Design Group -- Bearhawk Non-Modifications

Russ Erb

Originally published September 1997

Bob Thomasson sent me some excellent questions about proposed modifications to the Bearhawk a couple of weeks back. While I recommended NOT doing most of them, I think you may learn something useful reading this and realize what all becomes involved when you think about making a "small" modification. The big one to watch out for is what I call the "design propagation nightmare" which can happen with just about any modification. Please remember that unless you know at least as much as the designer about aircraft design (and some of us may), it's best not to mess with it.

I have included Bob's original questions with his permission. Hope you enjoy it. I enjoyed putting it together and it gave me something to do on a long transcontinental airline trip other than just "eating my way across the country."

Russ,

I was browsing the Chapter 1000 web page and discovered that you're building a Bearhawk. I've bought the plans, but haven't started on it yet. I have a couple of aerodynamic type questions. The purpose of my plane will be mostly long trips with heavy loads to Alaska, the Northwest Territories, etc., with floats eventually. I don't see trying to land on very many 100' long sandbars unless I have to. Efficient cruise speed is probably a higher priority than extreme STOL capabilities.

What do you think of the Bearhawk wing airfoil? At Oshkosh I noted that there are endless variations in airfoils for similar type planes. Some have flat profiles on the wing bottom, some are concave and some convex. The Glastar wing is the most interesting to me, with the same airfoil as the Glasair with constant chord. It seems to work well and the Fowler flaps evidently help with STOL. How come nobody else has used this type wing with a "bush type" aircraft? I don't have any specific problem with the existing Bearhawk airfoil, I'm just hesitant to start bending metal when I don't have much of an understanding of the wing. I'm trying to avoid expending 9 million hours of labor only to decide I really wanted another wing airfoil. Did you modify yours at all, or consider any changes?

Same goes for those huge 50 degree flaps. I asked why the flaps don't start until about a foot away from the fuselage when all the other flaps I've looked at start very close to the wing root. The Bearhawk people told me that the flaps are built like that for ease of construction and had a surprise benefit. The propwash flows through the gap to the tail when the flaps are deployed and gives better low speed controllability. What do you think? It also seems to me that the flaps might be too wide and with the way they are hinged will sacrifice too much in wing area when deployed - only conjecture, but what do you think? I'm thinking of spending the extra time to build Fowler flaps. Did you modify the flaps in your aircraft?

Other changes I'm considering are some windows in the roof and a Wittman type gear, since I don't anticipate too much real rough field landings.

Thanks for your time, and I understand if you're too busy to reply. I really enjoyed the Chapter 1000 web site. Good luck with your Bearhawk. What are you going to power it with?

Regards, Bob Thomasson EAA 474976

Bob-

You've asked several excellent questions. That shows you're thinking, which is a good thing.

The short answer:

Don't change anything (with one exception).

The long answer:

Glasair/GlaStar Airfoil:

The airfoil you refer to in the Glasair and GlaStar wing is either the GAW-1 or GAW-2. A friend of mine refers to it as the "Gawd Awful Wing," and everyone I've talked to with opinions I respect feels that this airfoil is highly overrated. It has a very high pitching moment, which will factor into the stability and control of the aircraft. For instance, it would probably require a readjustment in the incidence angle of the horizontal tail or the wing incidence angle. Are you ready to deal with that?

Another problem with the GAW series of airfoils is they are very sensitive to proper shape. In other words, the wing skin must be rather stiff to maintain the shape of the airfoil in order to get the expected performance. On the Glasair, the composite skins are probably sufficiently stiff to accomplish this. You may be thinking "But the GlaStar uses an aluminum wing." But that doesn't mean it's right. I'm guessing that Stoddard-Hamilton used the GAW airfoil on the GlaStar simply because that's what they were used to on the Glasair. Another concern I have with the GlaStar wing is that it uses primarily hat section stringers to maintain the shape of the wing instead of ribs. A local GlaStar builder told me that he did not find a single such stringer in his kit that was the proper shape. He had to make a forming block to force each of them into the proper shape. Since the stringers and skins are pre-punched, there is a reasonable chance that the wing will eventually get to the proper shape, at least to begin with. I'm not yet convinced about how stiff the wing skins will be.

