NTSB Factual Report

On July 2, 2003, at 2242 central daylight time, an experimental amateur-built, Debay Dragonfly Mark II, N25JD, owned and piloted by a private pilot, sustained substantial damage when it impacted trees and terrain near Flying Cloud Airport (FCM), Eden Prairie, Minnesota. The pilot reported a loss of engine power during initial climb from runway 18 (2,691 feet by 75 feet, asphalt). The personal flight was operating under the provisions of 14 CFR Part 91 without a flight plan. Visual meteorological conditions prevailed at the time of the accident. The pilot was seriously injured. The local flight was originating at the time of the accident.

The airplane incorporated a tandem wing configuration with a tail wheel type landing gear. The airplane was configured to seat two occupants in a side-by-side arrangement. The airframe utilized composite materials over a lightweight foam core. Power for the craft was provided by a Revmaster 2100D engine. This engine is a modified derivative of a Volkswagen automobile engine. The engine incorporated a dual electronic ignition system with two spark plugs per cylinder.

A report submitted by the pilot's mother on his behalf states that the pilot has no memory of the accident event. The written report was submitted about one month after the accident. During a telephone interview at a later date, the pilot stated that he had experienced a loss of engine power shortly after takeoff and was executing a forced landing when the airplane struck trees. He stated that he had aircraft control, but due to the loss of power, he was not able to make it back to the airport and executed the forced landing. The engine was equipped with a Revflow slide-type carburetor and the pilot stated that he had previous problems with the throttle slide sticking. He said that this problem had been resolved prior to the accident flight. The carburetor was not equipped with a carburetor heat box and induction air entered through a cowl opening straight to the carburetor. When asked what he thought may have caused the loss of engine power, the pilot stated that he felt the carburetor had something to do with the power loss. A search of the internet revealed that the pilot had documented, on a website, the refurbishment, and modifications made to the airplane since he purchased the airplane in January of 2003. The website states the following:

5/13/03 - After lots of high speed Taxi Tests - First Flight is made - see the Videos & [Pictures]

First 10 Hours of Flying were interesting to say the least:

* Improved [carburetor] mixture problems with a new needle - had been running too rich at idle

* After doing a dead-stick landing due to having to shut engine down to land after throttle

bound up, I tried the Aerocarb from Aerovee - it has worse binding problems than the

Revflow from Revmaster. I finally resolved the throttle slide binding problems with

help from [person's name] to machine relief [sic] into the backside of the throttle slide, reducing

contact surface area. Probably the most significant change is I removed the old

Bowden throttle linkage cable and replaced with a new heavy duty 10-32 Push-Pull

cable. My recommendation is not to fly with anything less than a 10-32 Push-Pull

cable on the RevFlow, POSA, or Aerocarb (all similar designs).

6/15/03 - After I was more confident in reliability I made a 100 mile Cross-Country flight to visit my

Grandfather on Father's day to Sauk Centre, MN. See photos of first Cross-Country flight

* Mixture is still VERY sensitive. I have to keep EGT at 1300 in cruise by leaning mixture

or engine sputters from being too rich! Gets exciting! :-) Still trying to figure out best way

to set mixture and throttle on landing. Often during rollout on landing the engine seems

quite rich and likes to die out. Bringing throttle up to just above idle seems to help, but I

don't like doing this until speed is below 40 IAS so brakes can be used without using too

much runway. Still searching for the best combination of settings.

6/29/03 - Made Vortex Generators from 3/4" x 3/4" Clear plastic 90 degree angle material used to

cover sheetrock wall corners. It took some time, and I did not have enough made to install every

3" on center as [person's name] directions say, so I installed what I had at 6" centers so I could

add the others later. Simply glued on with 100% Silicone Adhesive. Within one hour it was ready

to fly. I took it around the pattern 3 times, but since it was dark, my approaches and landings were

not as consistent as in the daytime, but I could definitely tell they improved slower airspeed

handling and performance.

6/30/03 - Test flew the VG's more. INCREDIBLE RESULTS! I wonder if having every 3" is

neccessary or if every 6" on center is sufficient. I can't imagine the results getting much better!

I would consider these HIGHLY RECOMMENDED if not mandatory simply for safely of flight and

reduced likelihood of damaging a Dragonfly in a bounce situation!

Before VG's

After VG's

Repairs / Improvements to be made but are not stopping flight

* Fix cosmetic dings in tail and canard bottom skin

* Add better instrument lighting for night flight

* Get a new [Tachometer] or have old one repaired - off by a few hundred RPM

* Rewire something related to Oil Temp - gage shows higher for every electrical

system / device that is turned on.

* Continue to refine mixture / [carburetor] settings

* Weather Striping around Canard to seal water out better

* Add VG's to Canard to help reduce lift loss in rain or with bug buildup

A representative of the engine/carburetor manufacturer was contacted via telephone. During the interview, the manufacturer's representative stated that their records showed that the engine/carburetor was manufactured in July of 1991 and the engine serial number was 2639. The engine was a model R2100D with capacitive discharge ignition. The ignition system is crankshaft triggered and consists of an eight-pole permanent magnet alternator and four ignition coils. The representative also stated that the carburetor was not susceptible to carburetor icing because the slide type carburetor does not have a venturi.

A postaccident examination of the airplane was conducted by representatives from the Federal Aviation Administration. The location and conditions at the accident site prevented an in-depth on-scene examination of the airplane. A subsequent examination of the airplane was conducted after the airplane had been moved to an airplane hangar.

The airframe was extensively damaged during the impact. All of the major airframe components were located in the immediate vicinity of the accident site. The fuel tank still contained fuel. Control system continuity could not be determined due to the extent of the damage; however, no anomalies were found that could be attributed to a pre-impact defect of the control system.

The engine was examined and was able to be rotated. Valve train continuity was confirmed. Each cylinder produced "thumb compression" when the engine was rotated by hand. The engine incorporated an electronic dual ignition system. The ignition system was not tested. The linkages to the carburetor were damaged during the impact; however, no pre-impact defects could be found.

The weather reporting station located at the departure airport recorded the temperature and dewpoint at 2253 as 13 degrees Celsius and 7 degrees Celsius, respectively. A carburetor icing probability chart shows this to be in the range for serious icing at any power setting.

NTSB Probable Cause

The pilot's intentional operation of the airplane with a known deficiency with regard to an improper carburetor calibration that led to the loss of engine power. The improperly calibrated carburetor and the trees were contributing factors.