RV-9 pilot reported a fuel system malfunction in flight required a diversion to a nearby airport. Inspection revealed an unsecured fitting on the gascolator had allowed air enter the fuel pump; leading to fuel starvation.

Narrative

There were two pilots on board the flight; I am the pilot not flying and the pilot flying was an experienced RV pilot with >500 hours in type. During a routine flight at 7;500 ft a sudden and unexpected engine roughness and partial loss of power was encountered on a Lycoming O-320. This prompted quick troubleshooting action to address the issue. Fuel tanks were switched and the mixture was advanced to full rich. This did not immediately fix the issue but after approximately 15 seconds the engine returned to normal operation. Upon resumption of normal engine operation the decision was made to continue on towards our destination airport of ZZZ instead of diverting to ZZZ1 which was approximately 12 miles away. We were not sure what the problem with the engine was or if it would happen again so decided the best thing to do was maintain altitude as long as possible and set up for an approach into ZZZ that maintained gliding range to the runway at all times. While on final for Runway XX at ZZZ the engine trouble started again but this time it was much more severe. Immediately before the engine trouble began again the aircraft was configured for landing and power was reduced. The engine became stuck at idle power. Fuel tanks were switched; mixture was verified full rich; carb heat was turned on; and the boost pump was activated. None of these corrective actions had any effect. The throttle was moved through its full range of motion with no effect; the engine was completely non responsive to any control inputs. The decision was made to raise the flaps and glide to the runway. While crossing the threshold of the runway the engine came partially back to life and was stuck at 1180rpm; this was too much power to safely land so the decision was made to shut off the engine by bringing the mixture to idle-cutoff. Bringing the mixture to idle-cutoff appeared to have no effect on the engine so it was shut down using the magnetos and the aircraft was safely landed with the engine turned off. Once clear of the runway the engine was quickly restarted and the aircraft was taxied to the ramp where it unexpectedly shutdown on its own.Upon investigation of possible causes it was discovered that the sump fitting on the bottom of the gascolator had become loose and fallen out. This system consisted of a sump fitting screwed onto an AN bulkhead fitting with a flared end screwed into the bottom of the gascolator with no provisions to prevent vibration loosening. It was determined that the bottom of the gascolator being opened to the air caused the engine driven pump to begin sucking in mostly air instead of fuel causing fuel starvation and possibly severe carb icing.Upon replacing this gascolator fitting the engine was tested on the ground and in the pattern with no issues. The decision was made to fly the aircraft to a nearby airport where a highly experienced mechanic could inspect it for any other possible causes of the malfunction. No other discrepancies were discovered.Neither pilots on board the flight were the builder of this aircraft. I am writing this report to bring to light what could possibly be a dangerous shortcoming in Vans aircraft. I am not sure if the gascolator sump installation on my aircraft is standard to Vans aircraft or a one off by the builder of my aircraft. I plan to replace this fitting with the proper type and include provisions for safety wire to prevent loosening in the future.