Scheevel Aerodynamics of Q-2 Tri-Q-2 Design
- Category: Q2/Q200
- Published: Saturday, 02 May 2015 17:41
- Written by Jay Scheevel
- Hits: 16999
[EDITOR'S NOTE: Jay Scheevel has been working on the aerodynamic forces on the Q-2, Q-200, and Tri-Q almost as much as he has been working on his own highly modified Tri-Q-200. He has spent years gathering aerodynamic data from flying examples of these planes . . . and below he shares Part 1 and Part 2 of his findings.]
Aerodynamics of the Quickie Aircraft Corporation (QAC) Q-200 and Tri-Q-200
Part 1 – Introduction/Modeling Basics
This is Part 1 (of 3) of a follow-up to my 2009 study discussing Q-2 decalage (Q-Talk 138). This study focuses on aerodynamic computer modeling of the Q-200 tandem wing configuration and comparison of those models with flying aircraft. -- Jay Scheevel, Grand Junction, Colorado
For many years I had a pet theory that the range of cruise and landing behavior observed in flying Q-200's were primarily the result of variations in relative incidence angles between the wing and canard (decalage angle). This led me to hands-on measurements the decalage of more than 20 flying Q-200’s or Tri-Q200’s over a number of years and an attempt to study the impact on performance of those same aircraft using informal data supplied by the pilots/owners. Those results were summarized in 2009 (in Q-Talk #138). I will not discuss that paper here, other than to say that once I finished that study, I still felt like I lacked sufficient quantitative answers that would allow me to make sense of the influence of decalage to the flight qualities of the Q-200. Also, around this time, I became more keenly focused on landing and take-off behavior of the Q-200 as preparation for (hopefully soon) test-flying my own Tri-Q200. In addition to the influence of decalage, I also wanted to know the contribution of other key variables such as gross-weight, center of gravity, and landing gear configuration particularly as it impacts landing and takeoff behavior.
In order to answer these remaining questions in a satisfactory way, I decided to numerically model the aerodynamics of the Q-200 and Tri-Q200 configurations. Others have attempted to analyze the Q-200 using PC-based “flight simulator” programs such as X-plane, but that approach only yields a simulated flight experience, the real value of which depends on one’s ability to translate the computerized flight experience to real life. I do not find much similarity between actual flight and any PC “simulators”, so I see no the direct value in using a flight simulator version of the Q-200 to try to answer my detailed questions. Also, I wanted some numbers to crunch, so I went in another direction, one that would also allow me to communicate useful information to other Q-200 enthusiasts.
I decided to use a “2D” aerodynamic numerical modeling tool based on the well-known X-foil “panel” approach. X-foil was developed years ago at MIT. X-foil has been reworked by Martin Hepperle into a comprehensive Java-based user interface called JavaFoil. JavaFoil incorporates analysis, graphical editing and custom interface tools that are very powerful and complementary to the numerical modeling routines. In addition, JavaFoil can model multiple, mutually interacting airfoils (so called “multifoils”). Javafoil has the option of modeling any configuration and also has the capability to simulate ground effect. JavaFoil can run scripts allowing automated repetition of experiments with changes to parameters such as decalage, elevator deflection, reflexor deflection or dynamic parameters such as Reynolds number, wing loading. The scripts allow me to re-run models any number of times to test sensitivity of single parameter changes. The use of scripts also limits the possibility of manual entry errors resulting from one-at-a-time runs. Javafoils capabilities align nicely with the modeling that I wished to perform on the tandem wing configuration of the Q2.
2D Panel methods like X-foil are insufficient to model true stall and post-stall behavior (they approximate this). A 2D panel method also does not model 3D flow (span wise flow and finite wing behavior). My analysis does not depend strongly on the details of stall behavior, because the Q-200 main wing really never fully stalls and, as I will demonstrate, the canard normally cannot be forced into a full stall in the Q-200 configuration. So the modeling approximations of Javafoil are sufficient for my purposes. To approximate any 3D effects, I break down the Q wing into thirds and model them independently, then recombine the results in a spreadsheet with corrections added there. In this way I can handle washout (twist), sweep and span wise efficiency approximations. In summary, the JavaFoil platform has proven very adequate for my needs when combined with spreadsheet-hosted post-processing.
