lithium charging performance

[seattleboatguy]

I have a question for those of you who have made the switch to LiFePo4. I recently added a couple of these batteries and 2 new alternators to my boat. I'm not sure if this new charging system is under performing, or if maybe I had unrealistic expectations to begin with. How do the numbers in the attached graph compare with the amperage you are seeing in your boat? For this test, I first drained the batteries down to 50% capacity, then ran the 2 engines for 20 minutes with no load on the house batteries, then continued the test for an additional 15 minutes with a 24 amp load on the house batteries. Thanks for any feedback.

 

[twistedtree]

You should be getting more alternator output, and it should be more consistent. A few questions:


What are the programmed charging voltages in the MC 618s? Bulk/absorb/float, and absorb time?


Where are the the MC 618 voltage sense wires connected?


BTW, you diagram and chart are excellent and make it SOOO much easier to help with something like this.

 

[twistedtree]

Oh, and what RPM were you running the engines? The alternator output will vary quite a bit based on RPM. And output will drop as the alternators heat up, which will happen quickly at lower RPMs.

 

[DavidM]

Do your alternator regulators have Li specific settings? Li batteries need a constant 14.4 to 14.6 volts to charge adequately. Your graph indicates to me the regulator is following a traditional FLA charge profile.

David

 

[Comodave]

With both an AGM and the lithium batteries on one alternator through the isolator, which charging profile are you using, AGM or lithium. On our last boat I put in a lithium house bank. I had one engine charge the lithium bank and the other engine charge both starting batteries and the genset start battery. That way I could optimize the lithium charging and not effect the other chemistry batteries.

 

[rgano]

An aside: The long article on LifePo batteries recently linked here on TF highly recommended alternator protection circuity in the event the battery management system dumped the load. Do you folks have such?

 

[Comodave]

I had a Sterling Alternator Protection Device. Got it from marinehowto.com. Not sure they are shipping now or not with Rods illness.

 

[rgano]

I had a Sterling Alternator Protection Device. Got it from marinehowto.com. Not sure they are shipping now or not with Rods illness.

I thought Rod's article to be most thorough. It made me glad that my current boat's needs would never tempt me to go LiFePo.

 

[DDW]

Actually that looks kinda normal to me, depending on a lot of unknown factors. There is no voltage trace shown, it looks like a CV charge dropping to a timed absorb (or a background load), then to a LFP float (essentially off). Where is the voltage sense for the regulators attached?

The only thing that doesn't make sense, is if the battery was really at 50% SOC the CV should have taken longer. Whoever set it up would have to have the maximum alternator output turned down, Battleborn recommends a max charge of 0.5C if I recall, which would be 100A total. Also, the shunt wired as shown would not show charge to the batteries, only house load usage.

 

[seattleboatguy]

You should be getting more alternator output, and it should be more consistent. A few questions:


What are the programmed charging voltages in the MC 618s? Bulk/absorb/float, and absorb time?


Where are the the MC 618 voltage sense wires connected?


BTW, you diagram and chart are excellent and make it SOOO much easier to help with something like this.

I'm attaching the settings from my Balmar cell phone app for the port side regulator settings. The data is the same for the starboard regulator except for stuff like "Device ID" and "Device Name".

My regulators are configured for LiFePo4 batteries.

All my data was collected with the engines running between 1200 and 1300 rpm. I normally cruise at 1500 rpm, but this testing was done at the dock.

The +12v wiring to the 2 regulators is a little complicated, so I made a sketch for you.

Also, at no extra charge, I'll throw in the actual data I collected during the test. I got the data values for the port regulator, the starboard regulator, and the house shunt via the Balmar app. I had a bad temperature sensing cable running to the port alternator, so no data for that column.

Thanks for your help.

 

[gsholz]

According to your advanced programming menus, your bulk charging time is set to 0.1 hours and your absorption time is set to 0.3 hours. This is exactly what you see in your diagram. You'll need to increase these times to fully charge your batteries.

