Hybrid Battery Bank Idea

[CatNewBee]

I would install 2 separated banks of batteries when using LiFeYPO4, because if the BMS gets screwed or one cell fails, you will lose all power - and this is no fun during a night passage. That can just simply happen, if you got an over-voltage spike (hit by an lightning) or just an electronics failure during the night.

So I would use a hybrid system separated by 4 Diodes / Isolators for the board electronics / navigation / log / lights...

Alternatively I would charge the AGM / GEL Batteries via an Diode Isolator in parallel (keep them just happy in the float phase at 13.6...13.8V) with the LiFeYPO4 bank and have an emergency battery switch to switch over to the smaller Pb bank (150..200Ah) if the lithium bank fails (OVP / UVP / electronics failure) if necessary. It's like an insurance....

[john61ct]

Note redundancy for critical load devices is off-topic, best to stick to the core idea:

"topping off a lead bank to 100% from a separate LFP bank after active charge sources are off"

I believe most people here are talking about the Main bank being lead, with a smaller LFP Reserve, but the core idea works either way.

Diode-based isolators historically suck, but it is true some vendors have released (very expensive) modern ones based on MOSFET that apparently are efficient and suffer little voltage drop.

Examples include Victron ArgoFET, and
​ http://www.perfectswitch.com/isolators/

For that sort of money, personally I would go with a full-fledged B2B charger.

But many have indicated that nothing fancy needs to be between the banks, at least if you're keeping an eye out manually.

A SoC monitor on the lead bank with a relay to close the connection would IMO be a good idea.

[hellosailor]

But once you revert to split or isolated banks...you can't properly monitor voltage and amperage to each bank, unless you've got dual charge sensors and controllers as well. And, some provision to save the alternator if there is a failure and the one bank being charged get cut off.

Once you layer in the complexities and the extra equipment and the added permutations of possible failure modes? It seems cheaper to buy a spare battery and forget the rest. Just replace it all every 8 years instead of 10 because you're not attaining perfect SOC.

This is like the argument that new light fast turbodiesels are superior to old clunkers. Sure they are. And now there's an entire electrical and computer control system that can fail. And a new wire harness. And sensors and solenoids and critical lube oil. All well and good...but it won't go 10,000 hours with just a wrench and screwdriver any more, will it?>

[john61ct]

No one's talking about splitting banks to get this functionality. I would have Main/Reserve anyway, just as others have dedicated Starters or whatever.

Also no issue with Alt, hardwired to one bank.

And **again** the question as to practicality or complexity is off topic, the OP clearly states it is a theoretical thought experiment.

[Dockhead]

Originally Posted by CatNewBee

I would install 2 separated banks of batteries when using LiFeYPO4, because if the BMS gets screwed or one cell fails, you will lose all power - and this is no fun during a night passage. That can just simply happen, if you got an over-voltage spike (hit by an lightning) or just an electronics failure during the night.

So I would use a hybrid system separated by 4 Diodes / Isolators for the board electronics / navigation / log / lights...

Alternatively I would charge the AGM / GEL Batteries via an Diode Isolator in parallel (keep them just happy in the float phase at 13.6...13.8V) with the LiFeYPO4 bank and have an emergency battery switch to switch over to the smaller Pb bank (150..200Ah) if the lithium bank fails (OVP / UVP / electronics failure) if necessary. It's like an insurance....

I never thought of that, and it seems to me to be a really good point.

It's yet another reason to go with a hybrid bank.

But I don't know why you would do it through diodes. Because LiFePo's don't have Peukert and don't care about getting full charges, there's no imperative to make the biggest possible bank out of them, like there is with with lead-acid. You could have two entirely separate banks and use and charge them separately -- as alternatives. Two battery chargers etc., and switch your main alternator between them (I guess).

OR, have a full lead acid bank in parallel with a full LiFePo bank, and also use the the LiFePo bank to finish charge the lead acid bank as I originally suggested.

Actually I really like this. This maybe goes beyond thought experiment and starts to be something I might even think about implementing.

[hellosailor]

"Because LiFePo's don't have Peukert and don't care about getting full charges, there's no imperative to make the biggest possible bank out of them, like there is with with lead-acid"
Lithium of any kind still gets more cycles if you make them shallow cycles. And putting cells into fewer, larger banks (i.e. one big bank) is the way you keep the shallow discharge cycles, regardless of the chemistry.

