X-Virus-Scanned: clean according to Sophos on Logan.com Return-Path: Sender: To: lml@lancaironline.net Date: Wed, 06 Nov 2013 10:28:13 -0500 Message-ID: X-Original-Return-Path: Received: from mail-pb0-f54.google.com ([209.85.160.54] verified) by logan.com (CommuniGate Pro SMTP 6.0.7) with ESMTPS id 6578650 for lml@lancaironline.net; Wed, 06 Nov 2013 09:01:41 -0500 Received-SPF: pass receiver=logan.com; client-ip=209.85.160.54; envelope-from=legacyl2k@gmail.com Received: by mail-pb0-f54.google.com with SMTP id ro12so3556711pbb.41 for ; Wed, 06 Nov 2013 06:01:05 -0800 (PST) X-Received: by 10.68.225.9 with SMTP id rg9mr3439879pbc.122.1383746465463; Wed, 06 Nov 2013 06:01:05 -0800 (PST) X-Original-Return-Path: Received: from [192.168.0.7] (71-35-70-47.phnx.qwest.net. [71.35.70.47]) by mx.google.com with ESMTPSA id zq10sm48420581pab.6.2013.11.06.06.01.03 for (version=TLSv1 cipher=ECDHE-RSA-RC4-SHA bits=128/128); Wed, 06 Nov 2013 06:01:04 -0800 (PST) Subject: Re: [LML] More thoughts on canopy problems References: From: Mike's Gmail Content-Type: multipart/alternative; boundary=Apple-Mail-ADA5CD87-F8B3-4A18-9A06-2CB400F50DA0 X-Mailer: iPhone Mail (9B206) In-Reply-To: X-Original-Message-Id: <9401FFF5-0F75-4291-9D7C-BA3CC1EA1DDB@gmail.com> X-Original-Date: Wed, 6 Nov 2013 07:00:59 -0700 X-Original-To: Lancair Mailing List Content-Transfer-Encoding: 7bit Mime-Version: 1.0 (1.0) --Apple-Mail-ADA5CD87-F8B3-4A18-9A06-2CB400F50DA0 Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset=us-ascii Fred, Get some more experience in flying vs engineering. Their are so many system= s on many varieties of aircraft that will kill you if you over look your pre= flight responsibilities. You can put a band-aid on this problem but how are= you going to remember to put the band-aid on. Also since your not able or w= illing to test the airplane in a realistic way. How are you going to valida= te that your band-aid is up to the task. BTW, what is the force required to= hold the canopy closed. =20 In my humble opinion, a warning system would have saved everyone of these pi= lots. I have flown 3 4P turbs, 2 4P's, 4 ES's, 13 Legacy's, 7 320/360's, 1 235. D= o you know how many had a canopy/door/ cargo door unlocked warning system. 2= ! And one was not function when I started the test program. Many aircraft will not fly if not properly prepared for flight. Over 3000 a= irline flights are performed a day in the US. All of these airplanes requie= r proper preparation prior to flight. All of which will not fly without fla= ps set for takeoff. The primary causal factor to these accidents are lack o= f pilot discipline, poor or no procedures in place, and poor or inadequate i= ndication/safety systems. You can't engineer discipline. Mike Larkin Sent from my iPhone On Nov 6, 2013, at 5:31 AM, "frederickemoreno@gmail.com" wrote: > > A number of our group have written me privately in response to my prior no= te about the Legacy canopy problem. Together we have given it a lot of thou= ght, so here I will summarize thoughts, ideas, and concerns all in one place= , and ask somebody/some group (LOBO comes to mind) in the USA to take the le= ad in a suggested corrective action program.=20 > =20 > We have lost two Legacy aircraft in Australia in less than two months, one= having a definite oscillating canopy problem, and the other looking suspici= ously the same, but with fewer good witness reports. I think we are up to s= ix fatalities attributable in some way to open canopy difficulties. The cur= rent situation suggests more may be inevitable without change. > =20 > Theory > =20 > Here is an example of what we want: Fly an RV with the canopy unlocked, a= nd it will rise up a few inches at which point the down force from the air b= last counters the up force from the gas springs, and the canopy just floats o= pen a few inches staying in one position. The pilot returns to the airport,= latches the canopy, and continues on his way chastised by his oversight.