T-34 Mentor Matters

 

 

November 2, 2005

To the members of the T-34 Association Board

Dear Sirs:

I am an Aerospace Engineer (ret.) with 36 yrs. experience in structural analysis, and I have an F33A Bonanza. I recently read George Braly's article in the September Mentor Monitor in which he described his "Airframe Life Extension Cable". I believe your organization should review the following information regarding the claims for this proposed modification. To understand my interest in this matter please see my web site http://mysite.verizon.net/dick.wilson . (I am adding this letter to the site.)

1. ON REDUNDANCY

At 6 Gs the 2 front spar wing bolts apply an upward force to the T-34 fuselage of about 5400 lbs per side, most of this vertical load is a shear force in the lower bolt, reacted in the huck bolts in the carry-thru lower spar cap and from there by the fuselage skin. The spar cap is attached to the web of the I-beam with 60 hucks and rivets from one side to the other. Yet George says it is a pure tension member, holding the wings together, "that's all it does." He goes on to say that the ALEC will provide "total redundancy" and would "fly the airplane at 4 Gs with the bolt cut into 2 parts". He left out the major task of lifting 4 times the weight of the fuselage. The cable would have essentially no capability of providing a "fully redundant load path" in the event of a bolt or spar cap failure. If either of these occurred, say at 6 Gs, the structure would probably crumple due to the vertical wing lift load, and also the ALEC would instantly "see" a tension load of over 33,000 lbs. and it's rated for only 22,000.

2. BOLT STRENGTH

The original NAS certified bolt, as with all tension bolts, is critical in strength at the root of the threads, that reduce the crossectional area and also create stress concentrations. To compensate for these effects it has rolled threads, an appropriate heat treatment to balance strength against brittleness, a preload torque to minimize magnitude of load swings, and on top of all a large margin of safety. The high-strength NAS nut is also a highly sensitive component. With the ALEC, according to the article, the original bolt is replaced by a new bolt with a hole thru it and the original nut is machined to provide a ledge for the "transfer collar". A rough measure of the adverse effect of the threads on fatigue might be the penetration of the threads as a percent of thickness. In the original bolt this is 5% and in the tube-bolt 11%. Another fatigue problem is heat treatment. Standard high strength aircraft bolts are generally limited to a maximum of 160 ksi. because of lower plastic yield (brittleness) between yield and failure, of higher values. The ALEC tube-bolt, at 228 ksi, is said to be stronger than the original bolt. It is difficult to imagine that these redesigns could be approved as replacements for the original standard bolt and nut.

3. LOAD SHARING

The claims for the ALEC are based on the concept of reducing the 33,000 lb. lower spar cap tension at 6 Gs by about 1/3 by means of a steel cable between the left and right wing bolts. In the following I will work backward from the data presented in George's article, and derive critical numbers that are left out such as the necessary cable preload.

In the article there is a table of load factor v.s. total tension (cap + cable) for G meter readings from 1 to 12.26 in irregular steps. A column called "Reduced Loading as result ALEC" is explained in the text using the 5th row as an example: I quote "If you are flying your T-34 at 2.93 "g" then your lower center spar structure is essentially only loaded to 1 g (0.95 g)". From that clarification I was able to calculate the missing cap and cable loads with the formulas D = (C/A) x B and E = B - D. The decoded table is presented below. Columns A, B, and C are taken from the table in the article. (Disregard the blue first row for a moment):

A

Actual Gs		 B

Total Tension		 C

"Reduced Loading"

D

Cap tens. #

 E

Cable tens. #
0 0   -8684 8681
1.0 5333 -0.75 -4000 9333
1.85 9891 0.00 0 9891
2.57 13720 0.63 3363 10357
2.93 15634 0.95 5069 10565
3.29 17549 1.26 6721 10828
6.16 32864 3.78 20167 12697
6.88 36692 4.41 23519 13173
7.24 38607 4.73 25223 13384
9.03 48179 6.30 33613 14566

About a year ago Braly asked me how to calculate the total tension and I e-mailed the equations to him. The numbers in column B are in agreement with what I sent, and are a uniform progression of lbs. per G . The numbers in my added D and E columns based on column C also progress in (approximately) equal amounts per G. The load in the cap rises (33613+ 4000) = 37,613 lbs. in 8.03 Gs or 4684 lbs per G and the load in the cable rises (14566- 9333) = 5233 in the same 8.03 Gs or 652 lbs. per G. To put it another way, the spar Corp used a ratio of stiffness between the cap and the spar of 5233/652 or 8 to 1.  The results, extrapolating from the 1 G numbers back to 0 Gs yield the following for the missing 0 G flight condition.:

0 G: Total load = 0 Cap load = - 4000 - 4684 = - 8684 Cable load = 9333 - 652 = 8681

I have added this 0 G first row in blue to the table. So from the Spar Corp formula the cable load at 0 Gs must have 8681 lbs tension to put the cap into 8681 lbs. compression. To review the above; the Spar Corp. assumed elasticity for both cap and cable, specified cap fraction of total load in column C , and from these I derived the necessary tension in the cable at 0 Gs flight of 8681 lbs. But the ALEC will be installed and torqued with the T-34 on jacks, so the next question is: how much tension must be applied by torquing?

