wreaked havoc with the UI for the power system, keeping things running when you'd expect them not to, underpowering a big inverter with a tiny battery for a few minutes at shutdown, and on and on.

Basically the system was an absolute boondoggle made only slightly better through some iterative improvements in a few of the components. It is a textbook demonstration of how Not to build a mobile power system.

[Maine Sail]

Originally Posted by anthem00

Mainsail - thanks for the information. I know everyone talks about cutting the regulator before cutting the bank load. 'm curious - I spoke with Balmar awhile bank about this, and I remember succintly asking the the question of how long before B+ or power (brown or red) cut off to the regulator before its safe to cut battery load from the alternator circuit. They told me a minimum of 6 seconds, and I should be good at 10 before the diodes/magnets would be safe from a disconnect. I havent seen this time delay mentioned but wanted to throw that out there. . Maybe its "safe enough" to cut and then break instaneously, but balmar is saying they want to see at least a 6 second delay before the break. I've programmed that in to mine, but just wondering whether others are properly accounting for that time delay.

If you cut the regulator brown wire there is a brief delay in shutting the reg down, as it saves some stuff to memory. If you cut regulator B+ the field shut down is instantaneous. Not field power, no blown diodes. That said your charge cut should ideally be occurring before the bank is open circuited by the main contactor.

[anthem00]

Originally Posted by Maine Sail

If you cut the regulator brown wire there is a brief delay in shutting the reg down, as it saves some stuff to memory. If you cut regulator B+ the field shut down is instantaneous. Not field power, no blown diodes. That said your charge cut should ideally be occurring before the bank is open circuited by the main contactor.

I was always under the impression that the brown wire was a orderly shutdown and the power wire (red) was instantaneous. But Balmar techs said, even the power to the regulator, you need to wait 6 seconds to be safe and avoid damaging the alternator.. . Thats what I was getting at. So i've programmed my delay to be 10 seconds upon receiving a HVC before disconnecting the charge bus.