=20= > =20 > This kind of canopy behavior should be our goal for the Legacy: inherently= fail safe. If we can not get inherently fail safe behavior, then we need f= ail safe add-on mechanisms at work that absolutely prevent the dangerous beh= avior from occurring. > =20 > Because there are reports of canopy oscillations occurring at some flight c= onditions and in some accident reports, it may be that we have a confluence o= f mass, spring constants (in gas springs), aerodynamic forces pushing the ca= nopy down against the gas springs, some periodic forcing function (aerodynam= ic turbulence?) and inadequate damping all conspiring to create a resonant o= scillation of canopy up and down. Once started, it can not be stopped. =20 > =20 > An oscillating canopy would be hugely distracting and its up and down move= ments would no doubt badly disrupt the flow over the vertical stabilizer and= parts of the horizontal stabilizer. Yaw control would be compromised, and t= he changing forces on the horizontal stabilizer due to air flow disruption m= ay also cause the plane to oscillate providing the feedback effect which may= drive the canopy up and down. =20 > =20 > This theory is consistent with anecdotal evidence of the Geraldton crash w= here the canopy was seen to be substantially elevated and the airplane skewe= d sideways rather than making a coordinated turn in an attempt to return to t= he airport.=20 > =20 > If resonant oscillation is the problem arising from a spring mass (canopy)= system with periodic excitation, then engineers in the group will recognize= that if we can not remove the exciting forces (aerodynamic), we can apply l= ots of damping to kill the oscillation and maintain a stable condition. In t= heory, this could be obtained by substituting new gas springs having air/oil= working fluid like a nose strut. The damping orifice and oil viscosity are= selected to create a lot of viscous drag when the canopy is moving thereby c= reating the damping needed.=20 > =20 > If theory is correct, we could go directly to a fail safe solution: a cano= py that will not rise excessively and then go into destabilizing oscillation= s that doom the pilot to a bad outcome. Use of heavily damped gas springs w= ould be a simple bolt-in solution. > =20 > If the oscillation theory can not be proved, then a secondary fail safe me= chanism is needed. > =20 > With this overview, let's evaluate possible solutions. =46rom this vantag= e point I see at least two action pathways to pursue, and quickly before som= ebody else gets killed. > =20 > Solutions > =20 > Some have incorporated safety switches that illuminate warning lights or a= uditory alarms or EFIS warning messages if the canopy is not correctly seate= d. Given that an error in indication (failure to alarm) could lead to fatal= results, one builder added redundancy with dual switches.=20 > =20 > All of this is a vast improvement over having nothing, but in my view is i= nadequate because it is not yet inherently fail safe. If a switch or connec= tion fails, you are toast. Such arrangements are dependent on correct posit= ioning and setting of switches and latches. I am unfamiliar with Legacy lat= ches, but it is conceivable that one could construct a latch system that clo= ses the canopy adequately to trip the switch. Then perhaps with some wear o= ver the years, the latch could release without warning and when air loads ar= e applied allowing the canopy to jump up. For this reason, switches and be= lls and whistles are an improvement, but are NOT good enough. Remember, it i= s a life or death matter. > =20 > If we can establish that highly damped gas struts were up to the job, that= would be inherently fail safe as long as the springs and damping were up to= snuff. Adding a safety switch is then a nice and helpful warning feature,= but not essential. > =20 > If one can not control the oscillation any other way, then a fail safe sec= ondary safety latch is needed like the hood latch on your car. I understan= d there are objections to an ugly, centrally located spring-loaded safety ca= tch, not the least of which is that in the event of a crash for some other r= eason, rescuers may not know how to open both the primary latches (which mus= t be marked) and then the secondary latch. A counter argument is that if th= e canopy only pops up 1.5-2 inches when the primary latch is released, it qu= ickly becomes self evident where the safety latch is located, and a quick in= spection would disclose that just pushing it forward will release the canopy= . I leave it to others to judge this. > =20 > Another proposal is to use an inertial reel and belt arrangement presumabl= y fastened to the center rear of the canopy and bulkhead immediately beneath= , and set to let the canopy come up only slowly, and then lock if the canopy= tries to accelerate upward. One correspondent reported riding with another= pilot who has something similar, but with no inertial reel. He reported it= is impossible to ignore the strap not being locked down when it is blowing a= round in the cockpit during taxiing. However, same concern: what about resc= uers during a crash incident? > =20 > Proposal > =20 > Given the tragic loss of life and severity of the problem, I propose a dua= l track program which I suggest LOBO leads and coordinates given the skill o= f its management and members and location in the middle of the community in t= he USA.