That would be already answered, 8681 lbs., if the torque was applied during 0 G flight. But this would be reduced somewhat by installing the ALEC in the hangar. This would improve working conditions as well. Deriving the cable preload required to make all of the above come about:

The compression at the lower bolt with the airplane on jacks is the wing weight moment divided by the distance between the upper and lower bolts (11.06 in.). This is 56120/11.06 = 5074 lbs. compression for the T-34. (These data are from my web site, May 20 addendum, generated for another purpose). On the G scale that would be: -5074/5333 = - 0.95 G. That results in the 3 rd row in the table below. Therefore the cable tension to be applied by torquing the new 7/16 nut during assembly is 8060 lbs.

The following table summarizes the details derived under "Load Sharing" above. I added the first and second rows for later discussion:

  A
Actual Gs		 B
Total Load		 D
Cap Load		 E
Cable Load
Ultimate - Gs -4.5 -24,015 -29,762 +5,747
Limit - Gs -3.0 -16,011 -22,736 +6,725
On jacks -0.95 -5,074 -13,134 +8,060
  0 0 -8,681 +8,681
  +1 +5,333 -4,000 +9,333
Limit +Gs +6.16 +32,864 +20,167 +12,697
Ult +Gs +9.03 +48,179 +33,613 +14,566

I have plotted these numbers in a graph; click on ALEC

The cable loads in the above table result from the following choices by the ALEC designers:

(1) Assumption of elasticity, after "slack" has been taken up
(2) Selection of a fixed and repeatable ratio between cap stiffness and cable stiffness
(3) Selection of one particular number in column C (1st table), probably the 6.30, in order to assign about 2/3 of the total load to the cap at ult. design load.
(4) Assumption that the preload 8060 lbs. can be accurately applied, measured, confirmed, and sustained in service.

(One very serious ALEC problem is revealed in the above table; excessive compression loads in the lower spar cap at negative Gs due to cable tension. The cap was designed and certified to withstand 16,011 lbs. compression at neg. 3 Gs. The ALEC cable raises that load by 42 percent, to 22,736 lbs. And aluminum channels don't make very good columns.)

Assumptions (1) thru (3) above are highly idealized perceptions about the nature of cables. Actual cables don't act that way. The ALEC cable consists of 19 wires. The outer 9 strands are helixes at about 26 deg. to the cable axis. 9 more helixes underneath are wound in the opposite direction. The center wire is straight. Between all the wires are air spaces. This combination could be preloaded to 8,060 lbs. but it is a wild assumption that the cable tension would then rise to about 1/3 the total at 6 Gs. Likewise there is no reason to believe that it will return to the same preload on jacks.

The ALEC design proposal uses a 10 mm 1x19 Dyform cable, The website http://www.s3i.co.uk/1x19Dyform.php contains information about these cables, and cautions about usage in applications where stretch is important.

Finally, a comment about Assumption (4) above. The photos in the article show flat spaces on the inboard shaft, and a large integral flange on the new 7/16 in. nut to transfer the load to the "compression transfer collar". To apply the preload the friction at the threads would need to be opposed with about a 3/8 in. end wrench on the shaft flats. Then the unknown friction at the large diameter collar would need to be countered during torquing. George may claim that a desired preload can be accurately attained by applying a calculated torque. I don't think the FAA will believe this.

According to George's article, he is trying to take $2,000 down payments for ALECs in advance of FAA approval and before the parts are manufactured. I hope the T-34 association members will carefully consider all of the above, to avoid further disappointment.

Sincerely,

Dick Wilson

April 29, 2006

________________________________________________________________________

U.S. Department Small Airplane Directorate
Of Transportation Wichita Aircraft Certification Office
Federal Aviation 1801 Airport Road, Room 100
Administration Wichita, Kansas 67209

December 1, 2005

Mr. Dick Wilson
6900 Los Verdes Dr. #7
Rancho Palos Verdes, CA 90275

Subject: Model T-34 Proposed Spar Modification "ALEC".

Reference: Your letter dated November 2, 2005, to members of the T-34 Association Board

Dear Mr. Wilson:

We recently received a copy of the letter you sent to the T-34 Association Board. I have been asked to review your letter and respond.

We have not received an application for any type of project for this proposed spar modification nor any data pertaining to it. We did receive a conceptual briefing on it, but no drawings or analyses upon which to make any assessments.

