- Topic Moderator
- Posts: 13
- Joined: Fri Dec 31, 2004 11:55 pm
- Location: Ottawa, Canada
The follow text is extracted from the incident reports he wrote after the flights (edited to remove information about the aircraft model, as the design will change to fix these issues, and the company has not given permission to release the details):
Incident #1: Brake Line Binding
Certification flight testing of a light plane prototype was ongoing. The flight was devoted to exploratory spin testing, conducted at a mid weight and CG. The aircraft was fitted with a jettisonable door and the sole occupant was wearing a parachute. Test point #6 was a power ON spin entry, conducted at 7000 feet with pro-spin inputs held for one second. The brief entry and recovery were uneventful, however immediately after recovery to level flight at 75 KIAS it was observed that the slip ball was displaced fully to the left. Centering the ball required approximately 50-70 lb left rudder force. A controllability check was performed. Elevator and aileron response were normal. The rudder was compliant, but exhibited a strong static right force bias. In flight, possible causes were considered to be either (1) a stuck nosewheel, resulting in a rudder force bias from deflection of the nosewheel steering mechanism, or (2) structural deformation or incipient failure of the vertical stabilizer. In order to minimize structural loads, it was decided to limit speed to 70 KIAS, maintain the current configuration, and to perform a flapless landing. In consideration of a possible uncommanded nosewheel deflection, it was decided to perform a shallow, power-on approach in order to hold the nosewheel off the runway for as long as possible. A straight-in approach was set up to landing ATC was advised. The flight was passed to the local FSS, and an emergency was declared. Winds were 240 degrees at 10 to 15 knots. Upon landing, with the nosewheel still in the air, the aircraft sharply veered to the right. Neither full rudder nor left differential braking was unsuccessful at arresting the rate of turn. The aircraft departed the runway and came to rest in the grass at a 90 degree angle to the runway. The aircraft was shut down without further incident and FSS was advised. The aircraft was undamaged.
Post-flight investigation revealed that rapid, full deflection rudder inputs could result in interference between one of the rudder pedals and a brake line. This condition would account for the right rudder pedal force bias and the right heading divergence upon landing. Attempts to apply compensatory left pedal upon landing had the effect of applying uncommanded right brake. Slight chafing was found on the affected brake line, confirming the nature of the problem. The manufacturer is modifying the prototype to address the deficiency.
Lesson - Pay very close attention to anything that could possibly interfere with the flight controls. Consider that things may move out of their normal position during the unusual motions in a spin.
Incident #2: Spin Testing
A prototype light plane was engaged in developmental spin tests, progressing slowly from extensive exploratory spin testing. The loading configuration was heavy-forward. At 7200 feet the aircraft was set up for a one-turn, clean, power ON, spin to the left,. This was the tenth repeat of this test point, in an effort to characterize the precise recovery interval following spin recovery inputs. The nine previous events were nominal with recovery occurring within approximately one turn. Certification criteria required recovery from a one-turn spin with "no more than one additional turn". In accordance with Advisory Circular AC-23-8B, it was decided that the definition of a "one turn recovery" could begin following the completion of recovery inputs, and needed a more precise record of the recovery response. So...entered a one turn left spin, then idle power, full opposite rudder, progressive forward elevator.....hmmm....to the forward control stop. Nothing. The spin accelerated to a rate of approximately 1.6 seconds per rev. The engine and propeller stopped. It was noted that the briefed bail-out altitude was approaching. The pilot changed hands on the stick to jettison the door, and in so doing momentarily released full forward elevator pressure. In response there was a slightly pitch oscillation. The technique was recalled where pilots have recovered from spins by applying an in-phase pitch rocking technique to break the stall. This technique was applied by modulating the elevators within approximately 25% control deflection around the full forward stop. Recovery occurred after another 2-3 turns. Engine was restarted. The aircraft returned to the airport without further incident. A time history of the event is depicted in Figure 1, below. Note the briefed bail-out altitude was 4000 feet.
The above time history shows about 10 turns before recovery, and an altitude loss of about 2500 ft.
These incidents highlight the risks inherent in spin flight testing. Spins are not all exactly the same. Small differences in the entry conditions, control positions, weight and CG can affect how a spin progresses. In the second incident, Rob did a spin at a particular condition that had been completed nine times before, with the recovery occurring about one turn after the recovery was initiated. But, on the tenth test, recovery did not occur then the recovery controls were input. There is a big difference between a one-turn recovery, and an unrecoverable spin!
Lesson #1 - Don't assume that just because you did this spin several times before with no problems, that the next one will be the same. The second incident has caused me to rethink the scope of my planned spin test program. I haven't come to a conclusion yet, but I may drop some of the more obscure test conditions, and increase the number of repetitions on the more likely conditions.
He was eventually able to obtain a recovery by using a very different, unusual recovery technique.
Lesson #2 - Do some research on different recovery techniques, and be prepared to change recover techniques if the first one isn't working. But, don't be too quick to give up on the first technique. Your sense of time changes when you are under stress, and it is natural to believe that more time has passed than actually has. Also, as the angle of attach decreases during the recovery, the nose falls, and the axis of rotation becomes more aligned with the fuselage. The polar moment of inertia is less as when the axis is aligned with the fuselage, so the rate of rotation increases (exactly as it does when a spinning figure skater pulls in his arms). The increase in rate of rotation can fool you into believing that the spin is getting worse.
I'll do a second post for discussions about spin recovery techniques - this post is long enough already.
Rob had defined a bail-out altitude, and was prepared to abandon the aircraft if necessary.
Lesson #3 - You should also be prepared to bail out - the aircraft should have a means to jettison the canopy, you should wear a parachute and helmet, consider all the actions you need to take to egress the aircraft, physically and mentally practice those actions, and review them before every spin.
It may be a good idea to have a static line connecting the parachute D-ring to the aircraft. This will ensure that your chute opens, even if you hit your head as you egress.
About 2,500 ft was used from the altitude where the spin was started, to the altitude at the bottom of the recovery.
Lesson #4 - Give yourself lots of altitude when doing spin testing. The extra few minutes required to climb another few thousand feet are minor compared to the risk of losing the aircraft.
Edit - I want to make it absolutely clear that I am not criticizing Rob Erdos in any way. He was well prepared for the spin testing, the flight tests were well executed, and he did an outstanding job when things went off the rails.