=20 > =20 > 1) Oscillation tests. We DO NOT want to test by opening canopy in flight.= Rather, I suggest the following. Two people in the airplane, on the ramp,= brakes locked, one managing brakes and throttle, and the other pushing up o= n the canopy against the air blast that tries to close the canopy pushing ag= ainst the air springs. Adjust to different power settings (say 2000 RPM, t= hen 2100, 2200, etc) and continue with periodic pushes up adjusting the freq= uency of pushing to see if one can find and excite an unstable oscillating r= egime. Yes this is way short of a flight test, and yes, it will not induce v= ariable downloads on the tail which would wiggle the airframe up and down wo= rsening the problem. But it would be safe. Careful observation from outside= observers (hopefully with video cameras) could watch canopy behavior as wel= l as search for variable down load on the tail if the canopy starts to oscil= late up and down. Imagine putting a post immediately behind the trailing e= dge of the elevator with some marks to measure displacement, and see what ha= ppens. > =20 > If these tests uncovered a regime of canopy oscillation, we would not have= a smoking gun, but a pretty good outline of what the gun looks like. With s= ome measurement of spring constants and canopy mass and moment of intertia o= ne can construct a simple spring mass model and check to see if it predicts t= he frequency of oscillation. With that model, one can then calculate the cr= itical damping required to kill the oscillation, and go on to order test spr= ings with damping coefficients that have MORE than the critical level. Rete= st, and then perhaps we may have a fail safe solution. Full testing that in= cluded flight would be risky, but may not be required if we can get some goo= d test and vibration engineers working the problem that end up satisfied wit= h the model and results. If it works, we have a bolt-in solution that is f= ail safe. > =20 > 2) In parallel, assume that the oscillation theory can NOT be ground verif= ied. Flight testing is too risky. So the second pathway is plan B: a fail s= afe latching or restraint system. So far we have spring loaded hood latch d= esigns to consider and inertia belt constraints tuned up for the application= . Other ideas may arise.=20 > A) Develop and ventilate as many ideas as can be found. > B) Evaluate among the options and select the best approach measured ag= ainst some objective criteria including effectiveness, complication, cost, d= ifficulty of installation, etc. > C) Design, build, test, evaluate a prototype, prepare a brief report a= nd distribute to the community. Let lots of owners have a review for evalua= tion and gather additional suggestions. Revise based on suggestions. > D) Collect deposits, manufacture parts to retrofit the fleet, and then= chase up all Legacy owners for retrofit of safety hardware (latches or impr= oved high damping gas springs, assuming they can be shown to work). > =20 > Hopefully concerned operators will take up the challenge and make progress= on both pathways. Owners should be VERY concerned given that there is a lo= aded shot gun pointed at your head every time you take off. Some may still w= ant to practice denial and brush it off. They will likely be weeded out of t= he gene pool. =20 > =20 > I suggest a smart, thoughtful, and well informed course of action, and I p= ut the challenge to LOBO to push it forward, and if not, then a group of vol= unteers. > =20 > And with that I leave it to you. I have done all I can to get things unde= rway.=20 > =20 > I do not want to read more accident reports. > =20 > =20 > Fred Moreno in far off Western Australia > =20 > =20 > =20 > --Apple-Mail-ADA5CD87-F8B3-4A18-9A06-2CB400F50DA0 Content-Transfer-Encoding: quoted-printable Content-Type: text/html; charset=utf-8
Fred,

