Lithium Batteries installation

[Jen-in-Wellies]

1 minute ago, nicknorman said:

Yes however there are two undesirables - one being that modern boats generally have 2 alternators (although this issue can be overcome by rewiring) and the B2B is both very expensive and very inefficient, turning a lot of generated power into heat.

As also discussed in the quoted thread and why it is not as elegant as direct alternator control. 

[nicknorman]

37 minutes ago, David Mack said:

I note that different capacity batteries have the same case size. So do the lower capacity batteries for a particular case size just have more open air inside?

Different models seem to have significantly different lives. The 230Ah costing £700 has a life of 3500 cycles to 80% DOD, whereas the 280Ah version, £800, similar spec apart from the capacity and in the same size case, has a life of 6000 cycles. Why the difference?

The difference is probably how the cycle life was calculated. Guesswork by 2 different people gives 2 different results. With such long cycle lives, it all becomes a bit academic and depends on how it was cycled and under what ambient conditions. And calendar life comes into play as well.

16 minutes ago, Alan de Enfield said:

 

But this hybrid system is the one the insurance companies are not 'happy with'

 

As @BEngo posted earlier :

 

The other cloud  on the horizon is the BSS.  They have asked the IEE to look at the electrical parts of the BSS, with focus on Lithium batteries (and 240 V systems).  It seems that the review body are following in the footsteps of the USA's insurance led standards bodies and are not keen on mixed battery types ( hybrid installations), because a failure of the charge separator will lead to overcharging.  Overcharging Lithium batteries increases the risk of a battery fire.

 

I seem to remember that @matty40s has previously posted similar information

No a "hybrid system" is where 2 batteries of different chemistries are directly connected together. It is unreasonable to consider a Li battery charged by a charging system that just happens to be assisted by another battery, to be a "hybrid system".

13 minutes ago, Alan de Enfield said:

 

Just a thought - my car has a 170 amp water cooled alternator, apart from the price, would that be a suitable option for the high current, extended run times required by Lithiums ?

A vehicle alternator, if it is from a car built in the last 10 years or so, is unlikely to have a built in regulator. The regulation of the alternator is controlled by the vehicle ECU (engine control unit) as it adjusts the voltage according to what the car is doing (eg lower voltage during acceleration, high voltage during overrun or braking to recoup some of the energy) and according to the SoC and load.

So you would need an external regulator. Other than that, it would certainly allow for higher % of alternator load. There was a thread on here where someone had done that.

12 minutes ago, Alan de Enfield said:

 

If you tailor the alternator output to suit the alternator (ie keep it at a lower level so it doesn't overheat) does that not remove one of the major benefits of Lithiums over FLA batteries ?

 You would now need to run the engine for (possibly) just as many hours as you would for charging FLA's

No. For example our 175A alternator can run at about 125A continuously without getting too hot. When we had LA, the current would be below 125A after the first 5 mins or so, and then spend many hours gradually reducing with the final hours at below 20A in order to fully recharge and ward off sulphation.

With Li, we can put in 125Ah an hour every hour until 15 minutes before the battery is full.

14 minutes ago, Jen-in-Wellies said:

The failure of any Lithium charger could lead to overcharging, since they are all separators between the input source and the Li bank. This would include mains, solar MPPT's, or even alternator controllers. I'm not seeing how a properly engineered B2B from the likes of Victron would be any different. 

This is why you need 2 stages of protection. The first stage is that the charge source is designed not to overcharge the battery. The second and independant stage is that if the first stage fails, the BMS disconnects the battery. What is to be avoided IMO is to use the second stage routinely as a means of controlling the charging.

[dmr]

8 minutes ago, nicknorman said:

 

Yes however there are two undesirables - one being that modern boats generally have 2 alternators (although this issue can be overcome by rewiring) and the B2B is both very expensive and very inefficient, turning a lot of generated power into heat.

 

Victron do nice DC-DC converters with battery charging capability and about 90% efficient, but by the time you get up to 100amps or whatever the price is probably similar to something like the Arco Zeus.

Are two alternators a problem? just wire them both to the starter battery. As long as the voltage stays low enough to stop their internal regulators competing it should be fine. The advantage of something like the Arco Zeus is doing proper temperature control on the alternator.

 

As you can deduce I am sort of sold on the Zeus but want to wait till there is is some real world user feedback. Has anybody here got one? 😀

20 minutes ago, Alan de Enfield said:

 

If you tailor the alternator output to suit the alternator (ie keep it at a lower level so it doesn't overheat) does that not remove one of the major benefits of Lithiums over FLA batteries ?

