An EFI VUTEk H5 in Buena Park, California, stopped part-way through its initialization sequence and put a system error on the control panel. It would not get any further than that.
What was in the carriage
The error pointed at the printhead electronics, so the carriage came open. Inside it, lying loose, was a screw. It had come out somewhere above and fallen onto the printhead control boards, bridging them while the machine was powered. Four boards were damaged by the short.
A screw is a two-second find once the cover is off and an impossible one from the panel: the machine reports what stopped responding, not what is lying on top of it. That is the argument for opening the carriage early rather than replacing what the error code names.
Board-level repair
The boards came out for component-level diagnostics. On the first one the damage was a single failed transistor; it was replaced and the board went back into service.
The remaining three are still on the bench as this is written. Each one is being traced the same way — find what the short took out, replace the part, test the board on its own before it goes near the printer again.
Where it stands
One board repaired, three in progress. A set of four printhead control boards bought new is the reason this kind of fault is quoted as a rebuild; at component level it is a transistor and an afternoon per board.
The fastener that started it matters too. Anything loose above the carriage lands on the electronics under it, so a machine that has been apart is worth a look before it is powered up.