Drop in LiFePO4 batt

[Delfin]

Better to just use cell "balancing", IMO calling it equalization may confuse some.

And let go of the voltage/SoC idea, not applicable in this case.

When float is set lower than actual, then that should functionally be the same as Off, no current flow into the battery, but any loads carried by the charge source. Only extremely high-amp loads should cause any voltage drop.

Also, there is no "ideal SoC" except for when not cycling, which is way down near the bottom of what the BMS allows. While cycling you will go between "full" and "empty" as defined by the BMS, passing through 60% twice each cycle.

Do you know if your BMS initiates a new bulk/absorb charge cycle based on SoC via coulomb-counting?

While Li battery voltage doesn't taper like LA batteries, it does taper, so I'm not sure it is correct to say that voltage/SoC is not applicable. Just different, as the manufacturer says.

And no, the BMS doesn't have any impact on the behavior of the charging sources. It just sits between the battery and the load and is best thought of as part of the battery. All it can do to the charging source is disconnect while allowing the load to continue to draw if it is a two channel, or disconnect both charging source and loads if a single channel.

 

[Art]

Delfin, 61ct, tt, FW... and others... have been and are continuing to cue me into the hows, whys and what-ifs regarding improved batts and chargers. I thank you all.


Eventually, with appreciation owed to boaters like you guys, Li batts may become the boating norm. That is if some other even better electrical storage apparatus does not rush onto the scene via ever quickening High-Tech methods.


After a couple days out away from dock: Photo shows last Sunday's [7/9] multi meter's recorded 13.20 state of charge on our house bank [four - g31, parallel-link, la, deep cycle batts; installed Spring 2010] a few minutes after shutting down from a one hour morning run of gen set. I then left batts with no draw for two hours; their leveled off charge was 12.65. Soooo... you can see why I currently like to keep the DC power sysytem in complete KISS mode regarding being aboard our simple, little Tolly tri cabin.


We hauled anchor Monday 7/10 at 7 AM. Cruised a short hop back to our dock. As usual, we did not hook into shore power so there was no charger throwing current into the battery bank. Took us a couple hours to haul tow behind runabout into storage, clean the heads/galley/state room/salon etc. and place canvas in position. During that time we used DC power sparingly ,but, the batt bank power was somewhat used for sink water, toilet clean/flush and ceiling lights several times. Just before departing I checked the multi meter again... reading was 12.6.

Will be at least a few weeks before we can return for more boat fun. Same length of time away from boat that our current ongoing obligations have been keeping us. Understand... I never leave boat plugged into dock; for many reasons... so the batt bank that I isolate via Perko switches will have no charge into nor draw out of them.

I anticipate, such as is always the case, when I get aboard the multi meter reading will be 12.5 +/-. From that point another grand time aboard Tolly is about to begin!

Happy Boat-Batt-Charge Daze! - Art

 

[Delfin]

Thanks John, picture of one of two batteries replacing 12 below (assuming I am using Imgur correctly). 13" x 13.5" x 24", 135 pounds.

And apparently not, since the image doesn't post. Oh well. Trust me...they're big.

 

[john61ct]

While Li battery voltage doesn't taper like LA batteries, it does taper, so I'm not sure it is correct to say that voltage/SoC is not applicable. Just different, as the manufacturer says.

But at any given SoC the voltage will drop in inverse proportion to the load; a light load will keep the V curve very flat, a very heavy load will drop it much lower and more quickly.

But my main point was actually that you seemed to think having the float set to a particular voltage would mean that charge source "maintains" the bank at a particular corresponding SoC.

So I was letting you know it doesn't work that way during normal usage cycling.

And when not being cycled, a LFP bank should not be kept anywhere near full.

 

[Delfin]

Finished the installation of a 600 aH 24v LiFePO4 bank. Traded around 1400 # of AGM batteries for 270# of batteries with equivalent usable capacity. I'll be testing capacity over the next few days, but I have learned a couple of things.

For all who use the Balmar 612 or 624 regulator, I discovered something I didn't know, even after a decade or so of using their product. Specifically, their documentation doesn't provide a critical piece of information on how they measure the time settings. For Li batteries, the time charging should occur after bulk voltage is reached, or for how long absorption voltage should be maintained is as short as possible. If you read the manual, it states that time can be set in 6 minute intervals, but it turns out that a value of 1 for these time settings = 6 minutes, 2 = 12 minutes, etc. So, if you think you are setting the duration of time to hold at bulk for 30 minutes by entering 30 into the program, you are actually setting it for 180 minutes. I spoke with the tech support for Balmar and they sheepishly admitted they had wanted this changed for years, but it hadn't happened yet. Ah well.