 You would now need to run the engine for (possibly) just as many hours as you would for charging FLA's

Not really, if a rated 100 amp alternator can produce a continuous 70amps then that is still much better than the 15amps that goes into LA's towards the end of the charge cycle. It is really just ensuring the alternator works within its real world rating rather than de-rating it.

[Jen-in-Wellies]

21 minutes ago, nicknorman said:

This is why you need 2 stages of protection. The first stage is that the charge source is designed not to overcharge the battery. The second and independant stage is that if the first stage fails, the BMS disconnects the battery. What is to be avoided IMO is to use the second stage routinely as a means of controlling the charging.

Single point of failure. Always a bad idea, which is the problem with the long wire, plus BMS disconnect only method. 

[nicknorman]

8 minutes ago, dmr said:

 

Victron do nice DC-DC converters with battery charging capability and about 90% efficient, but by the time you get up to 100amps or whatever the price is probably similar to something like the Arco Zeus.

Are two alternators a problem? just wire them both to the starter battery. As long as the voltage stays low enough to stop their internal regulators competing it should be fine. The advantage of something like the Arco Zeus is doing proper temperature control on the alternator.

 

 

No, 2 alternators are not a fundamental problem, but some rewiring would be required.

Just get the Zeus, you know you want to!

[BEngo]

19 minutes ago, magnetman said:

 

I wonder if permanent magnet alternators may be interesting here. 

Generating the  watts while keeping the rotor and stator cool is not  the problem.  Keeping the diodes cool is. 

Heating the diodes comes from the rotor/stator losses but  mainly from  the forward bias  voltage across the diode x the current through it.  Not sure how a permanent magnet alternator will reduce anything other than the  rotor losses.  A remote, fan cooled diode pack would be a good idea.

 

N

[Jen-in-Wellies]

1 minute ago, BEngo said:

Generating the  watts while keeping the rotor and stator cool is not  the problem.  Keeping the diodes cool is. 

Heating the diodes comes from the rotor/stator losses but  mainly from  the forward bias  voltage across the diode x the current through it.  Not sure how a permanent magnet alternator will reduce anything other than the  rotor losses.  A remote, fan cooled diode pack would be a good idea.

 

N

Or bolt the remote diode pack to the boat base plate! 

[nicknorman]

Just now, Jen-in-Wellies said:

Single point of failure. Always a bad idea, which is the problem with the long wire, plus BMS disconnect only method. 

 

This reminds me of helicopters. Lots of warning lights for dire things like tail rotor gearbox chip. But then you have to have a means to be confident that the warning system would work in the once in a lifetime actual problem, because if it didn't lots of people are going to die. So you make it failsafe ie if the system fails, the light comes on. And then because you are not totally confident in that scheme, you add a test button to simulate the gearbox chip and bring the warning light on which tests the bulb and all the wiring, and hope that decreases the failure rate to "extremely remote". But then with so much complexity, you are plagued with false warning lights. And you get so fed up with that, you start to ignore them...

[IanD]

As far as boats are concerned, insurance companies are not in favour of directly paralleled LA/LFP systems because their safety depends completely on how they're built and controlled -- yes they can work OK if built and used by somebody who understands what they're doing, but usually they're a sign of a DIY bodge rather than doing an LFP system properly.

 

The much-discussed (and POSTED IN VLOGS!!!) lithium fire hazard is with NMC and similar chemistries not LFP, and the insurance industry seems aware of this going by the recent symposium.

 

Properly-installed LFP batteries are probably safer than LA which everyone has used for many years since there's no release of explosive gas when overcharged. However a bodged LFP system without proper charging control and protection can lead to battery destruction but also burned-out alternators, and the fact that the batteries have huge short-circuit currents (if not protected) can cause wiring fires. This shouldn't happen with "drop-in" batteries with an inbuilt BMS with protection, assuming that this works and is reliable, but this can't be used to give carte blanche to any DIY LFP install.

 

None of which is an issue with a properly-designed/installed LFP system, but there's the rub, there are new things to go wrong compared to well-understood LA batteries -- and the reason that some insurers (e.g. in the USA) are insisting on LFP systems being either professionally installed or vetted by a qualified electrician before giving insurance cover, or may even refuse to cover DIY systems at all.

