[thingstodo]

Here is a sketch of how the equipment was connected.

I started with a 6A battery charger on the deep cycle battery. I didn't get very far into the test before the inverter shut down on low voltage. (40% on the variac, under 200W output)

So I replaced the 6A 12V charger with a 40A 12V power supply, cranked the +/- 10% adjustment on the 12V as high as it would go - 13.33V, and connected it to the battery terminals. After the battery was charged up, I got through more trials but was not able to run the 900W heater all the way up. I got a trial at 94% output on the Variac, 714W output on the WattsUpPro to work. 95% on the variac dropped the voltage at the terminals of the DC to AC inverter below 10.8V, where the inverter shut down

As noted, this data includes the losses from the 140VDC to modified sine wave output stage. I have not soldered the additional wiring or fusing inside the DC to AC inverters to allow me to connect loads to the high voltage DC output. The WattsUpPro was available and gave me voltage, current and power measurements - it was convenient. It gives me confirmation that the previous efficiency estimate of 85% was a worst-case scenario. The losses are pretty consistent at about 10% from 366W output to 714W output. I have no reason to expect that relationship to change drastically as the output goes higher. As the transistors heat up, there will be some additional losses.

The output voltage at V2 dropped from 140 VDC to 118.7 VDC at the end of the test. Since the test stopped well under continuous rated power and far under peak power, I will assume that droop was caused by the battery voltage drop into the inverter. I'll try to get the test re-run with an additional parallel battery. The high voltage DC began to droop when the battery voltage dropped below 12.0 VDC.