Flip a coin.

[john61ct]

Actually it is lead's Peukert effect that makes consolidated banks mean shallower draws.

With LFP, as long as you have a good system for switching heavy loads to Fuller banks, there's no difference.

Plus fast CAR means a gennie can top up on demand as needed.

Personally since a well-treated LFP bank will likely last for thousands of cycles anyway, I would not be too concerned about coddling them too much, long as you're staying away from the shoulders on both ends.

[hellosailor]

John, I don't totally disagree with you. Just that all the lithium battery makers, who all still disagree about everything technical, also all manage to agree that a shallower discharge cycle will provide more cycles, in whatever their battery is.

They may disagree about whether is is 2000 @90% and 5000@50%, or 3000@90% and 6000@55%...and they all say SO many cycles even at 99% that nothing else matters...but they all still do say a lower DOD will provide more cycles and more power in the long run.

Do I believe anything any of them have to say? Not really. Without lab tests, without warranties, without track records (Psst! Hey, buddy, wanna buy a battery? We just got these off the truck!)...

I know, they've already paid off in spades for everyone who has bought them. Regardless of how they use or install them. For a decade or longer now. And I heard the crystal harmonics of aging lithium batteries attract unicorns and rainbows, too!

[john61ct]

I'm not saying the relationship doesn't exist.

Just that IMO it's not worth buying more AH to ensure shallower discharge, when even at 90% cycling the payback period's already measured in decades.

Solar plus a quiet little portable genny will ensure I can keep it as shallow as I like anyway, and more cost-effective, that benefit's just a side effect.

And nothing to do with splitting the Main bank, which idea was in itself off topic 8-)

[Dockhead]

Originally Posted by hellosailor

"Because LiFePo's don't have Peukert and don't care about getting full charges, there's no imperative to make the biggest possible bank out of them, like there is with with lead-acid"
Lithium of any kind still gets more cycles if you make them shallow cycles. And putting cells into fewer, larger banks (i.e. one big bank) is the way you keep the shallow discharge cycles, regardless of the chemistry.

Flip a coin.

Fair enough.

But if you split them into two banks, you can still keep the discharges shallow by timely switching between them. I like the idea of having fully redundant banks.

[john61ct]

Why my scheme is still just two banks, but Main / Reserve rather than House / Starter.

Not the same size, but Reserve big enough to run all but non-essential loads.

Cranking and essential loads routinely switched between them when all fully charged, to build in a real-life "everything working" test.

Reserve being lead, but for this thread concept that could go either way.

[hellosailor]

dockhead, try not to get upset but I agree with you again.(G)

OTOH, I also have seen a number of "human element" failures when the humans got too tired, or too sick, or too busy. I try to aim for the kind of engineering that prevents human failures from becoming system failures.

My preference is for one bank. No need to switch battery sensors, etc. No need to throw switches. And, to back that up, a low voltage alarm/cutoff, and a separate "engine start / emergency" battery that is being fed by something like a West (Yandina) Combiner.

Single big bank, but still an ace in the hole.

[john61ct]

Originally Posted by hellosailor

My preference is for one bank.
..
and a separate "engine start / emergency" battery

So now you're up to two anyway.

And none if the rest of what you're talking about is inherently any less complex than how I would implement the core idea here.

[hellosailor]

"So now you're up to two anyway."
Sure, but tat number two can be ridiculously small and cheap and still serve as an engine starter. I've used a $20 17AH AGM "alarm backup" battery to start a diesel, and neither the battery nor the engine had any complaints about it.
Big difference between throwing a <$50 battery in as an emergency SLI, and putting serious batteries in banks.

Oh, and the "jumper cables" weren't walletbusters either.(G)

[marlin driver]

Originally Posted by hellosailor

This is like the argument that new light fast turbodiesels are superior to old clunkers. Sure they are. And now there's an entire electrical and computer control system that can fail. And a new wire harness. And sensors and solenoids and critical lube oil. All well and good...but it won't go 10,000 hours with just a wrench and screwdriver any more, will it?>

Exactly. And why would any intelligent person fool around with this "internet" thing when the library is full of books that work reliably. Who is the smart one when the power goes out?