[dmr]

2 minutes ago, nicknorman said:

 

No, 2 alternators are not a fundamental problem, but some rewiring would be required.

Just get the Zeus, you know you want to!

Yeah, it does CANbus, two shunts for proper alternator current limiting (which I need) and temperature control, it looks a bit like a bigger commercial version of what you did. Its very new and currently no UK agents so a hefty shipping charge, but I am heading in that direction.

and like you, it can use the state of charge from the Victron BMV to terminate charging.

...I am surprised that no "drop in Lithiums" give a simple 2 wire "enable charge" output.

[IanD]

14 minutes ago, BEngo said:

Generating the  watts while keeping the rotor and stator cool is not  the problem.  Keeping the diodes cool is. 

Heating the diodes comes from the rotor/stator losses but  mainly from  the forward bias  voltage across the diode x the current through it.  Not sure how a permanent magnet alternator will reduce anything other than the  rotor losses.  A remote, fan cooled diode pack would be a good idea.

 

N

 

PMAC alternators always use a power MOSFET switch bridge to control them and convert AC to DC, and these have much lower losses than diode bridges because the voltage drop is many times lower. They also don't have rotor windings which need keeping cool, and the magnetic design also makes it easier to keep the stator windings cool. Also maximum rpm can be higher because there are no windings or slip rings, the rotor is just steel and magnets, so they can use a bigger pulley ratio and run faster which keeps them cooler and gets more power out of a given case size. Overall efficiency is usually around 90% compared to typically 50% for a normal 12V alternator, so they have to dissipate far less heat (maybe 5x less) and also load the engine less for the same output current.

 

So great for higher power levels and heavy-duty use and cooling and efficiency and fuel consumption, but also considerably more expensive to buy -- which is why few boaters use them... 😉

Edited January 25 by IanD

[nicknorman]

3 minutes ago, dmr said:

Yeah, it does CANbus, two shunts for proper alternator current limiting (which I need) and temperature control, it looks a bit like a bigger commercial version of what you did. Its very new and currently no UK agents so a hefty shipping charge, but I am heading in that direction.

and like you, it can use the state of charge from the Victron BMV to terminate charging.

...I am surprised that no "drop in Lithiums" give a simple 2 wire "enable charge" output.

Interesting that it has a shunt for the alternator current. Not sure why that would be necessary as one can calculate the current from field current and rpm to the necessary accuracy.

In any case, usually it is the torque load at low rpm that is the issue and this again is related to the field current, not to the output current. Same field current, same torque load, gives lower output current at low rpm and higher output at higher rpm. Or to put it another way, for the same current output, the torque decreases as rpm increases. So to adjust the torque load to match the engine output at low rpm, you just need to set up a simple max field current vs rpm function.

[dmr]

Interesting observation about field current. I need to think about that but you will be correct. I think the Zeus can do field control too but I will investigate.

I am pretty much stuck with a 100 amp alternator on a single V belt so torque and temperature limits are needed.

The alternator shunt is a bit odd as it usually has to sit in the postive wire which makes the electronicss a tiny bit more difficult and means that the sense wires might need to be fused.

[nicknorman]

1 hour ago, dmr said:

Interesting observation about field current. I need to think about that but you will be correct. I think the Zeus can do field control too but I will investigate.

I am pretty much stuck with a 100 amp alternator on a single V belt so torque and temperature limits are needed.

The alternator shunt is a bit odd as it usually has to sit in the postive wire which makes the electronicss a tiny bit more difficult and means that the sense wires might need to be fused.

Controlling the field current is of course what any regulator does! It is just a matter of whether a max field current limit can be set. But this is what the various existing devices do (Alpha Pro, Wakespeed). With modern precise PID digital regulation, trying to control current by adjusting the set voltage is very difficult because very small changes in voltage make a huge difference in output current especially into Li batteries, and it varies with varying load. And if you think about it, an alternator is a current source not a voltage source, a regulator is a current supply not a voltage supply - the main regulation loop adjusts the field current to adjust the output current such that the output voltage is at the desired value.

 

So no point in complicating it by trying to control the output current via the output voltage which is controlled via the output current which is controlled via the field current. Just set a max field current (max duty cycle on the controlling MOSFET) and that is job done.

[dmr]

I suppose with a shunt you can set an accurate current limit (in amps) straight off, but with field current limit you would have to do a small experiment to find the relationship between field current and output, or can you assume that 100% output correponds closely to 100 field?

