Return of the GremlinsI think the coffee spilling inside the lunch box cooler is what released them. They're probably offspring of the ones that ruined my Oshkosh trip a few years ago. That year the gremlins had been especially active — a flat tire, a plug wire coming loose, the Dynon screen shorting out internally and a string of other daily anomalies. It got bad enough that my wife convinced me to ground the plane and I drove to Oshkosh instead. That forced the little pests elsewhere for a while, but this spring's weather must have called them back.I'd broken the Condition Inspection into stages and flown several post-maintenance flights to verify everything and isolate maintenance-induced issues. Minor drag reductions included countersinking the cowl screws deeper so the heads sat flush. I was also keeping an eye on a prior oil cavitation event (momentary low oil pressure) linked to turbulence; we'd inspected the regulator spring and ball, installed a new crush gasket, checked the screen, changed the oil and filter, and verified the sender. Oil sample taken. The bottom cowl cooling ramps had been tweaked, steadily improving the CHT deltas to less than a 20 F spread.Camping gear went in Thursday night. Before dawn Friday, I packed clothes. Weather looked marginal but doable for an early departure. After breakfast and a final sky check, I texted LuAnn (at her brother's in Oklahoma) around 8 AM that I was heading out soon.As the hangar door rolled up, the sky had that "get going now" look. A student was in the pattern as I taxied. Broken clouds were already at pattern altitude, so instead of risking getting pinched between FTW Class D and DFW B, I departed Runway 14, turned crosswind, and climbed through a shallow canyon to get on top west of Bravo and head south then east. Takeoff felt great — little radio chatter while climbing to 11,500 feet. All parameters looked good, though oil temp was rising, so I enriched the mixture slightly and shallowed the climb.
Clouds from the south were solidifying, with another broken layer above 9,500 feet so I opted to level there for now. It was clear to the north, but graphical METARs on the HDX and FlyQ showed airports under the Bravo flipping from VFR to IFR quickly. I could still see terrain below, and things looked more scattered east. It's fine… until it's not.
Then the panel lit up: "Oil Pressure!" — 90 PSI in the yellow, sometimes higher. Nearest VFR options were Ennis or Tyler farther east. I modulated throttle/RPM; the pressure responded but stayed erratic. All other temps and indications remained solid. I'd seen a similar "glitch" before that we'd traced to turbulence-induced cavitation, and recent post-maintenance checks (screen, regulator, filter, oil color still amber after 30+ hours) had looked clean. Still, with a long trip ahead and fresh in my mind how quickly things can go south after recent work, the mental checklist started: How many strikes? Abnormal indication… marginal weather… spilled coffee in the cooler…I found a stable spot below 2,400 RPM where pressure held around 80 PSI. Rusk and Tyler were still VFR. I was already monitoring ATC and nearby airport frequencies, and I reviewed divert options, transponder codes, and possible radio calls. Rather than press on, I turned toward Tyler at 9,500 feet as a precautionary measure. Clouds were lifting in spots, some METARs improving, and ADS-B traffic was increasing. It didn't look terrible below, but I could see why fields were toggling IFR as clouds passed over.At Tyler, I decided the Florida trip was off. I had planned to fly down for the Sport Air Racing League events on Saturday and Sunday, then continue on to Lakeland for Sun 'n Fun. Stopping at Tyler risked being stranded for several days as the weather system moved north. Things seemed to be trending better, so I diverted toward McKinney (where James Redmon is based), descending to 8,500 feet while trying to keep the RPM under 2,400 and CHTs near 300°F. I was still weighing it: sensor glitch? Alternator debris in the accessory case? New filter shedding media? Weather holding? McKinney sits on the edge of DFW Class B. I didn't want to give up altitude so opted to skirt the Bravo on the North side. Is that still two strikes? The airports aren't IFR here…
Denton, Gainesville, I-35, and Hwy 287 were all within easy reach. I descended further to 6,500 feet, planning lower near Denton to avoid arrivals into DFW, but not too close with high density student traffic. Then the phone lit up with texts from my wife: "You're not on FlightAware. Everything ok?" The transponder showed it was receiving and I had ADS-B traffic on my screen. I'd moved the antenna to the copilot side during the CI, so I power-cycled the transponder anyway. I was now literally just minutes from home.THUMP! Like a bird strike, not mechanical. Then another. Nothing else changed. Copeland, Propwash, Alliance passed below. Approaching at 2,000 feet, T67 had no traffic and higher clouds. I announced five-mile final for Runway 14 as a helo crossed in front of me going to Copeland. I noted conflicting windsocks (not unusual due to the hangars, trees and RR tracks), got a bit slow, and touched down firmer than ideal — but safe.Back in the hangar, I texted LuAnn that I was down. Post-flight inspection told the story: A screw holding the temporary fence (lip) on the #3 cooling ramp had come loose, allowing the fence to flap around and cause the thump. This was from my recent tweaks to improve CHT deltas. I also found the cowl screws were not as tight as when installed (countersinking had disturbed the fiberglass surface where the washers seat, so it will need some cycles to restore its hard polish).EFIS data was uploaded to Savvy's site. This was flight #531. The #1 & #3 temp graphs indicated when the ramp extension had started coming loose. Early in the