Optimizing simultaneous alternator charging of LFP and LA

[dsteinfeld]

I currently have an AGM house bank (510Ah) and a flooded lead acid starting battery. I'm planning on upgrading the house bank with 2 paralleled Kilovault HLX+ 3600 batteries (600Ah). While the current system successfully charges with a 125A alternator (with a Balmar MC-614 regulator with alternator temperature monitoring/control) and ACR to parallel the two banks as needed, that won't work well with LFP and LA. I'm retaining the LA starting battery as the recommendation for drop-in LFP is not to use them for engine starting. I'm looking for feedback on this solution:

While I want to protect the alternator from the LFP batteries disconnecting due to a high voltage event (and with the internal BMS of the drop-ins I have no control over this), I don't want to limit the charging current of the LFP bank by charging the LA directly and then a DC-DC converter charging the LFP. That's just a waste of alternator capacity resulting in increased charging time. So I propose to charge both banks via a Victron ArgoFET isolator (which despite being better than diode isolators still has a 0.1v drop). This will leave the LA as a load on the alternator in the event the LFP disconnect. The alternator regulator will be set to Kilovault's voltage/time recommendations, which I know will somewhat undercharge the LA battery, something which I accept. To avoid an LFP high voltage disconnect by its internal BMS, I'm planning on monitoring the LFP battery voltage with a Victron BMV712, and use its high voltage relay feature to open a contactor (wired between the isolator and the LFP bank), diconnecting the LFP bank from the alternator in the event that the LFP voltage rises above a setpoint below that of the Kilovault's BMS HV disconnect voltage. However this introduces the following problem: the alternator regulator sense lead is either connected directly to the isolator's LFP output, leaving it floating there in the event that the contactor opens, or it's connected to the LFP positive terminal, leaving it at the LFP voltage which will begin to decay with the contactor open. This has the potential to overcharge the LA battery as the sense lead no longer relates to the LA voltage. My solution is to transfer the sense leads from the LFP battery to the LA battery via relays driven by an Arduino when the contactor opens. In detail, the Arduino will have control of the regulator (by switching the "ignition on" lead) and will be able to transfer the sense leads (positive and ground) from the LFP bank to the LA battery. When it detects the high voltage alarm (a switch to ground) from the BMV712, it will disable the regulator, transfer the sense leads, then re-enable the regulator. I chose this sequence because I don't know the behavior of the regulator during the switching of the sense leads and didn't want a transient causing a voltage spike. The configuration is fail-safe as the arduino must be running and sensing that the LFP voltage is ok before it sets the sense leads to the LFP bank and closes the contactor for charging.

As I mentioned above, the LA battery will be undercharged regardless of whether the LFP is disconnected or not as the bulk charge setting of the regulator will be 14.1v in line with the LFP manufacturer's charging recommendation, however I think this is acceptable.

Is this overkill? The Arduino implementation is simple as it's an Uno board with an off the shelf 4 relays shield in a box--no custom electronics. Is there a simpler way of charging 2 banks with different requirements and protecting against a disconnect event without using a charge rate limiting DC-DC charger?

[Statistical]

That could work but if you are going to use Victron Battery manager anyways and you trust it then I would just use the relay to cut the field on the alternator. Do you really need to switch it over to charger the starter battery?

On a related note if you are worried about short charging the start battery a simple low powered DC to DC charger solves that.

One other thing to consider is your lithium bank probably can start the motor just fine. Depending on the engine starter current tends to be 50 to 100A but inrush when the motor is not moving can be about 5x as high but that only lasts for a fraction of a second. So battery needs to supply 500A for a fraction of a second then 100A for a few seconds. The issue with small lithium setups is they can't supply the 500A. The BMS sees the current hit 500A and trips to protect the battery. A 600Ah pack though likely is capable of >500A continual and possible more than that briefly.

[dsteinfeld]

Originally Posted by Statistical

That could work but if you are going to use Victron Battery manager anyways and you trust it then I would just use the relay to cut the field on the alternator. Do you really need to switch it over to charger the starter battery?

I would agree if I didn't have an electrically dependent engine (Yanmar 4JH4AE) that has an electric lift pump and stop solenoid. Since this contrivance is for an abnormal situation, I don't want to have to worry about the engine quitting while I'm resolving the electrical issue.

Originally Posted by Statistical

On a related note if you are worried about short charging the start battery a simple low powered DC to DC charger solves that.

I'm not. I'm comfortable with the notion of charging the start battery at 14.1v instead of 14.4. A DC-DC charger would get me 14.4, but I'm still left with the electrical dependency problem (assuming the battery monitor cuts the alternator field as you say, with LFP connected to the alternator output, along with the DC to DC charger and then to the LA start battery).

Originally Posted by Statistical

One other thing to consider is your lithium bank probably can start the motor just fine. Depending on the engine starter current tends to be 50 to 100A but inrush when the motor is not moving can be about 5x as high but that only lasts for a fraction of a second. So battery needs to supply 500A for a fraction of a second then 100A for a few seconds. The issue with small lithium setups is they can't supply the 500A. The BMS sees the current hit 500A and trips to protect the battery. A 600Ah pack though likely is capable of >500A continual and possible more than that briefly.

It's actually 2x 300Ah in parallel, so the internal BMS only has to handle half the current. So you're right--it'll probably work. But I'm not willing to go without a separate start battery.