 

[john61ct]

Good info, and nice job!

Nitpicky I know, but Bulk stage has no voltage setpoint, that is the Absorption voltage setting.

And the regulator needs to be told how long to hold Absorption, it has no control over the length of Bulk stage.

I think Balmar may be one of the vendors that misuses those terms.

 

[Delfin]

Good info, and nice job!

Nitpicky I know, but Bulk stage has no voltage setpoint, that is the Absorption voltage setting.

And the regulator needs to be told how long to hold Absorption, it has no control over the length of Bulk stage.

I think Balmar may be one of the vendors that misuses those terms.

Have you ever programmed one of these?

The Balmar has a set point voltage for bulk (bv) that determines target voltage, as well as a time setting to hold the battery at bulk voltage after the battery reaches that target value (b1c). Absorption is required to be lower than the bulk set point by .1 volts, and the time to hold at absorption can also be set, just like bulk.

 

[john61ct]

I found one of Maine Sail/CMS' many screeds on Balmar's misuse of the terminology

http://www.cruisersforum.com/forums/showthread.php?p=2071974

 

[Delfin]

I found one of Maine Sail/CMS' many screeds on Balmar's misuse of the terminology

Balmar 614 "bulk charging" - Cruisers & Sailing Forums

"Bulk is constant current or a transition voltage point where we end bulk/CC and begin absorption/CV. " - Maine Sail

In other words, a set point of the voltage at which the transition occurs. And, in the case of the Balmar, there is a time setting that determines how long the regulator will maintain the set point bulk voltage in a CV state before transitioning to an absorption voltage, also CV, that has to have a different set point voltage and different time setting before shifting to float.

I would also point out that when someone refers to "constant current", that is also a misnomer. The current for all chemistries will start as high as the charging source and regulator settings will allow, but it isn't constant. What is constant is the target voltage Balmar chooses to call the bulk voltage target that is set when programming the unit, and the only sense in which the current is constant is that it is constantly changing as voltage rises to the target value defined in the regulator settings. So I guess inexactness of terminology is a generalized problem.

 

[kev_rm]

Be curious to know how many full cycles you get. That's my main concern - for most applications dollars per lifetime kw/h is what matters in the end, and its hard for lithium to win that contest today.

 

[john61ct]

No, for relatively gentle House usage, if properly cared for, LFP comes out cheapest overall long-term.

But risky due to early stages of adoption, such large investment up front.

 

[Delfin]

Be curious to know how many full cycles you get. That's my main concern - for most applications dollars per lifetime kw/h is what matters in the end, and its hard for lithium to win that contest today.

I think Lithium wins that contest, at least it will if the lifetime cycles reported are close to correct. Maine Sail seems to think that based on his testing, 2,000 cycles looks attainable, and I suspect that if the batteries are not discharged to 80% 100% of the time, that number can be stretched. But assuming 2,000 cycles at 80% discharge, that is double the cycles of carbon foam batteries and four times the best quality AGM batteries.

Given just the cost per cycle, Li and Carbon come out about the same, but it sure looks like Li have the edge in how they deliver the current stored, having little to no reduction in voltage until nearly empty. AGM turns out to be the most expensive on a cost per cycle.

In my case, here's the logic:

Amount of usable current desired: 480 amps @ 24v
Size of bank needed for Li: 600 amps @$15,000 - 80% discharge
Size of bank needed for Carbon Foam: 800 amps @ $7,700 - 80% discharge, recharge to 80% full
Size of bank needed for AGM: 1000 amps @ $4,800 - 50% discharge

Li cost per cycle @ 2000: $7.50
Carbon foam per cycle @ 1000: $7.70
AGM per cycle @ 500: $9.60

The AGM also require around 4 times the generator time to recharge fully so that you can get 500 cycles. If the Li get the 3,000 cycles the manufacturers say they can get at 80% discharge, they are more favorable.

 

[twistedtree]

Finished the installation of a 600 aH 24v LiFePO4 bank. Traded around 1400 # of AGM batteries for 270# of batteries with equivalent usable capacity. I'll be testing capacity over the next few days, but I have learned a couple of things.

