To increase the flutter speed for the big tail, I added three strips of 18 inches wide, style 7781 fiberglass cloth to the top and cottom of the fuselage and reran the flutter analysis. No flutter is observed for speed up to 600 kts for all altitudes. The fiberglass should be oriented at 0 (zero) degrees to the longitudinal axis of the fuselage.

I should mention that even though many of the Lancairs with the big tail have flown thousands of hours with no flutter, it does not mean that there is no flutter problem. Records show that some airplanes have flown 10 years with no flutter problem and suddenly it happens. And it only needs to happen once. Steve Wittman's fatal accident is a good example of that. He had flown his Tailwind over 10 years before he encountered a tail flutter problem. He and his wife were killed in their Tailwind and the tail of his aircraft has never been found.

So, regarding flutter, I’ll provide a brief history to describe this non-event. As we all know, the 320 has never ever exhibited any indications of flutter whatsoever. But ok, this is not proof that flutter can’t occur. Not being an expert on flutter myself, I have spoken to many "experts" in the field. One trait rings true, "ask ten experts (although you’ll be pressed to find ten flutter experts on this planet) and you’ll surely get ten different answers" - and each will say the other is of course wrong.

Last year, we contracted with one of the best respected flutter experts in all of Europe (Norbert Niedbal of Vibration Engineering Inc.) who has specialized in composite airframe flutter analysis for all the major makers in Europe. Europe, by the way, leads the world in composte airframe manufacturing.

We outfitted a 360 MKII (large tail) with 38 accelerometers to measure normal mode deformations. Phase resonance methods were used for the survey. Mr. Niedbal states that although this method may be slow, "you see the mode shapes and know better what you do". Thirty six different normal modes were measured. Flutter calculations were made according to the P-K Method. For testing to 1.2 Vd, the frequency range was enlarged by a factor of 1.2 to 58 Hz. Actually some vibration tests were performed for a frequency range up to 65 Hz.

And the 132 page survey says: None of the control surfaces showed any instability. There was no fuselage tail section coupling found either. In fact everything came up excellent with one finding of possible interest: The flap rotation / torsion mode coupled with the wing bending and the wing torsion modes. Short story is that the flap, being essentially unsupported at the outbd end showed some potential for coupling with the wing. This case, according to the engineer, "has a very shallow curve and no aggressive slope". With no service history to suggest any problem any service bulletin would be of the precautionary nature, however we are about to issue one since a very simple 4lb preload on the outbd flap (in the up position) fully cured the case on the computer modeling. With this preload, the overall aircraft flutter free speed exceeded 300 KTAS.

We could undoubtedly contract with another flutter engineer and perhaps receive yet another set of data and flutter free speeds. Why all the differences? It seems there are lots of possible reasons such as, proper methods of excitation and monitoring, number of accelerometers and interpretation and accurate recognition of exactly what your measuring. Mr. Neidbal saw some "strong contradictions" in the one existing report that we sent him. With those contradictions, one would have had to "fix" the FE model which would require a few "tricks". Tricks such as varying shear modulus or densities for example but such "fixes" can cause added problems with subsequent interpretation.

From: Martin Hollmann

Subject: Flutter Analysis on the Lancair 360 with the Big Tail

Date: 3 November 1997

Many years ago, maybe 50 years or so, before the days of finite element analysis, a ground vibration test was performed and the data read from accelerometers was used to determine the mode shapes and frequencies of a structure. These mode shapes and frequencies were then used to perform a flutter analysis. The flutter analysis was performed using charts and calculations and later using computer codes.

This method worked if all mode shapes could be recognized and the mode shapes that coupled were compared. Often mode shapes were not identified and often improper results were obtained.

Today we perform a finite element analysis and compare the results to a ground vibration test. Using both methods we compare results and make certain that a mode has not been overlooked. Often one method will find a mode that was not found by the other method.

That is exactly what happened with the BD-10 jet on which I performed the "post crash" flutter analysis. The initial analysis was performed by an ex-Douglas Aircraft engineer who [was] only comparing 2nd bending and twisting of control surfaces. He never checked vibration mode coupling with other structures such as the fuselage or coupling of other surfaces.

When I performed the analysis I set up a very detailed FEA model of the tail assembly and input the data into my flutter program (SAF). The results showed that the fuselage bending up and down was coupling with the horizontal tail bending. A critical flutter speed of 385 kts was determined and that is exactly at what speed the tail came apart on the BD-10. The CEO of Peregrine International was killed in that accident. After the crash, we also made a ground vibration test to determine the actual control surface responses. During those tests we did not find the fuselage bending frequency. however, from the FEA and flutter analysis, we knew that fuselage bending was the culprit. I knew that earlier ground vibration tests had been performed. I looked through the data and "Bingo", there it was, "the fuselage bending frequency" that the FEA model had shown. Had I just used the control surface responses from the last ground vibration test I would not have found the correct answer. The finite element model data had shown us what to look for. I have also found cases where the vibration test came up with modes (assymetric modes) that the FEA model did not. It is therefore important to recognize that finite element modeling and vibration testing go hand-in-hand. There is no flutter analyst that will refute this. No one except Mr. Niedbal. Mr. Niedbal's analysis of your Lancair 360 tail suggests that only a vibration test was performed.

The flutter results of your aircraft are very similar to those of the BD-10. My analysis of your aircraft shows that the aft fuselage bending couples with fuselage torsion. It takes a large shaker to excite the aft fuselage for a ground vibration test. Did Mr. Niedbal find the fuselage bending and torsional frequencies? Did he include these in his flutter analysis? If you send me a copy of his report I will be happy to look it over and give you my opinion of what has been done.

One comment that you make really bothers me. You state that Mr. Niedbal quotes: (saw "strong contradictions" in the existing report(I assume that this is my report) With those contradictions, one would have had to "fix" the FE model which would require a few "trickes".) This statement clearly indicates that Mr. Niedbal is not familiar with finite element modeling. Certainly he is not using it as we do in the USA today.

I might add that I worked for Messerschmidt Bölkow Blohm and other companies in Germany for several years, and that Germany is far behind the USA in the development and use of computer programs. I have also worked in most major US aerospace companies where the real progress and development of composite materials is being made. If it were not for those developments, your aircraft would not have been made out of those materials. The honeycomb, the glue sheets, woven graphite prepregs and vacuum bagging methods were developed by such companies as DuPont, Hercules, Fiberite, Hexcel, Convair, Lockheed, Northrop, and many others. I know because I was there when it happened. But that, again, is another story.

Sincerely yours,

In case you're wondering, I will be incorporating Martin's suggested modification into my aircraft. Details of the process will be included here on the site for all to see, of course.