If you have real time readout of field and output then this is a quick and easy experiment, but in reality it is likely to be temperature or belt slip that needs to be controlled so a bit of experimentation is required anyway.

In the recent re-work of my high current electrics I have left space for a second shunt but have not fitted one yet 😀

 

[IanD]

22 minutes ago, dmr said:

I suppose with a shunt you can set an accurate current limit (in amps) straight off, but with field current limit you would have to do a small experiment to find the relationship between field current and output, or can you assume that 100% output correponds closely to 100 field?

If you have real time readout of field and output then this is a quick and easy experiment, but in reality it is likely to be temperature or belt slip that needs to be controlled so a bit of experimentation is required anyway.

In the recent re-work of my high current electrics I have left space for a second shunt but have not fitted one yet 😀

 

Since the relationship between field and output current varies with battery voltage and rpm and temperature -- both ambient and as the alternator heats up -- it's actually a lot simpler to just measure the output current using a shunt and adjust the field current to set the required value (current demand), without doing any characterisation and curve fitting or relying on the belt not slipping.

 

If the controller also knows what the rpm is it then becomes trivial to do things like limiting the charging current (engine/belt load) at low rpm/idle, and if it knows what the alternator temperature is it can reduce the current demand if it gets too hot to stop it burning out when charging LFP batteries.

 

All this is what Wakespeed and the like do, because it's the right way to solve the problem 🙂

Edited January 25 by IanD

[dmr]

2 minutes ago, IanD said:

Since the relationship between field and output current varies with battery voltage and rpm and temperature -- both ambient and as the alternator heats up -- it's actually a lot simpler to just measure the output current using a shunt and adjust the field current to set the required value (current demand), without doing any characterisation and curve fitting.

 

If the controller also knows what the rpm is it then becomes trivial to do things like limiting the charging current (engine/belt load) at low rpm/idle, and if it knows what the alternator temperature is it can reduce the current demand if it gets too hot to stop it burning out when charging LFP batteries.

Yeah, the Zeus also measures alternator speed. It does appear to do just about everything that is required.

Its sad that Adverc never evolved their design to do stuff like this. They had a good market lead but just sat back and did no further developement, probably because they did not want to get involved with processors and software.

[Tony1]

5 hours ago, Alan de Enfield said:

 

If you tailor the alternator output to suit the alternator (ie keep it at a lower level so it doesn't overheat) does that not remove one of the major benefits of Lithiums over FLA batteries ?

 You would now need to run the engine for (possibly) just as many hours as you would for charging FLA's

 

There is some truth in that, but I think the reality might not be quite so bad. 

For example, if your alternator can only charge at say 40 amps, when it is charging lithiums they will absorb a consistent 40 amps for almost all of the charge period, until almost full. 

However (as I understand it) lead acids will start to 'resist' the incoming charge within maybe 20-30 mins, and it seems that resistance increases as they get increasingly full.

So it could be, for example, that in the second hour of charging they are only being charged at 20 amps, instead of the potential 40 amps available.

That is a guess, of course- an expert like Tony B will have a precise idea of what the 'fall off' rate is for the charging current as the battery gets full.    

The final 10% to get the lead acids up to 100% SoC might therefore take a good few hours. Maybe that's why people give them an 8 hour charge every week or so.  

So there is that advantage of lithiums- even if you can only charge them slowly, they will still  absorb that limited charge faster than lead acids will.

 

 

[blackrose]

Maybe I'll install lithium batteries when everyone's agreed on how to do it easily and safely while keeping my LA bank so I don't have to change all my charging equipment. To be honest I wouldn't really need my 70 amp alternator charging the lithium bank and could just use solar for that.

 

I've said this before and was shot down, but lithium battery technology on boats is obviously still in its infancy hence the multitude of different systems that people have come up with. Conventional battery technology on boats is so well established that everyone who's read a bit knows exactly what to do and all the experts are agreed. That's clearly not the case with lithium.

[Jen-in-Wellies]

15 minutes ago, Tony1 said:

 

There is some truth in that, but I think the reality might not be quite so bad. 

For example, if your alternator can only charge at say 40 amps, when it is charging lithiums they will absorb a consistent 40 amps for almost all of the charge period, until almost full. 

However (as I understand it) lead acids will start to 'resist' the incoming charge within maybe 20-30 mins, and it seems that resistance increases as they get increasingly full.

So it could be, for example, that in the second hour of charging they are only being charged at 20 amps, instead of the potential 40 amps available.

