I've only taken apart one after it failed, and although I agree with itsmoked's assessment that there was nothing worth the effort to repair inside, it appeared to me that on mine the little transformer in the boost converter circuit had shorted. For what they claimed the peak rating on the inverter was, I imagine that transformer was woefully inadequate and doomed to failure from the outset if the AC load connected to the inverter had any sort of inductive quality, ie where it required inrush current. That little transformer would have needed to supply all of that power and likely shorted it's windings in the effort, which would I imagine, show up as a short circuit if the damage had not also taken out the entire PCB power supply as well (as it did in my case).
"You measure the size of the accomplishment by the obstacles you had to overcome to reach your goals" -- Booker T. Washington