On going Lithium, I bought in enough 280Ah cells from china to do two 24V battery's.
And paired them up with the smart JK 8S 200A (2A balancer) BMU.
I did this after doing a quick and dirty total cost of ownership calculation.
Initially Lead Acid looked to be the cheapest buy in, but when you factor in really available capacity, lifetime in cycles etc the LiFePo4's come out on top. By a big enough margin to make them a no brainer. I guess though it depends on whether your circumstances will let you front load the investment or not though.
I have a prototype battery built on the kitchen table and have been using it to trial some kit with.
The kit I was trialling was a cheap grid tie micro inverter from china (180W).
A pair of XL4016 Buck CV, CC regulator boards from china (Supposed to be up to 300W, but no where near really)
A 36V 500W mains SMPSU again from china.
Realy I was using all this manually to try out the idea and it worked sort of. The plan then was to replace the pots on the XL4016 buck converters with digi pot's and bring it under software control, Linking it back to my household automation and the house Enphase Envoy Metered.
Observations
The XL4016 boards are great as a simple charge controler for the LiFePo4 but will run no where near the spec'd 300W as they thermal limit due to their construction and heatsinking. The leads on the XL4016 are a bit skinny for the rated 8A continuous too.
It did all work though till I powered up the charger half one morning and let the magic smoke out of the XL4016. Seems they are really not great with power on surges when operating in the upper part of their spec. Despite me having replaced the heat-sink with an ex zeon processor heat-sink.
The inverter I initially connected straight to the battery and it got quite hot very quickly but seemed to thermal limit. It did not like being fed by a low impedance (low internal resistance) source. it performed better when I put an XL4016 inline but see above notes re spec and wattage rating. The cheap Chinese micro-inverter seemed to settle to drawing at 16V when there was enough current limit applied. I guess this is what the micro-interver's MPPT voltage was supposed to be.
With it insisting on drawing down to 16V though the current limit had to be higher and again we were running into issues with the XL4016 regulator board. I think the micro inverter would be OK with a suitably matched solar array but was a poor match to being battery fed at 24V.
So overall a mixed bunch of results so far.
Having had a deep read around I have plumped for a Victron Multiplus, to do the inverter/charger bit at least for now while I experiment more. The price of energy in the UK has gone a bit mad.
The Victron kit though looks to be fairly easily controlled (it is for me anyway, as a linux and developer type). I have already built out a Raspberry Pi with the Venus OS image and am currently hooking up a DIY serial/USB converter to the battery BMS's serial port.
Next step is to enable the Venus OS's mqtt and integrate it with my houses automations MQTT broker. this should be fairly straight forward, Thereafter I will create a Nodered flow to control the charge/discharge levels using hte import/export values already available in my node red form the Envoy.
I have got root control over the Venus OS Pi so tweaking it's operation a little like this is doable.
Victron seem to have embraced open source enough for me to work my way around getting what I want using their multiplus unit, together with what I already have. They are not the cheapest on the planet, but when you factor in the cost of my time, effort, lost energy getting something going and all else, their prices start to look reasonable.
This is not the ned of experimentation though, just a gain of needed function way point. LOL
I can give my cash to the gouging energy company's in the UK or I can spend it on experimentation. Close but not that close.