WWI Replica Safety

[Machinbird]

I have hijacked Mustang's posting and placed it here for discusson.

Originally Posted by mustang

The recent series of articles on flutter in the replica aircraft thread has ceased for the time so I thought it might be timely to comment on a personal experience which had tragic results and invite further comments from members who may have had similar experiences.

I spent some time as a production test pilot on the GAF N22 and N24 Nomad, a small twin-engined STOL aircraft designed for outback operations on bush airstrips. The N24, the stretched fuselage version, exhibited on test longitudinal instability problems in some configurations of flap and power settings. The reasons were complex and I won't go into them in this post.

One of the "fixes" tried was to attach Al alloy "tee" strips to the trailing edge of the tailplan in incremental lengths, Initial tests showed some improvement and further tests were scheduled, progressively increasing the length of the strips until a satisfactory result was obtained, or some other solution to the problem had to be considered.

Accordingly, a stip of increased length was added and another flight test scheduled. The result was disastrous. As the aircraft passed through about 90Kts on take-off violent flutter developed in the tailplane causing severe structural damage and loss of pitch control.

The subsequent crash killed the senior experimental test pilot and the design engineer in the right hand seat. The flight test engineer survived, as a paraplegic. Flutter can be deadly and any aircraft which exhibits the slightest indications that it may develop should be either grounded or speed restricted until an aerodynamic and aeroelastic analysis is done and certainly no mass should be added to any control surface aft of the hinge line without expert engineering advice

Any pilot who ignores the initial symptoms of the onset of flutter will need to ensure that his will and insurance are up to date.

Mustang

As I understand it, the "T" shaped extrusion trick on the trailing edge of a control surface is used where there a wide deadband in a control surface effectiveness caused by a thick boundary layer. Obviously, riveting in a strip of this extrusion into the trailing edge of a control surface moves the C.G back quite a bit and would require adjusting the static balance of the control surface by adding additional counterweight. Perhaps the people running this test didn't appreciate how much they were affecting the static balance of the elevator.
Ideally, you would want every inch of a control surface balanced so that accelerations would not displace the control surface and would not twist the control surface. In practice, this degree of counterbalance distribution is impractical, and designers cluster counterbalance mass in a few "horns" forward of the hinge line on the control surface. When you do this, each balanced section of the control surface can torsionally oscillate relative to the adjacent counterbalanced sections. These torsional oscillation frequencies need to be evaluated for potential of causing flutter.
A crop-duster/FBO/IA Mechanic/sometimes ferry pilot told me of a scare he had ferrying one of Cessna's popular 4 seaters. The plane had been sitting unloved in South Florida for a considerable period and had then been sold. My friend was hired to ferry the bird and he checked for the obvious things to watch out for, particularly nests, and things that might affect the powerplant.
Very shortly after takeoff the ailerons began flailing wildly. He slowed to minimum airspeed, brought the plane around the pattern and landed.
The aileron counter weights were long strips of lead sandwiched between aluminum strips and riveted to the leading edge of the aileron lower skin. Dissimilar metal corrosion had caused the attaching rivets to fail and one of the counterweights had been lost. Proper mass balance of control surfaces is crucial.
Sid

[Machinbird]

Originally Posted by N4085B

Thanks for posting that nasa video. I learned how to fly in a Comanche 180 (single engine version of the aircraft in the video) ...and that was truly sobering to see!

Glad you found the video interesting. As I recall, there was an airspeed calibration problem on the Twin Comanche and it was actually flying in excess of red line.
Sid

[Machinbird]

If you had a plane with a flutter problem, how would you investigate it?
If you only had a problem at the high end of the airspeed envelope, the simplest fix would be to draw the red line on your airspeed indicator (if you have one) a safe amount below the problem area. Anything else is going to require the application of some Do-Re-Me $$$.
If you have a flutter problem that blocks you from the heart of your operating envelope, you either have a museum piece or you will have to figure out and fix the problem.
Modern electronics has brought the price of things like accelerometers down into the accessible range. You would wire up a bunch of light weight accelerometers to a computer to collect data. You would decouple the aircraft from the ground by loosening the bungee cords until the aircraft was resting only on the cords and not on either the up stops or the down stops. Also put some soft rubber under the tail skid. The final piece of equipment would be an electric motor driven eccentric weight that would need to be firmly attached to aircraft structure at various locations during the test. You would have to be able to control the speed of the electric motor accurately.
What you are looking for would be the resonance frequencies and modes of your particular aircraft. You would probably start by looking at your control surfaces and determining how they respond to vibration. You would then move on to major structural elements. You could also use a visible medium such as a light coating of dust to observe the nodes (relatively stationary areas) of the vibration modes. The first five modes in ascending frequency order are probably sufficient. You need to move the exciting eccentric weight system around just to be sure you didn't accidentally couple up to a nodal point.

