WWI Replica Safety

[Joe Perkel]

Originally Posted by hank jarrett

I have seen it tear major components off an airframe is as little as two cycles! It looks like a simple overload failure until you look at the high speed films and see that there actually WAS flutter.

The more commonly seen flutter (because the guys that experienced it actually had a chance to survive) is lower amplitude and doesn't diverge as fast.

If you EVER feel anything that feels like flutter the best thing to do (if you have time) is slow down and get on the ground.

Hank

The discussion and short films regarding flutter are fascinating, informative, and unnerving all at the same time.

Questions then in my mind would be,...

- If an existing design has flown a number of examples, (say at least several hundred), and there has been no indication of such tendencies, then adherence to the original design and prudent rigging practices should likely be enough to stay out of this kind of trouble??

-A prudent initial test flight profile, would likely not be the usual "once around the patch" type of flights seen so often by homebuilders, in favor of a full power climb to altitude with egress capability and a full preplanned test syllabus?

Clearly much to consider.

[j ferguson]

I, too, found the flutter discussion unnerving. Having a plane shed mechanical parts in-flight is one thing, (I'm used to that) but airframe components is entirely different.

As Joe suggests, accurately reproducing a known design seems the safest way to go, assuming we've heard of that design's concerns.

Clearly, geometry and rigging must be accurate. But how accurate? Is it possible that the tolerances within which a design is flutter resistant are very small?

Sid's observation about elevator stop settings on the Luscombe suggest that a change in the stops which produces 1/4 inch greater deflection at the trailing edge can cause more interesting stalls. This should provide a clue.

So, flutter experts. What would you look for in flight testing these planes?

[Blue Max Aviation DR1]

Originally Posted by j ferguson

I, too, found the flutter discussion unnerving. Having a plane shed mechanical parts in-flight is one thing, (I'm used to that) but airframe components is entirely different.

As Joe suggests, accurately reproducing a known design seems the safest way to go, assuming we've heard of that design's concerns.

Clearly, geometry and rigging must be accurate. But how accurate? Is it possible that the tolerances within which a design is flutter resistant are very small.

Sid's observation about elevator stop settings on the Luscombe suggest that a change in the stops which produces 1/4 inch greater deflection at the trailing edge can cause more interesting stalls. This should provide a clue.

So, flutter experts. What would you look for in flight testing these planes?

Perhaps Billy Bishop's mother was right when she urged her son to "keep it low, slow and level in the turns..."

Lynn

[hank jarrett]

Joe,
Actually it is GOOD that the discussion and short films regarding flutter are a little (or a lot) unnerving. I have also spent 20 years doing crash investigations and have buried a few friends in that time.
There is a lot to be said for flying a proven design with lots of examples out there. Some designs fly for years on the ragged edge of a flutter problem and no one ever knows till someone makes the tiniest change, like adding a stall strip to the leading edge of a wing or changing the wing strut streamlined tube for one with a different cross section. With many examples flying there is a better chance that someone else has already tried what you are thinking of, and hopefully it didn't cause a problem. One typical problem (the one most dynamic stability classes use as an example) is changing the counterweight on a control surface. This isn't usually something someone does on purpose (if they are smart). What kills pilots without warning is when it happens and they don't even know it has changed. Ever look at the control surfaces on many planes where the skin is made of a flat piece of aluminum with stiffeners pressed into it? If the plane has enough counter balance to be stable and convergent as designed, what do you think would happen if mud daubers built nests in the trailing edge? The CG of the surface would move back, potentially past the static margin limit, making the surface unstable and prone to flutter. I almost lost a friend in a Cessna because of ice in the elevator. He was lucky and slowed down and got the plane back on the runway, but there was control system damage from all the control shaking. One of the really scary things I have seen is someone adding a Grimes rudder light a the back edge of the rudder. This is REALLY asking for trouble! On many high speed planes with fabric tail surfaces you are required to balance the surface to VERY tight tolerances when you recover or paint the fabric. We had to tear all of the fabric off of a C-118 rudder that looked perfect because the CG was 1/4" to far aft. Just a "little" to much dope! The bad thing was the plane was on its way to Davis Monthan for scrap.
Another is messing with the control gaps. Putting tape over the control gap can increase sensitivity and smoothness of the control, but if the tape comes part way off, it can also cause turbulent flow across the surface increasing the chances of flutter.
As to not doing a "once around the patch" and going for a full power climb to altitude, I wouldn't do that. The full power climb results in higher speed and there is a pretty good potential for inducing flutter at higher power settings. Move out in the flight envelope slowly! Raise the speeds a little at a time and explore the corners of the VN diagram in small well planned steps. It IS a good idea to climb to altitude (in case an emergency exit is needed) but there is NO HURRY to get there!
This is why I think a good flight test plan is important. You are absolutely right. There is MUCH to consider. The place to start may be a clear understanding of the VN diagram. I was a little shocked to find out how many pilots didn't even know what it is! If you don't know either, you are actually in very good company.
Hank

