Mixed chemistry/Hybrid battery system

[fourlyons]

Hello all, I wanted to ask the brain trust here what the consensus is on some version of a "Drop-in" Fifepo4 and AGM Hybrid battery system. I know some has been written in the past on this, and it is contested, but I think a lot is still being learned about what is the ideal, and what is functional or ideal enough.

It is time for me to replace my 400 ah bank of LA batteries. I have had this capacity for 10 years and don't feel the need for more. I have never had a separate start battery and probably won't change that. I have a dumb alternator regulator, but would consider upgrading. I also have 500 watts of solar, but average about 200-250 watts on a sunny day at any given time.

My reasons for considering a hybrid system; minimal changes to the existing system i.e. lower cost, ballast or base for the system in the event of a bms shut down, fewer complex components to fail (b2b chargers, smart regulators, etc).

Below are some of the links that got me thinking about this. Some are about hybrid banks, some are about using lithium compatible combiners for the alternator charging phase to charge two banks, one of AGM and one Lifepo4. It's also my understanding that many of the outboard driven fishing boats are now coming from the manufacturer with a version of this.

https://www.zwerfcat.nl/en/lithium-hybrid.html
https://nordkyndesign.com/electrical...attery-theory/

My specific question is, has anybody here done this, how is it going so far, and what have you learned? What are the downsides to doing this?

Also if I break my lithium bank up into more than one battery, each with it's own internal bms, do I really need to worry about a complete shut down killing my alternator or causing a black out?

Thanks,

[sailorboy1]

I bet lots of people have a mixed system in 1 form or another. I have a LFP house bank and SLA start battery that are separated by a FET isolator.

A note to your worry, a BMS should NEVER disconnect your batteries, If it does it is a failure in charging system or a mistake programing.

[wholybee]

You won't find consensus, only start an argument. Battery manufacturers and standards organizations (ABYC and ISO) both prohibit mixing batteries of different chemistries. (edit: connecting directly together in parallel) If your boat is insured likely the insurance requires a survey, and so you need to follow those recommendations.

Every time this comes up I have the same answer. There is no reason to combine LFP and Lead in one bank. There just isn't a reason, and it doesn't solve any problems. So while we can argue about if it is "OK" you still shouldn't do it because it doesn't add any real advantages. Some others will chime in and say they have done it, and like it.

In your case, you have 3 considerations. 1) The inexpensive LFP options do not have a BMS capable of starting an engine. 2) LFP batteries charge way faster, and will pull a high current all the way until they are at 100%. This will overheat and burn up a stock alternator. 3) If the BMS disconnects (which should be very rare) the load dump can cause a voltage spike which can damage your alternator and much of your other electronics on board.

If you have a stock alternator and want a simple system and don't want to spend a lot of money upgrading your charring system: Add a lead-acid start battery. Connect the alternator to the start battery, and use a DCDC charger to connect the start to LFP so that the alternator will charge both banks. All other charge sources can go to the LFP. You can use fairly inexpensive drop in LFP batteries if you do this. In this setup, charging will be limited by the DCDC charger, which also serves to protect the alternator.

If you have previously upgraded the alternator with an external regulator that has an alternator temperature probe: In this case you can save a little money and take full advantage of the alternators charge capacity by not using a DCDC charger, but using a FET isolator like the Victron argofet. It would be wise to also add an "Alternator protection device" in the event of a load dump, but as a load dump is a very rare event, it isn't strictly necessary.

If you want to eliminate the start battery: You will need to upgrade the alternator if you haven't already. There are very few options for drop in's that will start an engine, and they are expensive. The most viable option is to build your own bank from individual cells, and use a BMS with a heavy duty contactor. There are lots to choose from, and which to use might spark an argument.

[fourlyons]

Thank for the detailed reply.

So if I upgrade to a smart regulator, I could have a lithium bank, and an AGM bank, and isolate them with the Argofet.

Or, if I upgrade to a smart regulator, I could buy lithium batteries that can handle the inrush/peak current of my starter and drop the AGM.

I own an oscilloscope and will measure my starter peak and duration, but my engine is a 20hp diesel so I doubt it draws a huge amount.

The drop in batteries I am considering, claim the ability to handle a 200 amp draw for 5 seconds. Is it correct to assume that four in parallel could handle 800 amps for 5 seconds?

[wholybee]

Originally Posted by fourlyons

Thank for the detailed reply.

So if I upgrade to a smart regulator, I could have a lithium bank, and an AGM bank, and isolate them with the Argofet.

Or, if I upgrade to a smart regulator, I could buy lithium batteries that can handle the inrush/peak current of my starter and drop the AGM.

I own an oscilloscope and will measure my starter peak and duration, but my engine is a 20hp diesel so I doubt it draws a huge amount.

The drop in batteries I am considering, claim the ability to handle a 200 amp draw for 5 seconds. Is it correct to assume that four in parallel could handle 800 amps for 5 seconds?

It is correct to assume that 4 can handle close to 800 Amps. However the load distribution will not be exactly equal, so it will be a little bit less. You should consider what happens one one battery fails. Will that cause a cascade causing all of them trigger an overcurrent disconnect? Just something to consider in your plan.