An increasingly more publicized "failure" of an implementation the GAW airfoils is the Piper Tomahawk. When Piper was designing the Tomahawk, they were looking for any advantage they could gain over the Cessna 150 while using the same engine. One of the things they did was to use the "new" GAW airfoil to try to get slightly better performance. This worked out reasonably well on the prototype, which was used to do the certification. Unfortunately, in a story repeated far too often in history, the production engineers thought they were smarter than the design engineers and built the production aircraft with about half as many ribs as the wing originally had. (During a sheet metal workshop, I had the opportunity to de-skin a Tomahawk wing, and was surprised at how few ribs it had. This was before I learned this story.) Sounds good: less weight, lower parts count, right? Wrong! Remember what I said about this airfoil being very sensitive to the proper shape? The end result was that the skins were not stiff enough and would "oil can" under air loads, disrupting the airfoil shape. Wing bending under aero loads would also distort the airfoil. The biggest problem was with the stall characteristics. In the prototype, they were acceptable. On the production birds, the stall was unpredictable and would change from time to time. I did my flight training in a Tomahawk, and have most of my logged time in one (again, this was before I learned about the wing problem). Stalls in a Tomahawk are not a nice, gentle g-break like a Cessna 150. Instead they were characterized by a fairly sharp (violent?) wing drop, which seemed like at least 45° of bank and was unpredictable in direction. I'm not sure about FAR 23, but I'm pretty sure that the stall characteristics would fail the appropriate Mil Specs. My flight instructor tried to convince me that it was designed that way to "improve training," but I don't buy that anymore. I certainly wouldn't want that sort of "training" low to the ground during a botched turn to final, whereas a g-break would probably be recoverable. The clearest indicator to me was that I was sufficiently scared of the stall characteristics that I refused to practice any stalls in the airplane after I received my certificate. On the other hand, stalls in other aircraft, such as Cessnas or even the Piper Cherokee series are non-events. I recently read that the NTSB is calling for a re-certification of the Tomahawk stall and spin characteristics. The stories I have read correlate well with what I remember. I may have close to 100 hours in the Tomahawk, but based on what I know now, I really have no desire to ever fly in one again. Are you sure you want this airfoil?

One last thing on the GAW airfoils: the undercamber on the lower surface (the concave section) will increase the difficulty of construction and require redesign of the flaps and ailerons. The ailerons on the Bearhawk are of a very nice Frise aileron design. You will note that when the aileron is deflected trailing edge up, the nose of the aileron will poke out the bottom of the wing. The added drag of this nose offsets the additional drag of the opposite aileron (deflected trailing edge down, increasing lift and thus induced drag) and results in a reduction of adverse yaw. That means less rudder is required to coordinate turns. I wouldn't want to mess that up.

Airfoil as designed:

Bob Barrows told me that the Bearhawk airfoil is essentially an NACA 4412 airfoil, which was actually what I expected. This is a very well known, time-tested and proven airfoil with well documented characteristics. You can find it in the classic airfoil text "Theory of Wing Sections" by Abbott and Von Doenhoff (Dover Publications). It's a blue paperback book about 1-1/2" thick that just about any aeronautical engineer will own.

The NACA 4412 has a turbulent boundary layer, which is actually a good thing for an aluminum wing. You may have heard a lot of hype about laminar flow airfoils and their lower drag, especially with respect to composite aircraft. The whole laminar flow business got started back in the early days of WWII, and is best remembered with respect to the development of the P-51. The P-51 was the first aircraft to be designed with an early laminar flow airfoil, which was supposed to give large gains in performance. However, it didn't turn out quite like it was planned. Laminar flow airfoils require very smooth surfaces to work properly. The P-51 had an aluminum wing constructed traditional techniques, and all of those flush rivets and any waviness in the skins were enough to trip the boundary layer to turbulent flow. If not in the beginning, certainly after a little bit of hangar rash and dirt got on the surface. Traditionally, laminar flow airfoils have not worked well with aluminum structures. They have worked with composite structures because the skin could be made extremely smooth and stiff.

The turbulent boundary layer actually helps keep the flow attached to the surface and improves the stall characteristics. Laminar flow airfoils tend to have lower stall angles of attack because of the sharper leading edge and the fact that laminar boundary layers are not as likely to stay attached. The NACA 4412 has a nice rounded leading edge, and a two-dimensional unflapped stall angle of attack of 12 to 16 degrees. Why do I keep harping on stall characteristics? First, a high stall angle of attack indicates that the airfoil will be able to operate at a high angle of attack, which is necessary for STOL. Second, because a STOL aircraft will land and take off with a reduced margin above stall speed, it is imperative that the aircraft have gentle stall characteristics so that you can get yourself out of trouble shouldst you get into it.