 

[twistedtree]

I think the primary problem is that your sense wire from the alternators is on the alternator side of the combiner, and it needs to be on the LFP battery side. That's why the voltage at the shunt is so much lower than the voltage at the alternators. The sense wire is terminal 9 on the MC 618, not to be confused with the main power supply on terminal 2.


I'd start with that. You want to see the voltage at the LFP battery climb to the bulk/abs voltage, hang there are a bit, then it should switch to float. During bulk/abs you should see the alternators got full output and stay there for the whole charge cycle until it switches to float. At 1200 rpm you might not get full rated output, but it should be close. And as the alternators heat up, output may drop a bit. But the regulators should be driving full field voltage, and with a rates output of 100A, I'd expect 80A or more out of each alternator. The current should be pretty much a square wave, not a curve like with lead batteries.



14.2V may be too low as well, but I would deal with that after you have the basic charge process working correctly.

 

[DDW]

But the regulators should be driving full field voltage, and with a rates output of 100A, I'd expect 80A or more out of each alternator.

Battleborn specs only allow 50A charge for each 100AH battery, so you will need to reduce the max charge to 100A total or risk a BMS disconnect.

This seems to be a little recognized problem with LFP conversions, the initial charge on an AGM bank may well be considerably more than the replacement LFP bank can take. It will still charge to 100% faster, but having to limit the charge created some other problems.

I still don't see where the shunt is wired in the system.

 

[Delfin]

No doubt there will be differences of opinion on this subject, but IMHO 14.3 max bulk voltage is too low. The b1c setting is the minimum time to hold bulk, which will switch to absorption after that time IF the target voltage of 14.3 is reached. It is switching to absorption at 14.1 based on your data, and perhaps TT's observation about the sensing wire explains that.


My bank is five years old now, and my bulk (if 12vdc, not the 24vdc I have) is set at 14.5 volts, which I keep the charge rate at until the amperage being accepted by the battery drops off a cliff. I'll see a slow increase in charge voltage - generally holding at 13.5 volts for quite awhile, then a rapid increase to 14.5, with a concomitant drop off of amperage acceptance over a couple of minutes. I will then hold that 14.5 volts for no more than 10 minutes mostly to defeat memory issues, then charging to the LFP is disconnected, with further charge current then directed to LA start batteries.



Hundreds of cycles, and capacity is around 98%. I could be going about it all wrong, but my impression of LFP is that they should be charged at 14.5 - 14.6 volts until "full" as measured by amperage acceptance drop off, then disconnected from the charging source. Manage them like a fuel tank - use the fuel, refill, then stop pumping current to them.

 

[twistedtree]

Battleborn specs only allow 50A charge for each 100AH battery, so you will need to reduce the max charge to 100A total or risk a BMS disconnect.

This seems to be a little recognized problem with LFP conversions, the initial charge on an AGM bank may well be considerably more than the replacement LFP bank can take. It will still charge to 100% faster, but having to limit the charge created some other problems.

I still don't see where the shunt is wired in the system.

Good point. The spec sheet says charging at .5C (50A per battery in this case), but it also says 100A continuous current. Either way, .5 C is a more comfortable charge rate.

 

[seattleboatguy]

According to your advanced programming menus, your bulk charging time is set to 0.1 hours and your absorption time is set to 0.3 hours. This is exactly what you see in your diagram. You'll need to increase these times to fully charge your batteries.

I like your idea. I will try adjusting those settings.

 

[seattleboatguy]

I think the primary problem is that your sense wire from the alternators is on the alternator side of the combiner, and it needs to be on the LFP battery side...

That sounds like an idea that should be easy enough to test. Thanks very much for the feedback.

 

[DDW]

Good point. The spec sheet says charging at .5C (50A per battery in this case), but it also says 100A continuous current. Either way, .5 C is a more comfortable charge rate.

Deep in their stuff somewhere it says discharge at 1C but charge at 0.5C. I'm running into this problem looking at converting the sailboat to LFP. It takes a pretty huge LFP bank to absorb the charge from my pretty huge alternator. Lifeline AGMs, old tech, but there is no charge limit for current, and when down they will take 1C for awhile no problem. Logically you want about half the LFP capacity as AGM, so now I can charge it at 2C - except most vendors spec against it.