The spreadsheet is also where other calculations specific to the Q-200 geometry such as loading, and wing layout can be varied. Simplified non-wing drag moments (form, parasite, etc.) are introduced in spreadsheet calculations in order to achieve a match of the modeled performance to that of the entire flying aircraft with flight tests. All of the JavaFoil models and spreadsheet manipulations represent many tens of thousands of aerodynamic “experiments” on the virtual Q-200 “model”. Part 1 of this study (this section) is an introduction to the modeling procedure and unique considerations required to model the Q-200/Tri-Q200. Part 2 of this study is a summary of the Javafoil model results and how these results relate to take-off, landing, and level flight for various decalage values on the Q-200 and Tri-Q. Part 3 of this study is a review of flying aircraft from online videos and comparison of the Javafoil modeling results to actual flight behavior of those aircraft. The comparisons in Part 3 are done by dissecting photos and videos that have been posted on the internet by a number of Q pilot/owners. I use photos and performance data from the original QAC Q-200 prototype (N81QA) as well as photos and videos posted by Mike Dwyer (Q-200), Jerry Marstall (Tri-Q200), Sanjay Dall (Q-200) and Jean Paul Chevalier (Tri-Q200), all of whom have created nice point of view (POV) videos covering various facets of flight. My use of their videos IN NO WAY implies that any of these people have approved or any way endorsed any of my conclusions. Their videos are publicly available on the internet, so use of them to evaluate the validity of my modeling against actual flying Q-200’s or Tri-Q200’s is essentially use of public data.
I have chosen to present most of my conclusions graphically, so careful study of each graph, diagram and photograph along including its caption will aid the reader in the understanding of the parameters, model results and my conclusions.
DISCLAIMER: None of my results or conclusions should be taken as those generated by a trained aerodynamicist, which I in no way resemble. This study may provide qualitative insight into the unique flight characteristics of the Q-200 tandem wing configuration as well as the Tri- Q200 configuration during landing, take-off and level flight. Do not use any of this study’s results or conclusions to guide the building or test flying of any aircraft. This study is not advice for flight procedures for the Q-200 or Tri-Q200 unless you have personally and independently verified the results for yourself. This study has not been critically reviewed or verified so it cannot be regarded as scientifically reliable.
Aerodynamics of the Quickie Aircraft Corporation (QAC) Q-200 and Tri-Q-200
Part 2 – Modeling Results
Part 2 (of 3) of a follow-up to my 2009 study discussing Q-2 decalage (Q-Talk 138). This study focuses on aerodynamic computer modeling of the Q-2 tandem wing configuration and comparison of those models with flying aircraft. -- Jay Scheevel, Grand Junction, Colorado
Q-200 Flight Envelope Modeling
The underlying details concerning modeling methods and assumptions for the models discussed in Part 2 (this part) of this study are summarized in Part 1. If any concepts or methodology are unclear, please review the discussion in Part 1.
The goal in this part is to describe as much of the flight envelope as possible, including takeoff and landing. This includes varying control inputs and their influence on the resulting flight characteristics of both the Q200 and Tri-Q200 when calibrated to flying aircraft. As outlined in Part 1, the modeling procedure employs Javafoil to compute the flight forces resulting from the wing surfaces and using spreadsheet calculations to incorporate all drag moments including landing gear drag. Javafoil is appropriate for computing stable/equilibrium flight, so generalization of model results to dynamic flight, including rapid vertical accelerations and decelerations, or abrupt control input changes is not appropriate. Limited analysis of smooth transitions based on the equilibrium models may be justified, so limited discussion of smooth transitions to landing and takeoff is included in this study.
