from my side im happy with the test knowing if i only had my laptop + 2 routers during a day outage and the on screen says its using around 60 to 70 watts i could last around 10 hours or more
Ok so if we take 1280wh and subtract 25% loss. That would leave us at 960wh with 400w load that should give approx 2hrs 21mjn run time
You got close so capacity isn't too bad
No the runtime with lower load cannot be calculated bassd on short run time
Ie the inverter self load stays the same so with less load the runtime will reduce ie the approx 100w device load stays so 400/500w load = 2.25hrs
Low load 100/200w would be 5.6hrs not the expected 9hrs
The fact that it died with very little beeping imo
Means that one cell is dictating the runtime
Normally the beeping is based on pack voltage
Likely the one cell running low caused the bms to do a low cell voltage protection shutdown
Normally it should be easy to run to 24v and hear lots of beeping but device stays on (bar the device has an early shutdown ie 20% etc whichnisn't easy to do on LFP)
As LFP cells are only empty at 2.5v
And almost done by 3v
But yea hitting 24v should be possible
And often ex lead acid devices have an ultimate shutdown of 20v
But yea all of these calculations are based on approx loss % which can be way off or not who knows
Is the device brand new?
If it is i would make the effort to test charge run time
example
note the charge amps and battery voltage lets say 1hrs 40min cahfge time
use amps 29A x 26.5v = 768wh x 1.66hrs (1+ 40/60) =1280wh
If you get less than rated i would return it
As it looks like it either has one cell that is sub par or out of balance battery pack out of the factory
No amount of give the bms time to balance the pack is possible
That is just bullshit they tell people to sooth them and hopefully not insist on swol out
Often devices are sold on special and they would refund and you can't buy the same specs for the same price
In this case each has to decide if the discount is worth the caoacity sacrifice