flight, they were within 4 degrees and later went to a 20 F delta.The oil pressure sender seems to be the main culprit behind the erratic oil pressure reading (At least for now). Note that I had already replaced the original pre-2018 Kavlico sensors per Dynon’s Technical Service Bulletin (issued May 2020, updated August 2020) regarding potential issues with the older units. I swapped them for the newer Kavlico v2 sensors (P/N 103757-000 / 105492-002 style), which now have over 350 tach hours on them with no prior problems. The suspect sender was replaced with one of the known-good extra spares from another builder, and oil pressure is now rock-solid at 70 PSI with zero jitter. Finally, my avionics neighbor spotted a pushed-back BNC center pin on the transponder coax about 1/16", which we fixed by installing a new connector and moving the antenna back.I’d been distracted by neighbors during the earlier test run, and another gremlin had quietly arrived. After shutting the engine down and pushing the plane into the hangar, I turned the master and engine bus switches back on — and the relays didn’t pick up. My first thought was, “Did my EarthX battery just die?” The next morning everything was normal. Research suggests the BMS likely isolated due to multiple start/charge cycles until the cells cooled. Resting voltage measured 13.53 V, which is excellent for a three-year-old battery.The old sensor was bench-tested with an air pressure gauge, 5-volt power supply, and voltmeter. At 70 PSI, the green wire voltage fluctuated only a few microvolts (2.302–2.313 V), equivalent to just a few pounds — nowhere near the 100+ PSI swings seen in flight. I’ll still check grounds and connections closely. These piezoelectric sensors are notoriously difficult to diagnose with a simple ohm meter.No test flight was possible before the weather went IFR, so the trip to Florida — and the Sport Air Racing League events plus Sun ’n Fun — is officially scrubbed. Time to sweep the gremlins out of the hangar. There must be a high concentration of them southeast of Dallas.
 |
| Broken layers. In the distance, another above 11K' south of DFW. Photo was taken NE of Denton looking south. |
 |
 |
| McKinney TX |
Denton, Gainesville, I-35, and Hwy 287 were all within easy reach. I descended further to 6,500 feet, planning lower near Denton to avoid arrivals into DFW, but not too close with high density student traffic. Then the phone lit up with texts from my wife: "You're not on FlightAware. Everything ok?" The transponder showed it was receiving and I had ADS-B traffic on my screen. I'd moved the antenna to the copilot side during the CI, so I power-cycled the transponder anyway. I was now literally just minutes from home.THUMP! Like a bird strike, not mechanical. Then another. Nothing else changed. Copeland, Propwash, Alliance passed below. Approaching at 2,000 feet, T67 had no traffic and higher clouds. I announced five-mile final for Runway 14 as a helo crossed in front of me going to Copeland. I noted conflicting windsocks (not unusual due to the hangars, trees and RR tracks), got a bit slow, and touched down firmer than ideal — but safe.Back in the hangar, I texted LuAnn that I was down. Post-flight inspection told the story: A screw holding the temporary fence (lip) on the #3 cooling ramp had come loose, allowing the fence to flap around and cause the thump. This was from my recent tweaks to improve CHT deltas. I also found the cowl screws were not as tight as when installed (countersinking had disturbed the fiberglass surface where the washers seat, so it will need some cycles to restore its hard polish).EFIS data was uploaded to Savvy's site. This was flight #531. The #1 & #3 temp graphs indicated when the ramp extension had started coming loose. Early in the flight, they were within 4 degrees and later went to a 20 F delta.The oil pressure sender seems to be the main culprit behind the erratic oil pressure reading (At least for now). Note that I had already replaced the original pre-2018 Kavlico sensors per Dynon’s Technical Service Bulletin (issued May 2020, updated August 2020) regarding potential issues with the older units. I swapped them for the newer Kavlico v2 sensors (P/N 103757-000 / 105492-002 style), which now have over 350 tach hours on them with no prior problems. The suspect sender was replaced with one of the known-good extra spares from another builder, and oil pressure is now rock-solid at 70 PSI with zero jitter. Finally, my avionics neighbor spotted a pushed-back BNC center pin on the transponder coax about 1/16", which we fixed by installing a new connector and moving the antenna back.I’d been distracted by neighbors during the earlier test run, and another gremlin had quietly arrived. After shutting the engine down and pushing the plane into the hangar, I turned the master and engine bus switches back on — and the relays didn’t pick up. My first thought was, “Did my EarthX battery just die?” The next morning everything was normal. Research suggests the BMS likely isolated due to multiple start/charge cycles until the cells cooled. Resting voltage measured 13.53 V, which is excellent for a three-year-old battery.The old sensor was bench-tested with an air pressure gauge, 5-volt power supply, and voltmeter. At 70 PSI, the green wire voltage fluctuated only a few microvolts (2.302–2.313 V), equivalent to just a few pounds — nowhere near the 100+ PSI swings seen in flight. I’ll still check grounds and connections closely. These piezoelectric sensors are notoriously difficult to diagnose with a simple ohm meter.No test flight was possible before the weather went IFR, so the trip to Florida — and the Sport Air Racing League events plus Sun ’n Fun — is officially scrubbed. Time to sweep the gremlins out of the hangar. There must be a high concentration of them southeast of Dallas.
