For several years, we have been trying to obtain information and data on the characteristics of various canard-types at deep stall conditions. Data for the VariEze has been available since the late 70's when NASA conducted rotary-balance wind tunnel tests and concluded that the VariEze has no stable spin modes, i.e., that if forced to any angle of attack and spin rate, it will recover by itself. Also, the small model tests showed normal flat-plate drag at high angles of attack. These data and extensive stall-departure flight tests with N4EZ formed the basis for our confidence in the deep-stall safety of these general aircraft types.
Then, about 5 years ago, several accidents occurred with the Velocity aircraft. We think the problem could have been determined if extensive aft-CG departure testing had been done during development, like we did with the Long-EZ and Defiant. Two very noteworthy results from these Velocity accidents were
1). The descent
was a stable, non-rotating condition about 50 to 80 degrees AOA, not recoverable with
forward stick or by rocking the wings.
2). The descent was slow enough to allow impact in water without pilot injury.
Rumors were abound about a this slow, 1000 ft./min. "Parachute-like" descent, probably induced by a violent trapped vortex above the wing. Researching this, we found the rumors were just speculation, that their was no hard data on the descent rate. Even the test pilot who stayed with a Velocity to the ocean instead of using his parachute admitted he had not aimed the altimeter nor remembered the rate-of-climb indicator's data. He merely climbed partially out, but feeling the "light breeze" of the descent, elected to ride it down. We have been extremely skeptical that an airplane can descent at this low rate, even with the best possible vortex.
To put things in perspective, consider what would be required. The EZs and the Velocity have a "loading" of about 10 lb. per square foot of total planform area (including wings, canard, fusel I age strakes and cowl). If all this area acts like a ' flat plate" in the descent, the airplane would sink at 50 knots or 5000 ft./min. (flat plate Cd=1.24). The very highest Cd we have seen in aerodynamic research papers on trapped vortex is about I Using a Cd of 10 for the entire airplane (very unlikely, of course), the sink rate would be 17 knots or 1800 ft./min. If the airplane could descend flat at 1000 ft./min. (only 9.9 knots), it would have a Cd over 30!!
Our interest in this phenomena certainly was increased after the deep stall accident of a LongEZ at Kanab (CP 68). Now we had some data, but very poor data. Only a tiny image of the airplane during the last 2.8 seconds on a video tape. This airplane hit the dirt without killing the pilot so we believed it could not have been descending at 5000 ft./min. An attempt to analyze the video resulted in a very rough approximation of 2900 ft./min. which results in a Cd of 3.7. Our surprise, of course, was that forward stick did not recover from the deep stall. The surprise subsided when we later learned that the airplane was being flown with the CG well aft of the FS 103 aft limit.
While the 2900 ft./min. sink estimate seemed to make sense, it was not considered accurate due to the problem of measuring a fuzzy blip on the video. We then made a decision to try to gather full scale data on the Long-EZ. The previous full scale tests done in Florida on the Velocity did not measure drag and lift, only the more important data of recoverability with various airplane modifications.
Then, another Velocity deep stall accident occurred. This one descended inverted, hit land, not water, and killed the pilot. In this accident data was available - good, accurate radar and transponder data. Obtaining this data from the FAA is a story in itself.. Finally, after threatening a media expose about government cover-up, we received the data. This Velocity entered a deep stall at about 7000 ft. and descended at a nearly constant 4400 ft./min. (4'4 knots) for the entire 90 seconds to impact. Of course, this inverted descent data may not apply to an upright Velocity but, at least, for the first time it represented good data during a deep stall accident.
We proceeded to develop the rig to allow us to measure the Long-EZ. This turned out to be a much more difficult and expensive job than originally thought. It was made possible by the loan from Donald Douglas of Sherman Oaks, CA of his Long-EZ that is accurately built to the plans, without modifications. A 3-axis electronic balance was built to measure lift, drag and pitching moment and an accurate speed indicator was installed in front of an Isuzu truck. These "Truck-EZ" tests can only be done in dead calm winds, so after many delays, we were able to obtain data at 40, 50, 60, 71 & 80 degrees angle of attack.
The data are presented in this newsletter. Note that these are full-scale tests at near the same Reynolds number as flight, so they are much more accurate than the small scale model tests done by NASA in the 70's.
First, let's discuss the lift and drag data. The data show substantial scatter due to the truck riding over bumps in the runway. A line faired through the average of scatter is considered reliable. If we combine the lift and drag resolved to a total reaction that would support the airplane during a stable deep stall descent, we can calculate the sink rate. This data, sink rate vs. angle of attack, is shown. Note that this prediction is very close to the radar data of the Velocity (4400 fpm).
Now, how slow does a Velocity descend upright in the deep stall attitude? We don't know, but we now tend to suspect that it is relatively high, 3500 to 4500 ft./min. We reason that the low damage and pilot survival is related to the fact that the water impact is nose down and the bottom fuselage is curved, this allows a few feet of deceleration at impact which can explain the lack of pilot injury.
How slow does a Long-EZ descend in a deep stall attitude? First of all, our pitching moment data show that it cannot descend at the extremely flat attitude of 70 to 90 degrees angle of attack. The pitch data indicated that if the CG is aft of limit, say F.S. 106, the aircraft may hang up at about 40 to 50 degrees angle of attack. It would then descent at about 5000 feet per minute. Why did the Kanab pilot survive? Possibly the nose-low attitude allowed a couple of feet of "crush and rotate" deceleration that provided adequate protection.
Our concern now is that there are many Long-EZs with extensive modifications that can affect deep stall recovery (long noses, bigger strakes, baggage pods, etc.). While we do not approve these modifications and can't be expected to analyze or test each one, we do feel obliged to encourage everyone to conduct adequate testing to determine the safety of their own modified airplane. Conduct stall tests at the CGs you fly, with adequate altitude for a parachute jump (egress above 8000 ft. AGL). Do not ride it down, even over water.
Another concern is that many of you do not accurately know your CG position. Calculating weight and balance is a pilot's responsibility (FAR 21) for each flight. Be sure you fly within limits (your own test-verified limits for modified airplanes) and check CG when any changes are made.
(Click on charts below to enlarge)