[Joe Perkel]

Originally Posted by hank jarrett

The place to start may be a clear understanding of the VN diagram. I was a little shocked to find out how many pilots didn't even know what it is! If you don't know either, you are actually in very good company.
Hank

Here's one for an RV9.

I do not recall having encountered one before, as a pilot nor A&P (then again I was more "P" than "A"), and it seems to me to be well within the "designers" realm. However, I would say that a builder, is a liaison between designer and pilot. Here we wear two hats,..yes? Quite obviously, it would be more than useful to have a working knowledge of this.

Thank you for bringing this to light Hank, I will study it further.

As for the comment regarding full power climb, I had read (somewhere), that most test flight engine failures happen with the first power reduction.

More research on the matter is required, but it occurred to me that since Vy is the difference between engine power and the power required to overcome the aircraft's drag, then you would likely benefit from both the lowered speed and performance climb.

[hank jarrett]

Joe,
Your timing is perfect! I was looking for a nice VN diagram last night to use as an example. This one has all the complexity to show what is needed for a WW-I design and a few things that aren't (so we can explain why).
As a starting point, every design has a different VN diagram! DON'T try to use this one for your Eindeker!
Hank

OK, I really didn't think anyone would think they could take an Eindeker to over 200 mph, but you never know!

[Blue Max Aviation DR1]

Originally Posted by hank jarrett

Joe,
Your timing is perfect! I was looking for a nice VN diagram last night to use as an example. This one has all the complexity to show what is needed for a WW-I design and a few things that aren't (so we can explain why).
As a starting point, every design has a different VN diagram! DON'T try to use this one for your Eindeker!
Hank

OK, I really didn't think anyone would think they could take an Eindeker to over 200 mph, but you never know!

If you have a flutter problem the component parts might come close to that.

[Joe Perkel]

Hank,

Taking a moment to analyze the VN diagram, certain things are clear however, would you please take a moment to clarify a couple of items?

Clearly understood

4 g's @ 150 mph produces structural damage, 6 g's structural failure. Conversely, -2 g's @ 150 mph = structural damage, and -3 g's = failure.

Questions then

-What is the +-50 feet per second gust line, and how does that figure into this?

-How do you calculate and plot these failure modes? There is a lot of hidden work in this diagram, IE when does one simply deform any single component vs break another at X speed and G's?

[hank jarrett]

First, even though the V-n diagram looks like there will be damage at just under 4 g’s and 150 mph, what the diagram is REALLY telling you is the design was TESTED to that g load from ~115 mph to ~180 mph. I would read that to mean I am safe to pull 3.5 g’s at any of those speeds. Now for a caveat, (aren’t there always exceptions?) That assumes the plane has NO damage or degradation in any affected load path (now you know why those annual inspections are so important and why the A&Ps jump all over such minor damage). It also assumes that there is perfectly smooth air and the g’s are applied gently.

That brings us to your gust load question. Those lines are calculated using the equations from the FAR’s. We will pass on the math here because at this point you just need to understand the results and I don’t want this to turn into a math class.

From the Feds on V-n diagrams;
Gust load is computed from the FAR Part 25 equation (the one we won’t be looking at). The formula considers a vertical gust of specified speed and computes the resulting change in lift from the wing. The associated incremental load factor is then multiplied by a load alleviation factor that accounts primarily for the aircraft dynamics in a gust.

Nice and clear? Don’t worry about it. It confuses the engineers at first too.

What it is saying is that a vertical gust changes the effective angle of attack of the wing and has the same effect as pulling back on the stick and loading the plane with g’s. As you can see from the V-n diagram, this effectively reduces the number of G’s you can pull in gusty weather. Sound familiar? That was what you learned when you were under instruction. If it’s bumpy, slow down! Now you know why, and more importantly, HOW MUCH to slow down. Since it is kind of hard to get Mother Nature to oblige the test pilot with just the right gust for tests, this curve is calculated. Also remember, if the gust is HIGHER than 50fps the curve would be MORE restrictive. To know the total effect you would have to run new calculations for EVERY possible gust. A 50fps gust is a good working size and no one could remember (or carry copies of) all the gusts so 50fps is a good standard. Actually for certified planes below 20,000’ the calculations have to be done for 25, 50 and 66 fps.