The study includes models of flight in an unbounded air mass and flight in ground effect. Understanding how ground effect impacts flight behavior is critical to understanding landing and takeoff performance of the Q-200/Tri-Q200 configurations. Ultimately I focus on the Tri-Q200 because that is the model that I am building, however ground effect modeling also applies to conventional Q-200 design and is reviewed completely in this study. The most rigorous flight test data that is available (used for model calibration) was collected in QAC’s prototype Q-200 N81QA, the original Q-200 tail dragger. A complete evaluation of the tail dragger Q-200 behavior is a prerequisite for analysis of the Tri-Q200. Flight test data is used to calibrate the model and quantify real world drag for both the Q-200 and the Tri-Q200. Once calibrated, the models are used to compare the flight characteristics of the two configurations in detail. The two landing gear variants differ as a result of both drag and lift forces primarily because the Tri-Q200 main gear’s drag and lift components differ significantly from those of the original Q-200 gear design. Part 3 of this study elaborates further on both variants (Q-200 and Tri-Q200) through analysis of the flights documented by point-of-view (POV) videos that have been posted to the web. As one might expect, landing and take-off behaviors of each design also differ because the Tri-Q200 must rotate in order to take off, whereas the conventional Q-2/Q-200 takes off essentially from a 3 point stance.
The Q-200 Flight Envelope
Figure 1 is a graph containing a series of colored curves which are generated by the calibrated model of the Q-200. Each colored curve represents the flight profile for a different decalage configuration. Any point on each colored curve represents an equilibrium flight condition where forces of both canard and wing (lift, drag and moment) combined with non-wing drag moments result in stable and level flight. The aircraft gross weight (900 pounds) is applied as a vertical force located at fuselage station (FS) and water level (WL) corresponding to the center of gravity (CG) of the aircraft. Each colored curve represents all possible combinations of elevator deflections and alphas (~AOA) that result in stable equilibrium flight for that decalage and the listed loading condition. The vertical axis of the graph is alpha, and the horizontal axis is elevator deflection. The black contour lines are lines of constant airspeed (CAS), each marked with the corresponding CAS in MPH. These CAS contours are in increments of 5 MPH with the thick contours representing 100, 150 and 200 MPH. The point where each black airspeed contour intersects a colored curve is the point where equilibrium flight conditions correspond to that CAS, alpha and elevator deflection. For example, consider the light green curve (zero decalage). The 150 CAS contour intersects the green curve at elevator deflection of +2 degrees and alpha of -2 degrees resulting in level equilibrium flight if all those conditions are met. The yellow stars on this graph represent the reported elevator/CAS combinations for flight tests in the prototype QAC Q- 200 N81QA and plot on top of the yellow curve corresponding to the model results for decalage of minus 1.
Aerodynamics of the Q-200/Tri-Q-200:
Part 3 – Analysis of flying Q’s
Part 3 (of 3) of a follow-up to my 2009 study discussing Q-2 decalage (Q-Talk 138). This study focuses on aerodynamic computer modeling of the Q-2 tandem wing configuration and comparison of those models with flying aircraft. -- Jay Scheevel, Grand Junction, Colorado
The aerodynamic modeling methods and assumptions in this Q-200/Tri-Q200 study are covered in Part 1. Part 2 discusses the flight envelope, including takeoff and landing, by employing the models to generalize about the flight behavior both in and out of ground effect. This part (Part 3) makes extensive use of flight videos gathered from the internet in order to do an evaluation of the Q-200 and Tri-Q200 flight behavior in actual flight conditions, and then to compare those flights with the modeling results from Part 2 of this study. The modeling results compare very favorably with the actual flight video evaluations and also shed light on the dynamic/transient behavior of the Q-200 and Tri-Q200 designs including variability between individual aircraft and the aircraft response to techniques of various pilots.
