Lithium battery & component suggestions?

[twistedtree]

TF is fortunate to have TT report on his findings. Seldom will you find in the boating press or online sources an individual who has the budget, business and proven technical background to jump into the middle of today’s technological Marine offerings as it applies to instrumentation and energy arenas.

Adulation aside for a moment TT, will your new build be initially set up to minimize generator run time by utilizing a Li package along the lines of this thread? If so the need for a solar array comes into question whether initially or as an after addition.

Wow, you made me blush.


The N68 will also have LFP, assuming this whole deployment goes well. The requirements for the boat and pretty close to what I'm doing now, though not exactly the same. Reliability and spares availability is more of an issue with the boat.


The plan is to keep gen run time down, and one of the attractions of LFP is the quick recharge time. Interestingly, LFP actually can't be charged with as high a current as an equivalently sized AGM bank. Manufacturers prefer C/3 to C/4, so 25 to 33A per 100ah of capacity. C is the max allowed, so 100A per 100ah of capacity. I'll be pushing it a bit with 400A going into a 720ah bank, which is C/1.8. The real shortening of the charge time comes from not having an absorb state with a tapering current. That adds 2-3 hrs for a full recharge no mater how large a charger you have. I expect the LFP bank to fully charge in 2 hrs, which will be awesome.


As for solar, yes it's planned right from the start. A friend of mine is doing the same thing and is about a year ahead of me, but he thinks he can get over 1200W on the hardtop.

 

[john61ct]

Those rates are no issue, long as the wiring is sound, check for heat is all.

1C is actually conservative for LFP, while most AGM won't accept that long at all if ever

 

[Delfin]

Wow, you made me blush.


The N68 will also have LFP, assuming this whole deployment goes well. The requirements for the boat and pretty close to what I'm doing now, though not exactly the same. Reliability and spares availability is more of an issue with the boat.


The plan is to keep gen run time down, and one of the attractions of LFP is the quick recharge time. Interestingly, LFP actually can't be charged with as high a current as an equivalently sized AGM bank. Manufacturers prefer C/3 to C/4, so 25 to 33A per 100ah of capacity. C is the max allowed, so 100A per 100ah of capacity. I'll be pushing it a bit with 400A going into a 720ah bank, which is C/1.8. The real shortening of the charge time comes from not having an absorb state with a tapering current. That adds 2-3 hrs for a full recharge no mater how large a charger you have. I expect the LFP bank to fully charge in 2 hrs, which will be awesome.


As for solar, yes it's planned right from the start. A friend of mine is doing the same thing and is about a year ahead of me, but he thinks he can get over 1200W on the hardtop.

With 70% discharge and 400 amp charging, you'll be full in around and hour and a quarter. While CMS writes he prefers lower charging capacity than the 1.8 you'll have, my guess is that as long as charging ceases at the target voltage to achieve 98% or so of capacity there would be no problem. Just curious, but what target voltage does Calb relate to 98% charged? For my cells it is 28.8 volts, at which point the acceptance rate plummets. I generally disconnect from the charging source a bit before that.

 

[twistedtree]

Those rates are no issue, long as the wiring is sound, check for heat is all.

1C is actually conservative for LFP, while most AGM won't accept that long at all if ever

Interesting. That's counter to all the LFP spec sheets that I've read. 1C is pretty uniformly the max charge rate that I've seen. Discharge rate can be up to 2C.


Where are you seeing higher charge rates allowed? Is it specified by any vendors, or is it just a practice that some people are doing, presumably without noticeable adverse side effects?

 

[twistedtree]

Just curious, but what target voltage does Calb relate to 98% charged? For my cells it is 28.8 volts, at which point the acceptance rate plummets. I generally disconnect from the charging source a bit before that.

Documentation is sparse. But they say you can charge at 3.6V per cell until current drops to 0.05c. Float voltage is 3.40V. Recharge should start at 3.1V

 

[Wil]

Great summary--thanks much.

I bought them from Electric Car Parts Company in Utah. Warranty I think is 1 year, FedEx freight is the worst of them all and I won't buy anything anymore that ships via them...…
…...Yes, anyone can buy these, though availability comes and goes, and the selling "dealers" are tiny one man shops. It's very, very far from a mature supply chain.