That is a guess, of course- an expert like Tony B will have a precise idea of what the 'fall off' rate is for the charging current as the battery gets full.    

The final 10% to get the lead acids up to 100% SoC might therefore take a good few hours. Maybe that's why people give them an 8 hour charge every week or so.  

So there is that advantage of lithiums- even if you can only charge them slowly, they will still  absorb that limited charge faster than lead acids will.

 

 

Even if you could only charge the Lithiums at 30A, you could put in 100Ahr in only 3hrs 20mins and the batteries would be safe to leave for weeks on end without further engine charging, assuming you were not going to drain them to their allowable minimum in that time.

[Roberto Conigliaro]

10 minutes ago, blackrose said:

Maybe I'll install lithium batteries when everyone's agreed on how to do it easily and safely while keeping my LA bank so I don't have to change all my charging equipment. To be honest I wouldn't really need my 70 amp alternator charging the lithium bank and could just use solar for that.

 

I've said this before and was shot down, but lithium battery technology on boats is obviously still in its infancy hence the multitude of different systems that people have come up with. Conventional battery technology on boats is so well established that everyone who's read a bit knows exactly what to do and all the experts are agreed. That's clearly not the case with lithium.

yeah I see your point, from what I am reading here (thanks everybody by the way!), there are a lot of different theories and disagreement, so it might seem too much of a jump in the dark a this stage. Said that, energy efficiency is such in important part of life on the boat that I am gonna do more researches and hear more from people who did the switch, before deciding it isn't the right time for this investment, taking in consideration the pros and cons, financially and on a practical level 

[nicknorman]

46 minutes ago, IanD said:

Since the relationship between field and output current varies with battery voltage and rpm and temperature -- both ambient and as the alternator heats up -- it's actually a lot simpler to just measure the output current using a shunt and adjust the field current to set the required value (current demand), without doing any characterisation and curve fitting or relying on the belt not slipping.

 

If the controller also knows what the rpm is it then becomes trivial to do things like limiting the charging current (engine/belt load) at low rpm/idle, and if it knows what the alternator temperature is it can reduce the current demand if it gets too hot to stop it burning out when charging LFP batteries.

 

All this is what Wakespeed and the like do, because it's the right way to solve the problem 🙂

I disagree. Well I suppose the question is, why do you want to limit the current? Is limiting the current actually the aim (for example because the battery can’t safely take the full current)? If so then fair enough.
 

But in the context of Li batteries and narrowboats, this is almost certainly not the reason. The reason will be due to:

1/ avoiding overheating the alternator and/or

2/ avoiding overloading the engine, belt and or pulleys at low rpm.

 

In neither case does the actual current matter. What matters is how hot the alternator is, and how much torque is being applied to the alternator. Both these are best controlled by directly adjusting the maximum field current, not some convoluted mechanism to measure and then adjust the output current.

 

For 1: there is no direct relationship between temperature and current, it depends on ambient temperature and fan rpm amongst other things. Having a set limit to the field current allows less current at low rpm when the fan is going slowly, and more at higher rpm when the fan is going faster for the same temperature (roughly). Which is exactly what you want.

 

For 2, there is a direct relationship between torque and field current at a specified low rpm. Set a maximum field current, this sets a maximum limit on the torque to the alternator, which is exactly what you want.

 

We are not miles apart, but is clear that you have not designed alternator control software, and I have. Once you actually start writing the code, it all becomes clear and there is no need to complicate things by considering output current, when the more relevant parameter, and one that is already intrinsically measured by the control chip is field current. I was going to include a Hall effect current sensor in my design, before I realised it would be completely pointless.

 

30 minutes ago, blackrose said:

Maybe I'll install lithium batteries when everyone's agreed on how to do it easily and safely while keeping my LA bank so I don't have to change all my charging equipment. To be honest I wouldn't really need my 70 amp alternator charging the lithium bank and could just use solar for that.

 

I've said this before and was shot down, but lithium battery technology on boats is obviously still in its infancy hence the multitude of different systems that people have come up with. Conventional battery technology on boats is so well established that everyone who's read a bit knows exactly what to do and all the experts are agreed. That's clearly not the case with lithium.

The only thing wrong with your post is that you have this fantasy that at some point, everyone will agree on how to do it. Never going to happen!!!

As you say, it is in its infancy. However the only sceptics are those who have not tried it. Those who have tried it would never contemplate going back to LA.