 

[twistedtree]

Deep in their stuff somewhere it says discharge at 1C but charge at 0.5C. I'm running into this problem looking at converting the sailboat to LFP. It takes a pretty huge LFP bank to absorb the charge from my pretty huge alternator. Lifeline AGMs, old tech, but there is no charge limit for current, and when down they will take 1C for awhile no problem. Logically you want about half the LFP capacity as AGM, so now I can charge it at 2C - except most vendors spec against it.

All the banks I've dealt with are large enough to accept full charge capacity, probably because the owners are converting from larger lead banks, or are seeking even more capacity. But point taken. 200Ah @12V just seems like a very small house bank


Back to this installation, even if he limits total charge current to 100A, the system should run at 100A right up to when the batteries are fully charged.

 

[DDW]

Back to this installation, even if he limits total charge current to 100A, the system should run at 100A right up to when the batteries are fully charged.

Agree with that, the problem with the Balmar regulator has always been: how do you make it do that? With house loads that might be 20 - 100A? Again, the beauty of the Wakespeed.....

 

[twistedtree]

Agree with that, the problem with the Balmar regulator has always been: how do you make it do that? With house loads that might be 20 - 100A? Again, the beauty of the Wakespeed.....

With LFP all you need to do is watch the voltage. The end-of-charge rise is unmistakable.

 

[DDW]

With LFP all you need to do is watch the voltage. The end-of-charge rise is unmistakable.

Not really the problem. The OP for example has a 200A charge source, batteries that can accept only 100A, and let's say potential house loads of 50A. If the LFPs would accept a voltage controlled, unlimited amp charge like AGMs, no problem, just let it rip until you see the voltage rise. However they do not. You must limit the charge output to 100A because the worst case is no house load, and LFP in a low SOC which will accept more than 100A without hitting the voltage regulation point and risk disconnect. That means you may get only 50A charge into the batteries with a 50A house load. Complicating it further is the unknown current going into the AGM start bank.

The Wakespeed solves all of these problems as it can see - and can take action on - the actual current being delivered to the LFP battery. The Balmar cannot see that, as it only has field current and voltage. The new 618 and SG300 combination actually does have access to current information, but AFAIK has chosen not to use it.

With an LFP bank that will accept more than any available charge source can produce, then yes, voltage regulation like the Balmar has is sufficient.

 

[twistedtree]

Not really the problem. The OP for example has a 200A charge source, batteries that can accept only 100A, and let's say potential house loads of 50A. If the LFPs would accept a voltage controlled, unlimited amp charge like AGMs, no problem, just let it rip until you see the voltage rise. However they do not. You must limit the charge output to 100A because the worst case is no house load, and LFP in a low SOC which will accept more than 100A without hitting the voltage regulation point and risk disconnect. That means you may get only 50A charge into the batteries with a 50A house load. Complicating it further is the unknown current going into the AGM start bank.

The Wakespeed solves all of these problems as it can see - and can take action on - the actual current being delivered to the LFP battery. The Balmar cannot see that, as it only has field current and voltage. The new 618 and SG300 combination actually does have access to current information, but AFAIK has chosen not to use it.

With an LFP bank that will accept more than any available charge source can produce, then yes, voltage regulation like the Balmar has is sufficient.

I see what you are saying, and agree. I'm still thinking of systems where the batteries can always accept everything the system has to give...

 

[DDW]

Yeah, the bigger the boat the more likely that is. Battery capacity is limited only by displacement and pocketbook, alternators only get so big. But even on my sailboat it is a problem. 440AH 24V AGM battery bank which can and does take all of the 280A from the alternator. Logical replacement is 200 or 300AH LFP. Most specify 0.5C charge. I'd actually have to go up to 600AH in LFP to take all it can give, way more than I need. Or stick with the few that spec 1.0C - even most of those say less is better for long life. Or dial back the Balmar which is probably what I will do (or maybe upgrade to the Wakespeed).