For all who use the Balmar 612 or 624 regulator, I discovered something I didn't know, even after a decade or so of using their product. Specifically, their documentation doesn't provide a critical piece of information on how they measure the time settings. For Li batteries, the time charging should occur after bulk voltage is reached, or for how long absorption voltage should be maintained is as short as possible. If you read the manual, it states that time can be set in 6 minute intervals, but it turns out that a value of 1 for these time settings = 6 minutes, 2 = 12 minutes, etc. So, if you think you are setting the duration of time to hold at bulk for 30 minutes by entering 30 into the program, you are actually setting it for 180 minutes. I spoke with the tech support for Balmar and they sheepishly admitted they had wanted this changed for years, but it hadn't happened yet. Ah well.

Me wants one too.......

Re the Balmar, there is a barely perceptible decimal point between the last two digits for those time settings, so what it's showing is decimal hours. So 0.1 hrs is indeed 6 minutes, 0.3 would be 18, and 1.1 would be 1hr and 6 minutes. I'm sure it made sense to the engineer who implemented it...

 

[twistedtree]

What are the final settings you came up with for the Balmar MC? And what about other charge sources?

 

[CMS]

But at any given SoC the voltage will drop in inverse proportion to the load; a light load will keep the V curve very flat, a very heavy load will drop it much lower and more quickly.

This is an actual discharge curve at 2 min intervals...
Link to large version HERE

This discharge curve may be helpful to some. Of interesting note is how little capacity is physically stored above 3.32VPC or 13.26V as a 12V nominal pack..

In the graph the 400Ah rated battery has only delivered 2.83Ah's by the time the voltage curve levels out at 3.32VPC under a 30A load. When these cells are at 100% SOC they can have a resting voltage of about 3.38VPC to 3.40VPC or 13.52V to 13.60V, for a 12V nominal pack, but there is really very little stored energy between 3.4VPC and 3.32VPC about 0.65% of Ah capacity to be exact.

On this pack the stored energy between 3.4VPC and 3.32VPC was 2.83Ah's. This is the rather abrupt near vertical portion of the blue voltage line at the very beginning of the curve. Also, at low rate discharges, such as this .075C discharge rate, the curve is more gradual and it is not until about 2.9VPC that the voltage starts to hit the knee and drop rapidly, in a near vertical fashion.

On this 400Ah rated pack, at a 30A or .075C discharge rate, the working voltage range between:

99.3% SOC and 0% DOD = 1.66V
99.3% SOC and 50% DOD = 0.26V
99.3% SOC and 80% DOD = 0.51V

I also included the loaded voltage points as well as every 10% of capacity in delivered energy.

 

[CMS]

I would also point out that when someone refers to "constant current", that is also a misnomer. The current for all chemistries will start as high as the charging source and regulator settings will allow, but it isn't constant.

Correct and why I often refer to bulk as the "constant potential" with solar wind or an alternator. With these devices bulk current will be determined by such things as engine RPM, alternator temp, irradiance, wind etc. but it is still bulk or the maximum the charge source can produce, based on things such as RPM etc. up until the CV transition point.

If we were talking about an AC charge source the current will continue flat or at the same current until the CV set point is reached. Bulk is simply not a constant voltage stage but rather the most the charge source can deliver to the battery... It means the regulator has not been limited by target voltage..

 

[Delfin]

What are the final settings you came up with for the Balmar MC? And what about other charge sources?

The 624 I have is delivering .2 volt more than it thinks it is, and while Balmar has kindly offered to replace it, for now in order to get a final bulk target of 28 volts I've applied the following settings, all shown in Balmar's cryptic abbreviations:

dLc: 30 (delay before ramp up of current)
CL: 28.2 (temperature compensation limit)
bv: 27.8 (bulk voltage set point target)
b1c: 2 (duration at bulk after set point is reached. 12 minutes)
Av: 27.6 (absorption voltage after 12 minutes at bulk set point)
A1c: 2 (duration at absorption set point is reached. 12 minutes)
Fv: 26.4 (float voltage, set below resting voltage so they don't float)
F1c: 2 (minimum time at float, 12 minutes)
AP: 78% (max output of alternator)
FbA: 74% (percentage of alternator output that triggers shift to absorption voltage, after the minimum of 12 minutes. I may have this one set too high, but I am not sure about that....)
FFL: 74% (percentage of alternator output that triggers shift from float back to absorption voltage. Same comment....)

This is based on my understanding of how this device functions, so any corrections from more knowledgeable users would be much appreciated. Engineers should not write documentation....