When you finish this exercise, you would have a fair idea of your aircraft's vibration modes and frequencies. You probably would not be able to discover unusual coupling situations (like the 747 model wagging its tail and twisting the wings). Probably the best way to record this data would be on a series of maps of the structure with the nodes and anti-nodes drawn in.

Before you proceed further, you might want to investigate adding mass balance to control surfaces to reduce their response to vibration. You would then retest the new configuration.
The next step is to calculate what exciting frequencies are likely based on Reynolds numbers your plane would be operating at and the dimensions of the surfaces. You would compare these exciting frequencies to the resonance frequencies that you had just measured on the actual aircraft.
Ideally the exciting frequencies and all the resonant frequencies will be significantly different, but if not, you have identified a potential problem requiring correction or at least careful evaluation.
Finally, you would move to test flying using exciting devices such as the one described here: http://ntrs.nasa.gov/archive/nasa/ca...2002119573.pdf
Here is where the real dollars come in. Potentially, this type of equipment can damp oscillation as well as excite it, so it could add an additional layer of safety in the test flying program.
Warning, use the above information with care. I don't do this for a living, and I may have understated or mis-stated important information. To the best of my knowledge, this is what people who cannot afford flutter tunnel testing do.
Sid

[hank jarrett]

Sid,
Did you move to Chicago when Boeing moved their headquarters?
I have worked "with" a lot of vibration and flutter guys, and even had them work for me back in my pre NASA NAVAIR days. That was a VERY good explanation, and clear, understandable explanations of flutter dynamics are hard to come by.
You also made a pretty clear case why folks like us aren't going to be doing much flutter analysis and testing.

Hank
If you aren't a flutter engineer, are you married to one?

[Machinbird]

Originally Posted by hank jarrett

Sid,
Did you move to Chicago when Boeing moved their headquarters?
I have worked "with" a lot of vibration and flutter guys, and even had them work for me back in my pre NASA NAVAIR days. That was a VERY good explanation, and clear, understandable explanations of flutter dynamics are hard to come by.
You also made a pretty clear case why folks like us aren't going to be doing much flutter analysis and testing.

Hank
If you aren't a flutter engineer, are you married to one?

Sorry to disappoint you Hank, my wife was a librarian and school teacher.
I haven't ever worked at an aerospace company. I wanted to build a light jet at one time and studied how I might determine the flutter limits. There is probably a lot I left out in my explanation.
Actually folks like us can do an initial resonance surveys of an aircraft with some relatively simple test equipment and a little creativity. Just verifying that your control surfaces are unlikely to contribute to flutter is probably enough on all but the most flexible WWI airframes. Your eccentric weight could be driven by an electric drill driving a flexible shaft. You would watch for brief resonances as the drill accelerated/decelerated the eccentric weight assembly. As I mentioned earlier, small accelerometers are now affordable and could be used to determine the resonant structural frequencies. It is just a physics lab experiment on a little bit bigger scale.
Sid

[hank jarrett]

There is still a lot we don't know or understand about flutter and predicting it, that's why we still have to do flight testing. Even then we occasionally don't get it right, and sometimes when we get it wrong we loose a vehicle (and sometimes a test pilot). One nice thing is the planes we mess with here are pretty slow and the danger from flutter goes up FAST with speed. If you are flying and feel a buzz in the airframe, the best thing you can do is pull the throttle back and SLOW DOWN!
Seeing that you have teachers in the family explains a lot about your explanation. Teachers are the experts in taking complex subjects and making them easier to understand for the rest of us. NASA engineers are about the worst.
Hank

My wife started out as a teacher too. I know "the look" when I start getting too technical.

[j ferguson]

John McKensie's very wonderful reports on his manufacturing of an FE2b include photos of a "sight gage" on the day-tank.

This is a vertical glass tube, housed in a bronze fitting for protection and installed in a niche on the side of the tank so that the level of its contents might be appraised.

I've seen this arrangement before. How about you?