 

[Wil]

BTW, TT has a really good article about power usage on his website now.

 

[Delfin]

Documentation is sparse. But they say you can charge at 3.6V per cell until current drops to 0.05c. Float voltage is 3.40V. Recharge should start at 3.1V

I'd be nervous about floating them at all, but if so, 3.35V seems to be what is recommended. You'll only see the 3.6V for the last couple of minutes before acceptance drops below .05c. Mine get to 3.4V and pretty much stay there until the acceptance rate falls off a cliff. Then voltage rises very, very quickly to the rate I have bulk set at, but acceptance rate approaches zero pretty quickly at that voltage.

For what it is worth, here is Lithionics correlation of voltage to capacity during charging:

3.9V, high voltage cutoff with Lithionics BMS
3.65V, recommended voltage to ensure 100% SOC
3.6V, minimum charging voltage to ensure 98% SOC
3.5V, minimum charging voltage to ensure 90-95% SOC
3.4V, average voltage you'll observe during bulk charging, regardless of what the upper limit of max charge voltage is, as long as it is > 3.4V
3.35V, float voltage to maintain 100% charge. As noted, I see no point in this.

And after charging but rested <8 hours, and without a load:

3.4V, resting voltage of a fully charged battery
3.33V, SOC >80%, but not 100%
3.3V, SOC somewhere between 30% and 80% of charge
3.2V, SOC <30%
3V, SOC approximately 10%, reserve voltage cutoff with BMS
2.8V, dead battery, recharge immediately. Low voltage disconnect with BMS

I'm considering installing a voltage switch that activates a noise maker when pack voltage reaches 28.4V. Nice reminder on a boat to disconnect the charging source from the house bank and run on the starter bank.

 

[john61ct]

Just meant that your

> 400A going into a 720ah bank, which is C/1.8

will not in itself do harm, long as you have good overtemp protection.

.4C is routine.

Now, for longevity, I would use MC-614 to limit to even a bit less.

And while you could go to 3.50Vpc at such a high rate (zero Absorb time, just stop, no Float)

Personally I'd keep to 3.45Vpc as usual with lower (solar) rates.

Let the cells rest at room temperature for 3+ hours with no load then measure.

If at 3.40Vpc they are "daily use full"

If you want to go a bit higher (why would you want to?) then don't increase the V setpoint, but use endAmps, say .025C, to give a little Absorb time.

I only do this when calibrating the BM to 100%, or for testing / maintenance routines.

Of course you don't see this in vendor specs, why would they want us to get many thousands more lifetime cycles? 8-D tinfoil hat

Actually more likely it's the dominance of high-C rate EV usage, get every last bit of mileage-range out for marketing purposes.

 

[twistedtree]

Those voltages are what CALB has in their book. My plan is to start with pretty conservative numbers and work from there.


John61, I think you are correct that the vendor specs and recommendations are heavily influenced by the primary consumers of these cells, namely the EV market. Highest capacity sells, and that means charging to max capacity, and discharge as low as you can without damage. We have the luxury of more conservative use

 

[john61ct]

Conservative in this case means staying away from the shoulders

Lower V at the top, higher at the bottom.

10%-97% rather than 5%-99%.

Shoot for 20+ years, calendar aging wins rather than cycles. . .

IMNSHO 8-D

 

[john61ct]

here is Lithionics correlation of voltage to capacity during charging

At two decimal points, you'll get some variability depending on charge rates.

AH-counting is better for measuring SoC on the way down.

On the way up, all that really matters is stopping at "your defined" 100%.

 

[Delfin]

At two decimal points, you'll get some variability depending on charge rates.

AH-counting is better for measuring SoC on the way down.

On the way up, all that really matters is stopping at "your defined" 100%.

John, 100% charge is neither arbitrary nor personally defined as you suggest, unless you don't care whether you kill the battery or not. For LiFePO4 batteries it is 3.6 volts per cell, as you will find in all references to charging this chemistry.