=
Get some more experience in flying vs engineering.  Their are so m= any systems on many varieties of aircraft that will kill you if you over loo= k your preflight responsibilities.  You can put a band-aid on this prob= lem but how are you going to remember to put the band-aid on.  Also sin= ce your not able or willing to test the airplane in a realistic way.  H= ow are you going to validate that your band-aid is up to the task.  BTW= , what is the force required to hold the canopy closed.  
In my humble opinion, a warning system would have saved everyone= of these pilots.

I have flown 3 4P turbs, 2 4P's, 4= ES's, 13 Legacy's, 7 320/360's, 1 235.  Do you know how many had a can= opy/door/ cargo door unlocked warning system.  2!  And one was not= function when I started the test program.

Many air= craft will not fly if not properly prepared for flight.  Over 3000 airl= ine flights are performed a day in the US.  All of these airplanes requ= ier proper preparation prior to flight.  All of which will not fly with= out flaps set for takeoff.  The primary causal factor to these accident= s are lack of pilot discipline, poor or no procedures in place, and poor or i= nadequate indication/safety systems.

You can't engi= neer discipline.

Mike Larkin

Sent from my iP= hone

On Nov 6, 2013, at 5:31 AM, "frederickemoreno@gmail.com" <frederickemoreno@gmail.com> wrote:

<SENDER_EMAILfrederickemoreno@gmail@@com.png>
A number of our group have written me privately in response to my prior= note about the Legacy canopy problem.  Together we have given it a lot= of thought, so here I will summarize thoughts, ideas, and concerns all= in one place, and ask somebody/some group (LOBO comes to mind) in the U= SA to take the lead in a suggested corrective action program. 
 
We have lost two Legacy aircraft in Australia in less than two months, o= ne having a definite oscillating canopy problem, and the other looking s= uspiciously the same, but with fewer good witness reports.  I think we a= re up to six fatalities attributable in some way to open canopy difficulties= .  The current situation suggests more may be inevitable without change= .
 
Theory
 
Here is an example of what we want:  Fly an RV with the canopy unl= ocked, and it will rise up a few inches at which point the down force from t= he air blast counters the up force from the gas springs, and the canopy just= floats open a few inches staying in one position.  The pilot returns t= o the airport, latches the canopy, and continues on his way chastised b= y his oversight.
 
This kind of canopy behavior should be our goal for the Legacy: inheren= tly fail safe.  If we can not get inherently fail safe behavior,= then we need fail safe add-on mechanisms at work that absolutely prevent th= e dangerous behavior from occurring.
 
Because there are reports of canopy oscillations occurring at some flig= ht conditions and in some accident reports, it may be that we have a conflue= nce of mass, spring constants (in gas springs), aerodynamic forces pushing t= he canopy down against the gas springs, some periodic forcing function (aero= dynamic turbulence?) and inadequate damping all conspiring to create a r= esonant oscillation of canopy up and down.  Once started, it can not be= stopped. 
 
An oscillating canopy would be hugely distracting and its up and down m= ovements would no doubt badly disrupt the flow over the vertical stabil= izer and parts of the horizontal stabilizer.  Yaw control would be= compromised, and the changing forces on the horizontal stabilizer due to ai= r flow disruption may also cause the plane to oscillate providing the feedba= ck effect which may drive the canopy up and down.   
 
This theory is consistent with anecdotal evidence of the Geraldton cras= h where the canopy was seen to be substantially elevated and the a= irplane skewed sideways rather than making a coordinated turn in an attempt t= o return to the airport. 
 
If resonant oscillation is the problem arising from a spring mass (cano= py) system with periodic excitation, then engineers in the group will recogn= ize that if we can not remove the exciting forces (aerodynamic), we can appl= y lots of damping to kill the oscillation and maintain a stable condition.&n= bsp; In theory, this could be obtained by substituting new gas springs havin= g air/oil working fluid like a nose strut.  The damping orifice and oil= viscosity are selected to create a lot of viscous drag when the canopy is m= oving thereby creating the damping needed. 
 
If theory is correct, we could go directly to a fail safe solution: a c= anopy that will not rise excessively and then go into destabilizing oscillat= ions that doom the pilot to a bad outcome.  Use of heavily damped gas s= prings would be a simple bolt-in solution.
 
If the oscillation theory can not be proved, then a secondary fail safe= mechanism is needed.
 
With this overview, let's evaluate possible solutions.  =46rom thi= s vantage point I see at least two action pathways to pursue, and quickly be= fore somebody else gets killed.
 
Solutions
 
Some have incorporated safety switches that illuminate warning lig= hts or auditory alarms or EFIS warning messages if the canopy is not correct= ly seated.  Given that an error in indication (failure to alarm) could l= ead to fatal results, one builder added redundancy with dual switches. =
 
All of this is a vast improvement over having nothing, but in my view i= s inadequate because it is not yet inherently fail safe.  If a switch o= r connection fails, you are toast.  Such arrangements are dependent on c= orrect positioning and setting of switches and latches.  I am unfamilia= r with Legacy latches, but it is conceivable that one could construct a latc= h system that closes the canopy adequately to trip the switch.   T= hen perhaps with some wear over the years, the latch could release with= out warning and when air loads are applied allowing the  canopy to jump= up.  For this reason, switches and bells and whistles are an improveme= nt, but are NOT good enough.  Remember, it is a life or death matter. <= /div>
 
If we can establish that highly damped gas struts were up to the job, t= hat would be inherently fail safe as long as the springs and damping were up= to snuff.  Adding a safety switch is then a nice and helpful  war= ning feature, but not essential.
 