Thanks for your thoughts -- keep 'em coming!

[CharlieJ]

I suggest that you use the BMV712 relay output to control a Victron Solid Switch 104 (p/n BMS800200104). Install the 104 in the B+ supply to the Balmar MC 614. Easy, peasey.

[dsteinfeld]

Originally Posted by CharlieJ

I suggest that you use the BMV712 relay output to control a Victron Solid Switch 104 (p/n BMS800200104). Install the 104 in the B+ supply to the Balmar MC 614. Easy, peasey.

Which then shuts off the alternator leaving the electrically dependent engine (outlined above) running on battery until the issue is resolved. I'm looking for a solution that doesn't allow a load dump event to destroy the alternator, doesn't limit LFP charge rate (as a DC-DC charger would), and keeps the alternator charging the start battery and running the engine.

[CharlieJ]

It appears that you want a solution for two diametrically opposed conditions: save the alternator from destruction if the house bank internal BMS opens the ATC protection on one of the batteries in the house bank (solved in #4 above) but keep the alternator on line to charge the LA starting battery.

Questions:
1. What is the electrical load of the engine?
2. What is the Ahr rating of the engine start battery?

If the start battery will support the electrical demand of the engine for a reasonable amount of time, use the scheme in #4, install a SPST switch to bypass the Victron Smart Solid Switch 104 and route the output from the alternator to the common pole on a standard 1-2-Both battery switch.

In the unlikely event of a LFP bank failure, the Solid Switch saves the alternator. You switch the 1-2-Both switch from Both to select the start battery and close the Solid Switch bypass SPST switch. MC 614 is back on line, the alternator is on line and the start bank is being charged while you go about troubleshooting what happened to the LFP bank. Still easy, peasy and costs <$100 in material.

[Bellinghamster]

If you're already thinking of putting an Arduino or other custom brain in the system, use it to control both field and Li charge disconnect (set below BMS charge disconnect) in a coordinated manner.
Wire the alternator output to both Li and FLA, but with a contactor in the Li path controlled by the Ardunio (or other voltage controlled swtich). Put the alternator voltage sense on the alternator output (like on the alternator side of the Li disconnect contactor). When approaching full Li state of charge open field winding, then open Li charge path contactor, then reconnect field and you're now running off start battery with the alternator charging it.

I wouldn't mess with trying to "move" the alternator voltage sense from battery to battery - adds risk of it malfunctioning and overcharging something.

[SV Confianza]

Few comments:

First, I don't think I would use an Arduino for a mission critical system. I mean totally up to you, but I wouldn't want to introduce that as a point of failure.

Second, common wisdom of the day is to have the alternator charge the LFP, and the LFP charge the starter bat. Your proposed system seems like it's more complicated than it needs to be.

I went the "not really approved" route and have a relay set to just parallel the starter battery with the LFP when charging. I conferred with a few LFP system designers, including Rod Collins. The theory is that the starter battery only really needs a top up, and when the LFP is charging, the bus is basically at LA float voltage. So when the LA is put on the bus, it just sees float voltage, which is fine to top it up.

This also prevents disconnect dumps from the LFP. I've had 750 hours on the engines since it was installed this way and no issues.

I did a whole writeup on my system and what I learned here since I refer to it often enough: https://www.clevermariner.com/post/c...work-from-boat

[Garrettw]

Victron makes a DC/DC charger that (I think) can take your alternator’s LFP voltage profile and fully charge your starter batt. Presumably it also would provide a load to your alternator.

https://www.victronenergy.com/dc-dc-...t-non-isolated

[donradcliffe]

I heard today that Battleborn is coming out a new battery later this year with a warning light and horn for impending disconnects. Hopefully they will let us tie in to the warning circuit.

[lmxr]

Originally Posted by dsteinfeld

I currently have an AGM house bank (510Ah) and a flooded lead acid starting battery. I'm planning on upgrading the house bank with 2 paralleled Kilovault HLX+ 3600 batteries (600Ah). While the current system successfully charges with a 125A alternator (with a Balmar MC-614 regulator with alternator temperature monitoring/control) and ACR to parallel the two banks as needed, that won't work well with LFP and LA. I'm retaining the LA starting battery as the recommendation for drop-in LFP is not to use them for engine starting. I'm looking for feedback on this solution: [...]

As I mentioned above, the LA battery will be undercharged regardless of whether the LFP is disconnected or not as the bulk charge setting of the regulator will be 14.1v in line with the LFP manufacturer's charging recommendation, however I think this is acceptable.

In case of LA battery undercharge, you can install a small DC-DC from the alternator output to the LA battery. Lithium banks spend most of their time being charged in the bulk phase where their voltage hovers around 13.6 V or less, and 13.6 V tends to be insufficient to charge an LA battery to full. The LA's voltage will be slightly higher because its wires will have less voltage drop, though.

Is this overkill? The Arduino implementation is simple as it's an Uno board with an off the shelf 4 relays shield in a box--no custom electronics.

All together the system feels overly complex. Nothing wrong with some custom electronics, if it's well engineered it will perform just as well (if not better) than off the shelf solutions.

[dsteinfeld]

Sorry for disappearing from this thread -- we're busy becoming grandparents! I'll come back to this when I get back to the project...