 

[twistedtree]

John, 100% charge is neither arbitrary nor personally defined as you suggest, unless you don't care whether you kill the battery or not. For LiFePO4 batteries it is 3.6 volts per cell, as you will find in all references to charging this chemistry.

I think he's just referring to how full one chooses to charge their battery, vs how full it can possibly be charged. Since it's quite common to have a charge/discharge protocol that stops short of full possible charge, and stops short of full possible discharge, he is calling these "user-defined" full and empty. From this perspective, the battery is 100% full when the battery is as full as your charge protocol is ever going to charge it.

 

[Delfin]

I think he's just referring to how full one chooses to charge their battery, vs how full it can possibly be charged. Since it's quite common to have a charge/discharge protocol that stops short of full possible charge, and stops short of full possible discharge, he is calling these "user-defined" full and empty. From this perspective, the battery is 100% full when the battery is as full as your charge protocol is ever going to charge it.

Perhaps, and if so, I like that definition of 100%. On that basis I have reached 100% of my ideal weight, even though I am still 20# heavier. I can go with this....

I think you will find that as far as Li batteries go, there is a point where they are full, regardless of what voltage you charge them at. Perhaps that is why CMS writes:

"An LFP cell can be fully charged as low as 3.4VPC to 3.42VPC if the voltage is held long enough. " He goes on to note that the only difference between full at a lower voltage and full at a higher charge voltage is how long you spend towards the end of the charge cycle. But "full" is "full" with Li batteries. Although perhaps I am misunderstanding what CMS is saying...

 

[BruceK]

How do vanadium batteries compare to lithium?

 

[Delfin]

How do vanadium batteries compare to lithium?

https://www.sciencedirect.com/science/article/pii/S1876610216310566

 

[john61ct]

John, 100% charge is neither arbitrary nor personally defined as you suggest, unless you don't care whether you kill the battery or not. For LiFePO4 batteries it is 3.6 volts per cell, as you will find in all references to charging this chemistry.

1. Are you defining Full at 3.60Vpc as

a. charging setpoint ?

get there and stop ?

some endAmps spec?, or

resting voltage after isolation?

2. Actually allowing your bank to get to that point is "damaging" in the sense that you will only get rated cycle lifetime, a small fraction of what you could get if keeping below that top shoulder.

Therefore I will never go that high in normal use.

Pretty silly to use a definition of Full that I will never get to.

Instead, **my** precise "100% Full" is 3.45Vpc until endAmps of .025C

in daily use, 3.45Vpc as a "just stop" setpoint, maybe 3.50 at very high rates, 3.40 at tiny solar rates.

My 0% is 2.975Vpc.

You can set yours for your bank as you like.

 

[twistedtree]

2. Actually allowing your bank to get to that point is "damaging" in the sense that you will only get rated cycle lifetime, a small fraction of what you could get if keeping below that top shoulder.

Just curious... has this actually been demonstrated to be true? It's a long and arduous process to demonstrate cycle life on any battery, let alone one that supposedly has 3000 to 5000 cycles in it. I think one of the big challenges with LFP is sorting opinion from demonstrated cause and effect. It's not unlike oil where someone believes the reason their engine hasn't failed is because they use brand X of oil. As you well know, you can't prove a negative. Now I have no reason to disagree with what you are suggesting, but also no reason to agree. I simply don't know.

 

[Wxx3]

Delfin,

Thanks for your descriptive explanations. You certainly took the magic out of them for me.

I like your simple charging pattern.

Richard

 

[Delfin]

Just curious... has this actually been demonstrated to be true? It's a long and arduous process to demonstrate cycle life on any battery, let alone one that supposedly has 3000 to 5000 cycles in it. I think one of the big challenges with LFP is sorting opinion from demonstrated cause and effect. It's not unlike oil where someone believes the reason their engine hasn't failed is because they use brand X of oil. As you well know, you can't prove a negative. Now I have no reason to disagree with what you are suggesting, but also no reason to agree. I simply don't know.

Lithionics alleges that DoD determines cycle life. At 50% or so, they claim many, many thousands of cycles, assuming the batteries are charged correctly. If so, I guess I'll have to put them in my will.

 

[Delfin]

Delfin,

Thanks for your descriptive explanations. You certainly took the magic out of them for me.