What is being plotted on the diagram isn’t really failure modes. It is the maximum stress the design has been tested to. It is a lot more complex to predict which component will fail first. If engineers were perfect (and MAN ARE THEY IMPERFECT!) the V-n diagram would reflect exactly the maximum loads the plane could take, and when you put one gram too much stress on the plane, every rivet, bolt and piece of metal in the plane would fail simultaneously. The ultimate “One horse shay”. No wasted material anywhere and the absolute lightest design possible. V-n diagrams DO NOT show where the plane will fail. They show you where someone else has shown they WON’T fail

Hank

Anyone else who understands this stuff PLEASE look at it carefully. I can't exactly ask the guys at work to peer review what I am writing and remember my comment above, ENGINEERS ARE NOT PERFECT! If I make a mistake PLEASE let me know! This stuff can get people killed if it is misused or incorrect!

[Machinbird]

Sorry if I made you nervous with the flutter flicks. As Hank indicated, flutter is serious business. I thought I'd comment about the V-n diagram presented.

The right hand vertical line is there primarily for flutter reasons. Fly into the never exceed speed range and you become a test pilot because, supposedly, no one has investigated the aircraft performance in that speed range or possibly, the flutter experts have forecast problems there and they flew as close to that point as they dared in demonstrating the aircraft's performance. If you were to fly past that line in still air, you might get away with it, then again, you might not. The actual line where bad things start to happen might not be vertical since the flutter behavior of the structure might change as positive and negative g are applied. It really isn't hard to stay below Vne as long as you keep the aircraft under good control. Just remember that there are degradations of the airframe that can cause flutter conditions to happen on the "safe" side of the Vne line (like the out of balance control surface that Hank discussed).

The top and bottom of the V-n diagram are structural limits. Pull too many g's and something has to break. Historically engineers have determined ultimate strength of the airframe (where something actually breaks) and then they set the safe operating envelope at 2/3 of that limit. You can pull more g's than the safe operating limit, it is just that when you do, you start to use up the structural lifetime of the airframe in an accelerated manner. In fighters, it is common to have over g incidents, but you try to avoid them. Maintenance personnel and engineers monitor these incidents and may flag an aircraft for additional inspections based on its "g" history. In your own personal aircraft though, you won't get this type of support, so it pays to be conservative in your flying.

The intersection of the 50 fps gust line with the safe operating g limit defines the structural cruising speed. This is just really an engineering SWAG.
(Scientific Wild *ss Guess). If you should actually fly into a 60 fps gust at the structural crusing speed, you will be beyond the safe operating g limit and causing structural damage (fatigue) at an accelerated rate, but yes, you will probably safely complete your flight. It is just that 50 fps gusts are considered far more prevalent than 60 fps gusts for normal air mass turbulence. Fly into a mountain wave rotor though and all bets are off. The SWAG wasn't intended to fit that situation.

The intersection of the accelerated stall line and the safe operating g limit normally defines the maneuvering speed. In this chart, the maneuvering speed is defined slightly higher, ie up to a slightly higher g load than the safe operating limit. The maneuvering speed is considered the fastest speed that you can bring the stick back firmly and the aircraft will stall at a g level that will not cause airframe damage. One caution though, rolling maneuvers that result in a stall at maneuvering speed will apply excessive g because the load is not evenly distributed.

What do you use the maneuvering speed for? Simply put, maneuvering speed defines the top of the airspeed range where you can safely "horse" the aircraft around. If you pull too hard, the wing will stall before g becomes excessive. In effect, the stall becomes a 'safety valve' to relieve structural stress. Just remember that the V-N diagram is normally not symmetric and what you can do with positive g is different from the negative g limits.

Should you fly into a mountain wave rotor, you will probably experience gusts from all directions that are beyond the 50 fps figure. It is a sensation akin to being thrown into a cement mixer. In that case, a speed significantly below your maneuvering speed will be needed to stay intact. Not an experience I would want to repeat.

As you continue to decelerate along the V direction on the V-n chart you encounter your level flight stall speed. You can fly slower, but it won't be in steady state flight, more like a parabolic arc.

The intent of the above tour of the V-n diagram is to give a better understanding of the use of the diagram from a piloting viewpoint. I hope it is helpful.
Sid