Another nice benefit is that the bottom surface is almost flat, which makes the construction easier.

The 4412 has a design lift coefficient (i.e. minimum drag lift coefficient) of 0.4, which for a 2300 lb Bearhawk, would occur at 97 KCAS. I'm expecting with a 220 HP engine a cruise speed of 140 KCAS. If the aircraft was to get heavier, the profile drag of the airfoil would actually decrease as the lift coefficient increased toward 0.4. Of course, the induced drag will increase, but the overall drag wouldn't change much. With this airfoil, there is still "growth" room.

As it turns out, airfoil selection has very small effect on aircraft performance compared to other factors. Generally you don't turn to tweaking airfoil performance until everything else has been optimized, because the improvements realized are typically no more than a percent or two. (As a side note, wing tips are another area that seems to be popular for changing, but again usually with small or no improvement, and sometimes a degradation in performance. After all, if it was so easy to improve, why wouldn't the designer do it that way in the first place?) Read the August 1997 Sport Aviation article on the Nemesis and Shadow on page 75. Note how small a gain is expected for using different airfoils. In air racing where fractions of a knot matter, the results are noticeable. For the type of flying I expect to do, you'll never notice the difference.

Changing the airfoil leads to a design propagation nightmare. When you change the airfoil, you'll probably change the thickness of the wing at the spar location. Thus, the spars won't be same height, which means you would have to redesign both spars. Assuming that wing got thinner, the spar would be heavier to have the same strength (weight is almost always bad). Well, when you changed the spars, you changed the whole structural analysis, and pretty soon you’ve designed an entirely new wing with unknown structural characteristics. In addition, you would have to redesign the flaps and ailerons to fit the airfoil, along with the flight control cables, pulleys, bellcranks, and pushrods. The designer won’t recognize it as a "Bearhawk," and you'll be on your own as far as builder support. If you really want to do that, then design your own airplane. At least then you'll understand the entire system and the tradeoffs involved. If you "don't have much of an understanding of the wing" the last thing you want to do is to start redesigning major components!

You asked "How come nobody else has used this type wing with a "bush type" aircraft?" I think by now you should be convinced the reason is because it is not suitable for the "bush type" mission. Remember that the Glasair has very different mission in life. If one airfoil were truly the "best" for everything, then why do we have so many different airfoil designs?

Flaps:

Let's start this discussion by reviewing the purpose of flaps:
  1. To increase lift (reduce stall speed)
  2. To increase drag to allow steeper approaches without gaining airspeed
  3. To lower the deck angle on approach so that you have a chance of seeing over the cowling

Starting the flaps close to the fuselage would only gain about 9 percent more flap area, which would probably result in less than 1 knot stall speed reduction. The current design allows a good integration of the wing root with the fuselage. While I can’t verify the claim that it improves low speed controllability, it does not sound unreasonable based on my other experiences. Aeronca had problems in the past with flaps that came all the way to the wing root causing buffeting on the tail.

Deflecting the flaps does not result in a loss of wing area. While the planform area seems to be reduced (what you would see from above), the chord that the air "sees" doesn’t change when flaps deflected. It just exists in more of a curved path. Overall lift and drag still go up when the flaps are deflected, otherwise the flaps wouldn't be deflected that far.

The flaps may seem large to you, but they are only about 23 percent chord, and the typical size of flaps for most aircraft is 20 to 30 percent. According to Budd Davisson's article on the Bearhawk in the October 1995 Sport Aviation (that convinced me to order the plans), Bob Barrows originally considered Fowler flaps for the Bearhawk, but decided to go with plain flaps for simplicity (read: reliability) and the resulting small change in longitudinal trim. Fowler flaps cause a much larger change in pitching moment, which results in the pilot having to make bigger trim changes when extending the flaps. Using Fowler flaps would also require extending the upper surface of the wing back to where the flaps would roll back to. This extension would require designing additional structure (here we go again), which would take up room in the airfoil, forcing the flaps to be thinner, resulting in less strength. Then, how would you mechanize them? The two basic choices: 1) external hinges, which result in increased drag and weight, not to mention a hazard on the ground for you to bash your head into, or 2) Cessna style internal flap tracks, which are a design nightmare, and would probably lead to serious questions of strength and fatigue problems. I figured out how much reduction in stall speed you would expect to get after all of this work according to a mathematical model I have built of the Bearhawk: about 1 knot! That's way too much effort with no noticeable payoff for me, not to mention the maintenance headaches you've created for yourself. Remember the designer's credo: Keep it simple, keep it light.