Edited January 25 by nicknorman

[IanD]

19 minutes ago, Roberto Conigliaro said:

yeah I see your point, from what I am reading here (thanks everybody by the way!), there are a lot of different theories and disagreement, so it might seem too much of a jump in the dark a this stage. Said that, energy efficiency is such in important part of life on the boat that I am gonna do more researches and hear more from people who did the switch, before deciding it isn't the right time for this investment, taking in consideration the pros and cons, financially and on a practical level 

 

I don't think there's any different theories or disagreement about the right way to use LFP batteries on a boat with various charging sources like an alternator and solar and shoreline and generator -- a proper charge controller/BMS/inverter/charger that keeps the batteries and alternator and everything else in the best of health, and stops them being over or under charged. For a new installation it's a no-brainer nowadays, but as an upgrade it means a rip-up-and-start-again.

 

The problem is that this is more complex and expensive than a good old dirt-cheap alternator/lead-acid setup, and lots of people can't or don't want to cough up and do LFP properly so they try various ways of saving money and bodging it -- some safer than others, some distinctly iffy, some combining LFP and LA, and some that shouldn't be touched with a bargepole -- even a long one.

 

The arguments are about how safe/dangerous/satisfactory/dodgy these cheapo setups are, with proponents claiming they're absolutely fine and others -- including insurers -- disagreeing.

Edited January 25 by IanD

[Roberto Conigliaro]

31 minutes ago, IanD said:

 

I don't think there's any different theories or disagreement about the right way to use LFP batteries on a boat with various charging sources like an alternator and solar and shoreline and generator -- a proper charge controller/BMS/inverter/charger that keeps the batteries and alternator and everything else in the best of health, and stops them being over or under charged. For a new installation it's a no-brainer nowadays, but as an upgrade it means a rip-up-and-start-again.

 

The problem is that this is more complex and expensive than a good old dirt-cheap alternator/lead-acid setup, and lots of people can't or don't want to cough up and do LFP properly so they try various ways of saving money and bodging it -- some safer than others, some distinctly iffy, some combining LFP and LA, and some that shouldn't be touched with a bargepole -- even a long one.

 

The arguments are about how safe/dangerous/satisfactory/dodgy these cheapo setups are, with proponents claiming they're absolutely fine and others -- including insurers -- disagreeing.

 

Well I just finished chatting with a person in another group who was saying that hybrid solution is the best  : super safe and straight forward - just install a AGM batteries between the alternator and the Lifepo4 as a buffer and the trick is done. I feel like maybe some people would disagree with that, at least in terms of safety? 

[nicknorman]

1 hour ago, Roberto Conigliaro said:

yeah I see your point, from what I am reading here (thanks everybody by the way!), there are a lot of different theories and disagreement, so it might seem too much of a jump in the dark a this stage. Said that, energy efficiency is such in important part of life on the boat that I am gonna do more researches and hear more from people who did the switch, before deciding it isn't the right time for this investment, taking in consideration the pros and cons, financially and on a practical level 

Energy efficiency is a good point. This is the product of the voltage and charge put in, vs that coming out. Both LA and Li have similar charge efficiencies (ie Ah out is about 95% of Ah put in). However when you take into account voltage, the energy efficiency of Li is much better because LA charges at 14.4v and discharges at (say) 12.5v. whereas Li mostly charges at 13.6v or less, and discharges at (say) 13.2v. And this is before you take into account the need to run an engine or generator for hours with just a trickle going in to get LA batteries up to 100%.

10 minutes ago, Roberto Conigliaro said:

 

Well I just finished chatting with a person in another group who was saying that hybrid solution is the best  : super safe and straight forward - just install a AGM batteries between the alternator and the Lifepo4 as a buffer and the trick is done. I feel like maybe some people would disagree with that, at least in terms of safety? 

It works adequately and lots of people do it. It is reasonably safe.

However it is a bodge and may have a limited future because most Li battery manufacturers specify "do not mix with batteries of a different chemistry". And the BSS people don't like it and at some point it might (or might not) become a BSS fail. If you are non-compliant with BSS and have deliberately flouted manufacturers installation instructions, the insurance co are not going to want to pay out!

I explained earlier that although the batteries are fairly compatible and complement each other, the fine detail for long term is more problematic due to the differing charging requirements. So it remains a bodge that works after a fashion, but is not "the right way to do it"!

Edited January 25 by nicknorman