Grumman Geese (Gooses?) had such sight gages on the wing tanks located over your shoulder. and over your head above the cockpit aft bulkhead.

On first seeing this arrangement, I asked if maybe there might have been a "safer" way to gauge the fuel.

I was told that this was a very good method since you could actually SEE the fuel - no mystery there.

The plane's owner also pointed out that the maintenance logs were in the pockets on the seat-backs.

This was all part of the system devised by Grumman to ensure that any accident arising out a sudden stop on arrival could be blamed on pilot-error.

The cockpit fire accompanying the breaking of the glass sight gages and the sudden contribution to combustion of the escaping av-gas would assure the immolation of the pilots and the logs leaving no alternative to this conclusion.

Imagine. Some of you might have thought this was only to make it easier to guess how much gas you had.

[Machinbird]

Originally Posted by j ferguson

John McKensie's very wonderful reports on his manufacturing of an FE2b include photos of a "sight gage" on the day-tank.

This is a vertical glass tube, housed in a bronze fitting for protection and installed in a niche on the side of the tank so that the level of its contents might be appraised.
.................

Hi John. What is a "day-tank"? From a risk standpoint, the sight glass isn't such a bad deal. The worst risk would be an inflight fire from relatively minor fuel leakage. Accurately knowing your fuel level is extremely valuable information for preventing fuel exhaustion accidents.
My Luscombe had a small window in the fuselage tank with a moving scale attached to a float behind the window. About 40 min after takeoff, I would see the fuel level in the glass and could get a verification of how accurately the tank had been filled. When you got really low on gas, you would see the bottom of the scale and it wouldn't move around too much with rudder or negative g. Only time I had trouble with it was when the cork float filled with gas and sank but you knew before flight that it wasn't working.
Switching back to Test Flying your bird, I would like to suggest two very good books on the subject.
The first is "Flight Testing Homebuilt Aircraft" by Vaughan Askue ISBN 978-1-56027-628-9. This book would be appropriate to the majority of the test flying programs you builders would be accomplishing. It has a practical down to earth style and nothing worse than text and easy to understand diagrams. It even addresses the questions of who should test fly your aircraft, and how to handle flutter.
If you have some sort of serious/unusual flight test problem, or you just enjoy understanding the theoretical aspects of things, then you should read "Flight Testing of Fixed-Wing Aircraft" by Ralph D. Kimberlin ISBN 1-56347-564-2. This is an actual textbook on the subject and is complete with Greek letters and formulas.
Both authors have your personal safety during the test flight program in mind and a provide a number of excellent suggestions for surviving the experience.
Sid

[j ferguson]

Hi Sid,
A day tank is an intermediate tank which is connected to the engine. It is usually smaller than the mains and is the one which you pump fuel to from tanks which might not be as well equipped with gauges or glasses. The one on the FE2b is a gravity feed tank. My guess is the other tanks would not feed by gravity and accordingly all fuel going to the engine gets pumped -maybe by hand - into the day tank where it falls by gravity to the engine making the tank a kind of buffer. Perhaps John McKensie will enlighten us on how his system worked.

Another advantage of a system like this is that you can filter the fuel coming from other tanks before it goes to the day tank and so if you have any problems there, you will know and can land using what's left in the day tank. This does require that you pay attention, though.

I like this system because it's simple, almost as simple as what you had on the Luscombe, and I had on the 120 - simple gravity feed.

One of my personal bugaboos is fuel systems that no-one really understands. The real risk of a not-well-understood fuel system is that if it fails to deliver, you may try to troubleshoot it rather than putting the plane on the ground and while so-doing use up the time you might have spent looking for Aunt Millie's tea party on that really big lawn.

john

Correction: I've been told not to call it a day tank since it's on an airplane, not a boat. Try "Feed Tank"

[John McKenzie]

Hi there all , John and Sid , thanks for that opener.....Just come on to have a look here on Joes thread.......Actually its quite a bit out of my comfort zone for the most part so I'm unable to contribute to the discussions in general , but all very interesting none the less .......Ref. Johns bit on the FE2b fuel system . I have , just the other day , put it all down coincidentaly , on my FE2b thread . ..I think you all pretty much have it figured anyway , but , if anyone would be interested , it is explained , as best I can , there. ...I have left out things like "filters" etc..so as not to get bogged down in my own ramblings !..........Keep up the good work..Regards to all..John.