 

[Delfin]

Correct and why I often refer to bulk as the "constant potential" with solar wind or an alternator. With these devices bulk current will be determined by such things as engine RPM, alternator temp, irradiance, wind etc. but it is still bulk or the maximum the charge source can produce, based on things such as RPM etc. up until the CV transition point.

If we were talking about an AC charge source the current will continue flat or at the same current until the CV set point is reached. Bulk is simply not a constant voltage stage but rather the most the charge source can deliver to the battery... It means the regulator has not been limited by target voltage..

Thanks Rod. Any suggestions on my settings shown in the previous post?

 

[Delfin]

Me wants one too.......

Re the Balmar, there is a barely perceptible decimal point between the last two digits for those time settings, so what it's showing is decimal hours. So 0.1 hrs is indeed 6 minutes, 0.3 would be 18, and 1.1 would be 1hr and 6 minutes. I'm sure it made sense to the engineer who implemented it...

That makes sense, although the support tech gave me a slightly different version amounting to the same thing. He said "ignore the decimal point", although your explanation makes a heck of a lot more sense. Now, wouldn't it have been easy for them to state "time is measured in tenths of hours"?

Naw, too simple....

 

[Delfin]

This is an actual discharge curve at 2 min intervals...
Link to large version HERE

This discharge curve may be helpful to some. Of interesting note is how little capacity is physically stored above 3.32VPC or 13.26V as a 12V nominal pack..

In the graph the 400Ah rated battery has only delivered 2.83Ah's by the time the voltage curve levels out at 3.32VPC under a 30A load. When these cells are at 100% SOC they can have a resting voltage of about 3.38VPC to 3.40VPC or 13.52V to 13.60V, for a 12V nominal pack, but there is really very little stored energy between 3.4VPC and 3.32VPC about 0.65% of Ah capacity to be exact.

On this pack the stored energy between 3.4VPC and 3.32VPC was 2.83Ah's. This is the rather abrupt near vertical portion of the blue voltage line at the very beginning of the curve. Also, at low rate discharges, such as this .075C discharge rate, the curve is more gradual and it is not until about 2.9VPC that the voltage starts to hit the knee and drop rapidly, in a near vertical fashion.

On this 400Ah rated pack, at a 30A or .075C discharge rate, the working voltage range between:

99.3% SOC and 0% DOD = 1.66V
99.3% SOC and 50% DOD = 0.26V
99.3% SOC and 80% DOD = 0.51V

I also included the loaded voltage points as well as every 10% of capacity in delivered energy.

Because of this, a paper I read that measuring capacity of LiFePO4 batteries is best done by Coulomb/amp counting between the knees and voltage near the knees seems to make sense. Figuring out how much capacity is left on these is not straightforward.

 

[john61ct]

I think Lithium wins that contest, at least it will if the lifetime cycles reported are close to correct. Maine Sail seems to think that based on his testing, 2,000 cycles looks attainable, and I suspect that if the batteries are not discharged to 80% 100% of the time, that number can be stretched.

I think 2000 will turn out to be low, only time will tell, and of course only babied banks.

But that babying can include going a lot deeper than 20% SoC, just need to stay away from where the voltage drop steepens. Lower amp discharge rates help.

And avoiding full charging (at over 14V, or even lower V left going too long) is apparently just as important.

Another piece is temperatures, no charging when it gets freezing cold.

 

[sunchaser]

Carl & Rod
Does Balmar have the edge on sensing and load management devices or is there meaningful marine oriented competition to consider?

Peter
Has Nordhavn used battery weight as a stability factor thus necessitating some re-ballasting if lead removed and Li put in its place?

 

[Delfin]

I think 2000 will turn out to be low, only time will tell, and of course only babied banks.

But that babying can include going a lot deeper than 20% SoC, just need to stay away from where the voltage drop steepens. Lower amp discharge rates help.

And avoiding full charging (at over 14V, or even lower V left going too long) is apparently just as important.

Another piece is temperatures, no charging when it gets freezing cold.

Your comment opens another area that seems a bit unknown, although Maine Sail's testing seems to point to some possible conclusions. Specifically, what is the best "exercise regime" for these banks that will keep them "fit" the longest? Lithionics calculates that you can get 35,000 cycles at 5 - 10% DoD. Maybe so, but what is the capacity of the bank if that is only how you use them? And who lives long enough to use 35,000 cycles?

I'm guessing without knowing that running these down 60 - 70% with the occasional DoD to 90%, keeping peak voltage <=28 volts and no floating will result in more cycles than I will ever use.