If one can not control the oscillation any other way, then a fail safe s= econdary safety latch is needed like the hood latch on  your car. = I understand there are objections to an ugly, centrally located spring-load= ed safety catch, not the least of which is that in the event of a crash for s= ome other reason, rescuers may not know how to open both the primary la= tches (which must be marked) and then the secondary latch.  A counter a= rgument is that if the canopy only pops up 1.5-2 inches when the primary lat= ch is released, it quickly becomes self evident where the safety latch is lo= cated, and a quick inspection would disclose that just pushing it forward wi= ll release the canopy.  I leave it to others to judge this.
 
Another proposal is to use an inertial reel and belt arrangement presum= ably fastened to the center rear of the canopy and bulkhead immediately bene= ath, and set to let the canopy come up only slowly, and then lock if th= e canopy tries to accelerate upward.  One correspondent reported r= iding with another pilot who has something similar, but with no inertial ree= l.  He reported it is impossible to ignore the strap not being locked d= own when it is blowing around in the cockpit during taxiing.  However, s= ame concern: what about rescuers during a crash incident?
 
Proposal
 
Given the tragic loss of life and severity of the problem, I propose a d= ual track program which I suggest LOBO leads and coordinates given the skill= of its management and members and location in the middle of the community i= n the USA. 
 
1) Oscillation tests.  We DO NOT want to test by opening canopy in= flight.  Rather, I suggest the following.  Two people in the airp= lane, on the ramp, brakes locked, one managing brakes and throttle, and the o= ther pushing up on the canopy against the air blast that tries to close the c= anopy pushing against the air springs.  Adjust  to different power= settings (say 2000 RPM, then 2100, 2200, etc) and continue with period= ic pushes up adjusting the frequency of pushing to see if one can find and e= xcite an unstable oscillating regime.  Yes this is way short of a fligh= t test, and yes, it will not induce variable downloads on the tail which wou= ld wiggle the airframe up and down worsening the problem. But it would be sa= fe.  Careful observation from outside observers (hopefully with video c= ameras) could watch canopy behavior as well as search for variable down load= on the tail if the canopy starts to oscillate up and down.   Imag= ine putting a post immediately behind the trailing edge of the elevator with= some marks to measure displacement, and see what happens.
 
If these tests uncovered a regime of canopy oscillation, we would not h= ave a smoking gun, but a pretty good outline of what the gun looks like.&nbs= p; With some measurement of spring constants and canopy mass and moment of i= ntertia one can construct a simple spring mass model and check to see if it p= redicts the frequency of oscillation.  With that model, one can then ca= lculate the critical damping required to kill the oscillation, and go on to o= rder test springs with damping coefficients that have MORE than the critical= level.  Retest, and then perhaps we may have a fail safe solution.&nbs= p; Full testing that included flight would be risky, but may not be required= if we can get some good test and vibration engineers working the problem th= at end up satisfied with the model and results.   If it works, we h= ave a bolt-in solution that is fail safe.
 
2) In parallel, assume that the oscillation theory can NOT be grou= nd verified.  Flight testing is too risky.  So the second pathway i= s plan B: a fail safe latching or restraint system.  So far we have spr= ing loaded hood latch designs to consider and inertia belt constraints tuned= up for the application.  Other ideas may arise. 
    A) Develop and ventilate as many ideas as can be fou= nd.
    B) Evaluate among the options and select the best ap= proach measured against some objective criteria including effectiveness, com= plication, cost, difficulty of installation, etc.
    C) Design, build, test, evaluate a prototype, prepar= e a brief report and distribute to the community.  Let lots of own= ers have a review for evaluation and gather additional suggestions. Rev= ise based on suggestions.
    D) Collect deposits, manufacture parts to retrofit t= he fleet, and then chase up all Legacy owners for retrofit of safety hardwar= e (latches or improved high damping gas springs, assuming they can be shown t= o work).
 
Hopefully concerned operators will take up the challenge and make progr= ess on both pathways.  Owners should be VERY concerned given that there= is a loaded shot gun pointed at your head every time you take off.  So= me may still want to practice denial and brush it off.  They will likel= y be weeded out of the gene pool.   
 
I suggest a smart, thoughtful, and well informed course of ac= tion, and I put the challenge to LOBO to push it forward, and if not, then a= group of volunteers.
 
And with that I leave it to you.  I have done all I can to get thi= ngs underway. 
 
I do not want to read more accident reports.
 
 
Fred Moreno in far off Western Australia
 
 
 
 <image.gif>
= --Apple-Mail-ADA5CD87-F8B3-4A18-9A06-2CB400F50DA0--