I like your simple charging pattern.

Richard

Thank you Richard.

The light bulb went off for me when I tried to program the Balmar to match what I understood were the best parameters for charging and had to make compromises. For example, when charging, the bank voltage climbs so quickly and the acceptance rate falls so quickly that having no absorption phase made sense, but the Balmar has a minimum absorption time of 18 minutes. The way around that is just to disconnect the system from charging and run off the starter bank (in my case lead acid) and the alternator. Based on CMS's admonitions, no float is ever desirable, so the only time my batteries are connected to anything is when they are being charged after depletion or when powering the house loads. In other words, they are a fuel tank. Fill them up, shut off the hose, use the fuel, then refuel. Everyone struggling to make the charging system work just like it does with other chemistries seems far harder and more of a hassle than following what I am doing.

The controversy seems to be what represents "full". Since these batteries are so efficient, full to me means when they won't accept anymore current, or much. So I disconnect usually around 30 amps CAR. You will literally see a drop off of the CAR from 60 amps to 10 in about 2 minutes, so "full" seems pretty quantifiable for me.

We're heading out today and the bank is around 250 amps down out of 600. To experiment, I've detuned the bulk voltage on the Balmar to 27.6 volts and I want to see what difference other than time to recharge this adjustment makes, and whether the bulk CAR is affected at the lower constant charge voltage than the one I have been using - 28.8 volts. Since you don't see the bulk voltage at the batteries until the end of the charge cycle, it may have no impact but I am curious to see.

 

[Delfin]

First observation: When the Balmar and charger on the Trace 4000 are set to 28.8, the CAR is 205 - 213 amps when underway at 1400 rpm. Having reset both to what some are recommending - 3.45V per cell or 27.6, the CAR has dropped to 160 - 175 amps. Since at 28.8 volts and 200 amps I am recharging at what CMS seems to say is an optimal fractional charge rate of .35C, other than spending more time to recharge, I can't see how this regime is going to have the slightest impact on longevity. But, I'd like to see if the taper at the end of the cycle is also lengthened, which I suspect it will be.

p.s., I may not have cell coverage at the end of charging, and if so may not for a couple of weeks, so a final report may not happen for awhile.

 

[Delfin]

Quick second report: At such a low bulk voltage setting of 27.6, both the Balmar and Trace just went into the float setting, so charging under this regime just went from super quick to very, very slow. I've taken the Trace off line and repowered the Balmar to bulk, and we are now charging a 96 amps, or about 87% of what the engine + alternator would charge at if the Balmar were set to 28.8, which by the way is Lithionics recommendation.

 

[sunchaser]

Delfin, your charging analysis makes sense. Once a full season of cycles are accomplished the real world findings will prove helpful for those of us following along.

The discussion on number of cycles seems much like the question as to how long a Marine diesel performing recreational duty will last. To me the question is more how long will the Li energy overall system remain impervious to the elements, low grade wiring, sloppy installation, inadequate monitoring systems, user inattention and an overly complicated/baffling setup will determine the number of years a Li energy system will perform. Cycles be damned.

 

[Delfin]

Delfin, your charging analysis makes sense. Once a full season of cycles are accomplished the real world findings will prove helpful for those of us following along.

The discussion on number of cycles seems much like the question as to how long a Marine diesel performing recreational duty will last. To me the question is more how long will the Li energy overall system remain impervious to the elements. Low grade wiring, sloppy installation, inadequate monitoring systems, user inattention and an overly complicated/baffling setup will determine the number of years a Li energy system will perform. Cycles be damned.

I expect you're right Tom. One of the things I have been checking is battery temperatures when bulk is set to 28.8. Never above 85 degrees, which is solely because they are about 4 feet away from the CAT, but behind a styrofoam wall with cooler air being blown onto them. So I assume no heat build up from charging.

Third report: So about 20 minutes into this trip, the acceptance rate has dropped to nothing because the Balmar is now in float again. The only conclusion I can make is that from the armchair, 3.45volts charging might make sense, but in the real world, none whatsoever, at least with my charging equipment, which CMS says is the best there is. I guess I could set float at 27.6 volts, but then that risks cooking the batteries if I forget to disconnect the charge source since they'll fry at that voltage as well as at 28.8. I may be missing something, but I'll reset everything back to 28.8 volts and get on with life, as well as charged batteries.