 

[Delfin]

Carl & Rod
Does Balmar have the edge on sensing and load management devices or is there meaningful marine oriented competition to consider?

Peter
Has Nordhavn used battery weight as a stability factor thus necessitating some re-ballasting if lead removed and Li put in its place?

I wish Ample Power was still around, as they had some kick ass equipment a few years back, or so I thought. I used one of their regulators on my sail boat for 20 years without a hiccup. Beyond that observation, I am relying on alleged familiarity of the Balmar, and its ability to set a whole lot of variables that very few people seem to actually understand until they feel they have to.

 

[john61ct]

Your comment opens another area that seems a bit unknown, although Maine Sail's testing seems to point to some possible conclusions. Specifically, what is the best "exercise regime" for these banks that will keep them "fit" the longest? Lithionics calculates that you can get 35,000 cycles at 5 - 10% DoD. Maybe so, but what is the capacity of the bank if that is only how you use them? And who lives long enough to use 35,000 cycles?

I'm guessing without knowing that running these down 60 - 70% with the occasional DoD to 90%, keeping peak voltage <=28 volts and no floating will result in more cycles than I will ever use.

IMO size for what you need, stay away from extremes, but don't be afraid to exercise the bank much as you like when needed.

When not cycling, store at low SoC and cool.

Other than that enjoy being on the water!

 

[john61ct]

This is an actual discharge curve at 2 min intervals...

Yes your charts are always very informative. In this case, 30A is pretty high for me, but shows effective capacity of 90% is doable, at lower discharge rates will get even more without pushing the shoulders.

For LVD protection, I think I will shoot for 12V as a floor, in addition to derived SoC, actually AH out since full. IOW either condition will isolate the bank from the Loads Buss, and/or turn on the genny if that functionality is in place.

Feedback? (from anyone is welcome)

 

[twistedtree]

Correct and why I often refer to bulk as the "constant potential" with solar wind or an alternator. With these devices bulk current will be determined by such things as engine RPM, alternator temp, irradiance, wind etc. but it is still bulk or the maximum the charge source can produce, based on things such as RPM etc. up until the CV transition point.

If we were talking about an AC charge source the current will continue flat or at the same current until the CV set point is reached. Bulk is simply not a constant voltage stage but rather the most the charge source can deliver to the battery... It means the regulator has not been limited by target voltage..

Another way to describe it is that when in "bulk" mode, the charging source is Current Limited. It is putting out as much current as it can, irrespective of the voltage. And as you point out, "as much as it can" depends on a number of factors, so the actual current can vary. But the device is running balls to the wall and putting out as much as it can.

And in "Absorb" mode, the charger is Voltage Limited. It is putting out as much current as is required to maintain the desired voltage set point.

In reality these devices are all just voltage regulated. It just happens that during bulk charge, the device can't put out enough current to attain the desired voltage, so it's stuck current limited until the target voltage is reached, and which point it can begin actively regulating the current.

 

[john61ct]

In reality these devices are all just voltage regulated.

Many let you set a maximum current limit.

This is to me a critical feature, when

wiring or downstream devices can't handle high current safely

upstream supply doesn't handle being overloaded well

you just want to reserve some total capacity for other purposes (e.g. engine propulsion)

or you just want to be gentler with your bank, say respecting the mfg max charging spec

 

[twistedtree]

Peter
Has Nordhavn used battery weight as a stability factor thus necessitating some re-ballasting if lead removed and Li put in its place?

Not for stability, no. 1000 lbs of batteries in a 130,000 lb boat just isn't significant. I can't imagine any building using an accessory like batteries as part of it's stability calculations.

But swapping them for LiFePO will likely affect trim and require some adjusting of weight distribution. But it's frankly not much more than the difference between a full and empty black water tank.

 

[twistedtree]

Many let you set a maximum current limit.

This is to me a critical feature, when

wiring or downstream devices can't handle high current safely

upstream supply doesn't handle being overloaded well

you just want to reserve some total capacity for other purposes (e.g. engine propulsion)

or you just want to be gentler with your bank, say respecting the mfg max charging spec

Point taken. But in my experience such things are usually accommodated in the initial system sizing. In practice people are usually looking for more change current, not less.

The only cases I've seen for dialing back charge current are 1) to not overload an alternator that is not rated for continuous duty (this is what Balmar's feature is for), and 2) to limit shore power draw when you are stuck with less than you normally want/use.