offgridQLD wrote:Yes to me its crazy.... 8 days I'm sure a 8 day + absence from ones home isn't something out of the ordinary.
I know you have to draw the line at some point as you can never cover for every possible unforeseen chain of events. Though I think this ones isn't unforeseen. It's more or less a guarantee that if you have a LVD event you have a 8 day expiry date on you expensive cells if your not home.
OK. You and Coulomb have convinced me not to connect the DC-DC on the battery side of the battery contactor, despite the fact that measurement of the quiescent current when the new one arrived, showed that it would actually take more like 30 days for the DC-DC to discharge the last 10% of the battery capacity.
So I'll either be using the previously-posted capacitive bootstrap circuit, or Coulomb's suggestion of a green momentary push-button in addition to the latching red emergency stop button. The green button would simply replace the capacitor in the aforementioned circuit.
The component layout looks nice and simple, tidy. How much heat will that dc/dc give off. Is it ok in the closed box?
Good question. With the cover off, its surface temp is 25 degrees above ambient. This appears to be nearly independent of load, which is consistent with it being nearly all switching losses, due to switching at hundreds of kilohertz to minimise its physical size. The losses are only about 4 watts, so I think it will be OK.
An update on progress. I've had hands-on help from both Coulomb and BladeCar which is much appreciated. Thanks guys.
The monolith is currently powering my house, hoorah, with the mains fed in where the genset would go, and a 1.5 kW PV array.
It's all working fine. But so far it's all manually configured and the contactors manually controlled, not yet operating under BMS control. But Coulomb has translated the PIP CRC code to MSP430 assembler, and an optic fibre input has been added to the PIP, and the BMS has successfully sent a command to change the float voltage.
The float-at-76%-SoC idea seems to be working. With the PIP set for both bulk and float voltage of 53.1 V, yellow LEDs dance across the top of the cells as they bypass at 3.322 V per cell.
When they were around 80% to 95% SoC I could have sworn they were near-perfectly balanced, as they were all at 3.33x V. But it's just that this is an extremely flat plateau in the voltage vs. SoC curve. But when I let the SoC drop below that I saw a much greater spread open up between the cell voltages. Their "rested" or float voltages ranged from 3.30x to 3.32x V. And now the voltages are all together again at 3.28x V under light load, which corresponds to the 3.29 V (rested) lower plateau between about 45% and 65% SoC, as shown in this document:
http://www.cse.anl.gov/us-china-worksho ... %20BMS.pdf
So that little step at around 76% really is there, at least for my blue Sky Energy cells. It remains to be seen for the grey CALBs. Who knows, some new dopant might completely spoil this convenient float point.
I use 3.32 V and 76% SoC, not because it's midway between the two plateaux -- that would be about 3.31 V and 73% -- but because it's the point where the slope is steepest, and because it gives us more usable capacity. But it certainly does require the +- 3 mV accuracy that our BMS gives us.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).