As one might expect, landing and take-off behaviors differ between Q-200 and Tri-Q200 primarily because the Tri-Q200 must rotate in order to take off, whereas the conventional Q-2/Q-200 takes off essentially from a 3 point stance. However, observations from videos also show there are substantial differences between the two variants in cruise flight apparently resulting from aerodynamic forces induced by the landing gear design of the Tri-Q200. The first part of the discussion will review the Q-200, followed by a discussion of the Tri-Q200.
The modeling from Part 2 is strictly for equilibrium conditions, meaning constant alpha and airspeed. but the videos of actual flights include periods of dynamic or transient conditions in addition to equilibrium ones. The non-equilibrium (dynamic) portions of the flight videos are instructive, especially during takeoff, but my analysis necessarily recognizes that these conditions depart from the modeling assumptions. Discussion of dynamic flight conditions is speculative, but may add qualitative insight.
The Q-200 Flight Envelope Model
Figure 1 is the same as one discussed in Part 2. The graph contains a series of colored curves which represent the equilibrium flight conditions for calibrated models of the Q-200. Each colored curve represents the flight profile for a different decalage. Any point on a colored curve represents an equilibrium flight condition where forces on both canard and wing (lift, drag and moments) combine with non-wing drag moments to result in stable flight. The aircraft gross weight (900 pounds) is applied as a vertical force located at the fuselage station (FS) and water level (WL) corresponding to the center of gravity (CG) of the aircraft. Each colored curve represents all possible combinations of airspeed and alpha (~AOA) for that decalage and the loading condition. The vertical axis of the graph is alpha. The horizontal axis is airspeed (CAS). The yellow stars on this graph represent the estimated alpha/CAS combinations for flight tests in the prototype QAC Q-200 N81QA and plot on top of the yellow curve corresponding to the javafoil-based model results for decalage of minus 1. The measurements are from Michael Huffman’s 1983 report on the QAC prototype Q-200 N81QA. Flight tests were conducted at a weight of ~900 pounds and CG of FS 44” based on description of loading used for flight tests. All reported data was collected during flight out of ground effect.
Figure 2 is the equivalent of Figure 1, but is the model response when in ground effect. The yellow stars shown on this figure are the same as those in Figure 1 (measure out of ground effect). The difference between the yellow curve and the yellow stars shows the significant impact of flight into ground effect. This is most apparent at low airspeeds where alpha values are significantly reduced from those when flying out of ground effect. Maximum achievable alpha at pitch buck in ground effect is less than 7 degrees, whereas out of ground effect it is near 10 degrees for the modeled CG and gross weight.
Use of videos to evaluate Q-200 Flight Behavior
In this part of the study we evaluate two currently flying Q-200’s through the use of posted internet videos. The first of these is Mike Dwyer’s Q-200 (N3QP). N3QP has several decades and more than 1000 hours of flight time. Dwyer’s videos are recorded in Florida (near sea level), so provide very useful standard atmospheric condition comparisons to modeling results. The second flying Q-200 was built and flown by Sanjay Dahll. Dahll’s plane (N102SD) has been flying for close to 4 years and has an estimated 150+ hours. It is flown in Michigan (ground elevation 600-700 feet above sea level).
Some of the videos prepared by Dwyer for flights in N3QP include strip-type speed and altitude data synchronized with the video (using Dashcam), these numbers are informative for analysis of all phases of flight. Dahll’s videos for N102SD do not include speed or altitude data, but like other videos used here, unique runway markings can be identified and distances between those markings can be measured accurately using Google earth. Then interval-timing of these markings on the real-time video allows ground speed to be computed during the take-off. Airspeed (CAS) from groundspeed can be estimated based on assumptions of atmospheric conditions during takeoff and flight. This includes consideration of density altitude and wind conditions.
Using specific angular relationships between airframe features, as seen from the camera point of view (POV), an angular scale can be constructed and then superimposed on the video frame capture images. By comparison of this scale to the distant horizon, the alpha angle during any phase of level flight can be estimated. Combined with the visually measured alpha values, and in some cases with video observations of elevator and reflexor configurations, meaningful comparisons with the Javafoil models from Part 2 can be made.