 

[diver dave]

.. because they are about 4 feet away from the CAT, but behind a styrofoam wall with cooler air being blown onto them. ...

Foamed plastic either a FR grade (fairly rare for common styrofoam) or contained behind FR material (metal, etc). I've seen some genset sound walls homebrewed in the ER that fail survey.

 

[twistedtree]

It's interesting to see Delfin's charge voltages and compare to what other manufacturers recommend. Delfin, do you happen to know who's cells are in the Lithionics product? I wouldn't expect much difference in recommended charging between vendors, assuming LFP is both cases, but you never know.


You are charging at/to 28.8V, which is interesting. One system I have looked at a bunch, mostly because a friend is building one, and their Allow To Charge signal shuts off at 28.1V. When that signal goes away, all charge sources are supposed to shut off. And they recommend charge settings of less than that so charge sources switch off on their own, and ideally the Allow To Charge signal never shuts off.


Float is another area where I scratch my head. I've heard people say to charge to a certain voltage/return current, then STOP. And I have heard people also say not to Float. All this only makes partial sense to me. Maybe it's just unclear language - I'm not sure.


What makes sense to me is that you want to charger at an elevated voltage and positive current up to some point, then stop charging. But to me that doesn't mean no Float, it means no Absorb or otherwise holding the batteries at an elevated voltage for a prolonged period of time as you would with LA. That's the "Stop Charging" part that makes sense.


But float I think is still a very desirable thing, assuming to have the voltage set right. With Float correctly set to the battery's resting voltage when near full charge, the zero current will flow, and there will be no charging. However when loads get switched on and the battery voltage is otherwise drawn down a bit, the charger kicks in to carry the load. To me this is very desirable to prevent gratuitous cycling of the batteries. Otherwise while on dock power, the chargers would stop, you would run on batteries until they are drawn down, then you would go through another charger cycle, stop, and repeat again. Float prevents this, and lets the chargers carry the load when a power source is available.


So to me, "Stop Charging" when full, and "No Float" are two very different things. This first is good, and the second not so much.

 

[john61ct]

their Allow To Charge signal shuts off at 28.1V

÷8 =
3.5Vpc, good, not too high.

> Float is another area where I scratch my head. I've heard people say to charge to a certain voltage/return current, then STOP. And I have heard people also say not to Float.

Those are two different ways of saying the same thing wrt Float.

But you are correct, it also means no holding Absorb time, that concept is needed for lead, because those banks need to get to Full.

LFP does not, and in fact it is risky, only go there when a consumer load is active ready to start discharging.

> But float I think is still a very desirable thing, assuming to have the voltage set right.

Not for an LFP bank.

Just isolate the bank from the charge source, no need to turn that off.

A cheap lead batt can act as a buffer / load dump, power always-on loads dockside.

 

[twistedtree]

÷8 =
3.5Vpc, good, not too high.

> Float is another area where I scratch my head. I've heard people say to charge to a certain voltage/return current, then STOP. And I have heard people also say not to Float.

Those are two different ways of saying the same thing wrt Float.

But you are correct, it also means no holding Absorb time, that concept is needed for lead, because those banks need to get to Full.

LFP does not, and in fact it is risky, only go there when a consumer load is active ready to start discharging.

> But float I think is still a very desirable thing, assuming to have the voltage set right.

Not for an LFP bank.

You will have to explain this to me. A connected charger floating at the battery's open circuit resting voltage is not charging the battery. It's no different electrically than being disconnected. Zero current flow. This is how Victron tells you to set up their LFP systems. Recommended float is 3.3V which is the OC voltage for about 80-90% SOC.


Delfin took an approach similar to what you are describing, only using his LFP batteries when at anchor, and using a small LA bank the rest of the time. I'd like to eliminate that complexity. Also, with solar able to carry most if not all the daytime loads, it makes sense to use it rather than cycling the LFP bank. Gotta use that sun when it shines,