You make some good points in your letter that we certainly will consider when we receive an application to start the project. The detail design of the parts, assembly and installation will be important not only for the initial installation, but also, the continued airworthiness. The FAA will approve the design when we are satisfied that the applicant has shown compliance with the applicable safety standards.

Your assessment certainly has merit and it will be interesting to see the T-34 Association's response. Be advised that when we receive an application for such a project as this, any data that we receive from the applicant will be proprietary in nature.

Sincerely,

Eual M. Conditt, Jr.
Associate ACO Manager, Airframe & Services
Wichita Aircraft Certification Office

______________________________________________________________________________________________________

 

July 29, 2006 The Saunders Spar Strap

In the T-34 link, Nov. 2, 2005 addendum I analyzed the "Aircraft Life Extender Cable" (ALEC) and showed that it should never leave the ground. Here I will present similar arguments regarding the Saunders Spar Strap.

On the AviaDesignWebSite there are many soothing claims for the Spar Strap; two of them are:

(1) "It creates an independent redundant load path that eliminates the possibility of wing separation in the event of failure of the spar cap, bathtub fitting, bolt, nut, or center section carry thru structure" That is impossible. The wing attach bolts, bathtub fittings, etc. lift the fuselage. The strap has no capability to transfer any of the wing lift to the fuselage.

(2) "It drops the stress levels in the existing spar by 40%"

I learned in a phonecon with an Avia person that the strap is preloaded by installing it with 500 lb. weights on the wing tips. He said that they proved by analysis that this produces a negative (compression) in the cap equivalent to minus 1.5 Gs, and that results in about a 40% reduction in cap tension at 6 Gs. He didn't say how they managed to get the airplane re-certified for this new wing tip load condition, not covered by the Type Certificate. But the main problem is that Avia seems to have simply assigned 40% of the load to the strap, then they derived that number with an "analysis".

Without the strap, the T-34 spar cap tension is approximately 30369 lbs. at 6 Gs (see Fuselage May 20, '05). That's about 12 autos hanging from an aluminum bar. If the average stress is 25,000 psi at 6 Gs, then the strain is 25000 divided by 10,000,000 (Young's Modulus) which comes to 0.0025 inches per inch of length. The cap length between the fuselage center line and the strap attachment is approx. 70 inches, so the elongation of the spar (without the strap) is 70 x 0.0025 = 0.17 inches. It stretches from 70 inches to 70.17 inches when the G meter reads 6. The task they assigned to the strap was to reduce that 0.17 inch stretch to 0.10 inches

The strap is installed slack, wrapped around `the soft underbelly of the fuselage, screwed to a doubler which is riveted to another doubler which is blind-riveted to the 0.020 inch thin outer wing skin. The skin is riveted to the aft segment of the "piano hinge" and the mating segment of the hinge is riveted to the spar cap. This circuitous load path is supposed to be so rigid that, at 6 Gs, it will reduce the cap stretch by 40% of 0.17 inches or 0.070 inches, with a force of 40% of 30369 = 12,000 lbs. That's about 5 of the 12 autos hanging on a string of hardware comprising a foot or two of the piano hinge (adjacent to the bolts), cherry-riveted to a 0.020" thick aluminum strip etc. etc. and the whole strap load path stretching only 0.070 inches per side under that load.

The web site states that the doublers "spread the load over a large area through the existing wing skin, ribs and stringers" and "the load per rivet is less than 20 lbs." (12,000 lbs / 20 lbs = 600 rivets.) This is an imaginary, roundabout part of a load path from the strap back to the spar cap. The wing skins fore and aft of the front spar are not tension/compression members. They are simply designed as stiff panels to apply suction on top, and pressure on the bottom, to the spar. Note that the main gear doors are just covers.

The attempt at a preload during assembly will result in the following: If the airplane is on jacks the lower spar cap between the jack point and the strap attach. will be compressed approx. 0.031 in. due to the 500 lb. weight. Then after the strap is screwed on and the weight is removed the cap in that length will extend back 0.031", taking 1/32 of an inch out of the looseness of the 70" strap assembly from the centerline thru the piano hinge.

To keep this discussion simple, I haven't mentioned the strap attachments forward of the spar. There are actually 2 hinges, and they would carry approx. equal portions of any strap load.

I don't believe that the strap can significantly effect the spar cap tension at any load factor. A good project for the T-34 Association might be to remove a strap from a T-34, place strain gages on both sides, calibrate them in a test rig, then reinstall the strap and take readings in flight. That probably won't happen, but I wanted to suggest it.

---------------------------------------------------------------------------------------------------------------

A note regarding aerodynamics: Beech engineers carefully designed the wing airfoil to NACA series specifications. And they built in -3 deg. of washout (twist) to initiate stall inboard; an important safety feature. This twist results in a reduction of wing bending moment.. The disruption of the inboard airfoil shape by the strap might shift the center of pressure outboard and cause a greater spar cap load than it was designed for.