Windows in roof:

Here's the one area that I would agree with you. The April 1996 issue of Beartracks (the Bearhawk builder’s newsletter) has already approved installation of a skylight back to the rear spar. I still haven’t decided if I want to do this. Flying around in fighter-type aircraft with big bubble canopies I find that the sun still makes it awful hot. It's wonderful if you want a greenhouse, but bad for comfort. Living in the desert, this is a significant problem. I've noticed many Navions and other aircraft with canopies that allow you to see straight up have some sort of sun shade over the pilot and passengers.

Wittman gear:

I wouldn’t recommend it. While it is true that it is very simple, it has virtually no inherent damping (shock absorption). As a result, when you try to plant the mains on the ground in any sort of landing, you'll probably bounce a few times. The landing gear as designed does very well for this type of aircraft. It is very clean and very appropriate. It is not just a bungee system like a Cub. The Bearhawk gear as designed has an integral spring and damper in the strut that keeps the gear from spreading. Thus, you can plant it reasonably hard without bouncing. This system is very similar to what is used on the Maule, and has worked very well for that aircraft. If you are concerned about the additional drag of the additional support arm, then add a lightweight non-structural streamline fairing over it.

Another point to consider is that the Wittman gear will transmit the loads to the fuselage in different amounts at different points than the gear as designed. I'm not sure how this would affect it, but I'm pretty sure it wouldn't be good.

As for rough field landings, it doesn't matter if you don't expect too many. One rough field landing is all you need to bounce the landing and possibly damage something. The landing gear as originally designed is much more tolerant of this type of operation. I still recommend it.

In general, as an aeronautical engineer with a background in aircraft design, and having a father who has much more experience in aircraft design and is more concerned with analyzing the design than I have been, you would expect me to critically evaluate the design of any aircraft I considered building. I have been pleasantly surprised and impressed with the design of the Bearhawk, finding it to be very sound. In fact, so far the only modification I have made is to add a landing light in the wing leading edge outboard of rib 10. Prior to doing this, I checked with Bob Barrows for the best location to do that. I realize this will reduce the wing strength slightly, but we are working hard to minimize the impact and the propagating design changes. It has resulted in moving one of the nose ribs in each wing about 2 to 3 inches, which I wasn't real thrilled about. I also plan to add navigation lights to the wingtips and tail, but those should have minimal impact on the structure of the aircraft.

Floats:

In the April 1996 issue of Beartracks, Bob recommends using thicker wall tubing in a couple of fuselage locations if floats will be installed. I plan to use the thicker tubing in these locations, even though I don't plan to add floats at this time. If I decide to later, I'm set. If not, it's only a weight penalty of a pound or two. It's far easier to build in the thicker tubes than to uncover the fuselage and cut it apart to retrofit them later.

Engine:

I'm not a big fan of auto conversion engines. I'm not much of a gear head. I can do maintenance on engines, but I don't know much about modifying them. I don't care to get into propulsion development, which is what would happen unless I could buy a turnkey system. Additionally, I have yet to see a conversion that is as light as an aircraft engine, or that would fit into the current cowling design. I really like the sleek cowling as it is, rather than the "flat wall" behind the propeller typical of Cessna 172s and many auto conversions.

However, Lycoming and Continentals are ridiculously overpriced. I don't care to search around, hoping that the right engine will pop up in salvage. I'm currently leaning toward the Franklin 220, which is a 220 HP engine of American design which is now being produced again by PZL in Poland. It's price new is about half of a new Lycoming. There are US distributors, so I won't have to deal with the mess of importing it.

Even so, it is still years until I have to make an engine decision, and I'm keeping an open mind about it. New and exciting developments are currently ongoing.

I'm also leaning toward a constant speed prop. I do expect to do a lot of operations on a very short (660 - 975 feet) grass field, and the increase in takeoff performance will justify the additional weight, complexity, and cost.


EAA Chapter 1000 Home Page
E-Mail: Web Site Director Russ Erb at erbman@pobox.com

URL: http://www.eaa1000.av.org/desingrp/bearhawk.htm
Contents of The Leading Edge and these web pages are the viewpoints of the authors. No claim is made and no liability is assumed, expressed or implied as to the technical accuracy or safety of the material presented. The viewpoints expressed are not necessarily those of Chapter 1000 or the Experimental Aircraft Association.
Revised -- 8 April 1998