 

[seattleboatguy]

I think the primary problem is that your sense wire from the alternators is on the alternator side of the combiner, and it needs to be on the LFP battery side. That's why the voltage at the shunt is so much lower than the voltage at the alternators. The sense wire is terminal 9 on the MC 618, not to be confused with the main power supply on terminal 2...

I ran a few tests over the weekend, which are summarized in the graphs below.

The first graph is just a copy of the charging profile that I posted at the beginning of this thread. This was the only test I did where I introduced a 25 amp load on the house batteries after 25 minutes.

The second graph had the same testing environment as the first EXCEPT I ran the 2 voltage sense wires from terminal 9 on both regulators directly to the LiFePo4 battery post. I was surprised at the increased amperage it generated.

The third graph had the same testing environment as the second graph EXCEPT I doubled the "Bulk Time" (B1c) setting from .1 hours to .2 hours, and I doubled the "Absorb Time" (A1c) setting from .3 hours to .6 hours. I was surprised that the graph didn't change much. After I saw that third graph, I went back to the boat to see if my B1c and A1c increases had somehow disappeared, but they were still there.

The spreadsheet with all the colors contains details of my charging environment that I captured while doing that third test.

Also, for what it's worth, I called Battle Born, told them I had 2 of their 100 amp LiFePo4 12v batteries connected in parallel, and asked them how much charging amperage they could accept. Their answer was 150 amps, until either the batteries reached full charge or the temperature limits were exceeded.

At this point, I'm pretty happy with the initial amperage I'm seeing, but I think it's a little strange it doesn't last longer. Do you think I should accept these results as "good enough", or continue tweaking the parameters in "advanced settings"?

 

[DDW]

It looks to me like it is dropping to float between minute 13 and 15. Prior to that you were still accepting 90 amps and even in float accepting 20 amps. The reduction in charge is due to the reduction in charge voltage, not because they are charged. Unfortunately this is what the Balmar regulator likes to do. They are famous for premature switch to float.

Ideally what you should see is the charge amps drop, and the voltage increase until it hits the 14.3 limit. It should stay there with the amps dropping further to near nothing, then drop to the float voltage of 13.5.

In the advanced settings there are % field current switches that can be used rather than the fixed times for bulk/absorb, but they don't work all that well either.

I'm still confused as to where your shunt is wired. In one diagram you showed it in series with the house loads. It should be in series with the battery ground.

 

[seattleboatguy]

...I'm still confused as to where your shunt is wired...

One side of the shunt connects to the -12v house battery terminal. The other side of the shunt is connected to all the -12v house wiring.

 

[twistedtree]

This is looking better, but there are still a few things off, I think.


- With LFP, you really want 2 stage charging, not 2 stage. So do that you need to make the absorb cycle either a part of the bulk cycle as it more typically is, or make it part of the float cycle. Right now with a setting of 13.6V it's essentially part of the float cycle. That's an OK approach, but you will need to increased the bulk time to get the 20 min of bulk/absob hold time that Battleborn calls for. Alternately you can increased the absorb voltage to match bulk, and use the combination of bulk and absorb times to get the desired 20 min. Keep in mind that Balmar won't let teh absorb voltage be more than .1 or .2V less than Bulk, or at least that's how it used to be. Because of this, I'm probably do all teh bulk hold time at teh bulk voltage, then set absorb to be the same (or as close as it will allow) as the float voltage.


- You will need to adjust so that once in float, the charger stays in float.


- The details about each regulator's operation is interesting. One continues to show full field even with reduced output current. I suspect this is an artifact of how the centerfielder works.

 

[DDW]

One side of the shunt connects to the -12v house battery terminal. The other side of the shunt is connected to all the -12v house wiring.

That is what your diagram showed - but it also showed the alternator connection as on the battery side of the shunt, not the house wiring side. Where is the alternator connected? It should be on the house wiring side, probably is or your numbers would make no sense.