Seat Construction

I was trying to figure out what would be the most comfortable seat position for a pilot in a reclined position.  My Goal was to reduce the height of the fuselage to reduce the projected frontal area. The other problem was the carry through of the main wing spar. I was aided in this problem because I decided early on to place the spar at 33%, this would place the spar aft of the center of mass of the pilot. In the design of an ultralight, because the Pilot weighs as much as the airplane (At least in my case!!) its a good idea to place the pilot in the center of the CG range of the plane so as to avoid the need for ballast.. One day I was reading my latest addition of Soaring magazine and I saw an ad for Jantar Sailplanes. In the ad they were boasting that their seat was the most comfortable in the industry. They had a inboard side profile of the cockpit and seat. Well if it is good enough for Jantar, then it was good enough for me!!  I scanned the ad and did an underlay trace on my Cad System (Rhino 3D) and I then had the seat curve. I orientated it to clear the spar and place the pilot on CG and then designed the structure around it.



Seat back up structure being installed.



The seat back up structure is designed to the bottom contour of the seat. One of the design criteria I adopted was to ensure that the seat did not deflect under max G loading such that it interfered with the aileron control torque tube that runs under the seat itself. In this picture temporary jigging is fixed to align the 3 seat supports. Most of the secondary seat support structure is made from Spruce Scrap and or Birch plywood. The contours and shapes are laid out from dimensioned drawings on the plans. For ease of build in areas like this, I plan on having laser cut card board templates made.



Seat center support intercostals

 There are two parallel seat support intercostals that not only serve to keep the pilots behind from deflecting the seat and jamming the controls, but these also provide the mounting provisions for the aileron torque tube. The Intercostals are built up of 1/32nd plywood and 1/2" spruce square stock. All glue joints used T-88.









seat supports completely installed

 The center intercostals were mounted to the fwd face of the main wing box. External stiffeners tie the intercostals to the wing box, like wise on the landing gear frame. In the final plans I will clean up this area a bit. I had to add plywood fillers that were not planned for originally. I will modify the gussets of the landing gear frame to extend them such that they act as fillers. I will also redesign the truss members to eliminate some redundancy . The intercostal stiffeners could become a part of the wing box internal truss, likewise for the landing gear frame. It will probably save 5 to 6 ounces.

One of the reasons I love to design ultralight airplanes, is the discipline it gives you in the area of light weight and tight design. I remember when I hired on at Beech Craft in Wichita, I was making a bit of a name for myself locally with my Wren Design, and I started receiving calls with job offers. When I was interviewing with the various engineering groups, the guy running the design group for the Starbarge (Star Ship) looked down his nose at me and said, "we don't hire Ultralight designers on the Starship" That was a fact!!!  It was 5000 lbs over weight and out of the class. Beech petitioned the FAA for a special category so the lead sled didn't have to meet Airline design criteria. I guess they had all the "Ultraweight Designers" they needed. I ended up doing manufacturing R&D, but that's another really good story!!!


another view of the seat supports.

 a short intercostal that spans the gap between the center intercostals will be added later to support the torque tube. There is a cutout for a Delrin Shoulder bushing.

Seat Flat pattern being trial fitted

I did not want to actually mold the seat. So I hit on the idea of laying out a seat flat pattern from 4.5 oz PVC foam . I pieced together the cores bonding them with a mixture of Micro balloon and Epoxy. The actual pattern was taped together Poster board cardboard. I have seen this forming technique used to make ECS ducting on commercial private jets. I filed this technique away for future use. The future is NOW!

Only one side of the foam is actually laid up, and that is the surface that will be in tension. Now the seat curve reverses and because of this, you do not lay a layer of glass the full length on the bottom surface. The glass stops where the panel curvature reverse. The seat is really nothing but a wide thin beam. The fiberglass laminates on the faces alternately resist tension and compression forces. If only one surface (Tension side ) is laid up and the seat is deflected, the foam itself experiences compression. And since the foam is relatively weak and is of low modulus (Stiffness) it readily forms to the curve of the underlying seat structure. This is why its important that all of the seat structure be in align and level.


Seat being pushed into position.

 In this picture the seat is pushed to the support structure. After a trial fit, I removed the seat and taped up all of the seat support structure and the side truss. The seat was then put back into position and hot glue was used to temporally hot the seat into position.

seat being held while the hot glue cools.

The front of the seat lower surface was left bare and not laminated. I used 1 single ply of 120 style glass for the bottom surface. Since this surface was in tension, there was no need to have a thicker laminate because thin section stability is not an issue . I have experimented quite a bit with foam and glass construction, its against intuitive thought, but the lightest laminates I have made were with 4.5 lb/cu/ft PVC foam. PVC foam is a closed cell foam, this is important because wet resin will not continue to wick into the core, and in the process draw in air. This of course will add extra weight. Laminates over foam usually start with a slurry coat of micro balloon and resin, usually Epoxy, that is trowelled onto the surface. The idea here is to fill the exposed cells with a lighter weight material than the raw resin. When the foam is cut at the factory, the cut cells are exposed, the lighter the foam, the larger the cells. I have found that 3 lb density foam has too large a cell structure, this is why the 4.5 lb density is lighter.



Another view of the fwd edge of the seat


 I elected to form the fwd reverse curve with a heat gun. I will not do it this way in the future, I would tape the lower surface with aluminum duct tape and then hot glue the foam to the supports.

once the seat core is formed into position, a single ply of 8 oz 285 weave style glass was laid up. You could also use a  8 oz plain weave because of the absence of compound curvature. This would be a little cheaper.



Seat after final trim

This is the seat after it cures and the edges are trimmed. The edge treatment is to lightly sand into the exposed foam and a mixture of Micro balloon and resin mixed to a consistency of bread dough is pushed into the edge. After it cures the edges as sanded smooth. The final seat weighs 15 oz. The seat needs to be removable to gain access to the wing attach pins, so it is held in place with twing strips of Velcro tape

Landing gear construction.

Landing gears are an especially tough component to design. They add useless weight to the aircraft, but are absolutely essential. Imagine designing a landing gear that was designed to strike the ground at the VNE (Velocity never exceed) of the airframe and then stroke to a point where all of the energy is absorbed and the airframe is neither over stressed or the landing gear breaks. Congratulations!! you have just designed the Peregrine Falcon talons. This  remarkable piece of engineering  like the design of  Spruce wood, was the work of a genius! and I adore him as a matter of fact!!! I have a design manual set by Dr Jan Roskom of the university of Kansas. The sections on preliminary design are the bible for the industry. He has accumulated an immense amount of information, or rather his Grad students have LOL!!!. In the section on landing gear he has a carpet plot that equates the weight of the landing gear to the gross weight of the aircraft. There is a line that holds true from the lightest (J-3 Cub) to the heaviest (C5 Galaxy) aircraft sampled that has the gear weighing around 3.5% to 5% of the gross weight. My landing gear complete with wheels, tires, axles, springs and all fittings weighs 7.2 lbs for the main and 1 lb for the tail or 1.3%. Because of this, this is actually an area of risk on the Robin.

The landing gear spring is the heart of this design, its constructed as a sandwich composite beam.



bonded gear legs

 Spruce wood is used as a core and strips of Graphlite rectangle flats are used as a spar cap. The whole bond assembly is then encased in a woven 45 degree knitted 8 oz Graphite sock. This resists the torsion loads induced by braking. The Graphlite sections were purchased from Goodwinds in Mt Vernon Washington. They are a refinement of the original Nepco pultruded carbon rod product developed many years ago for the Copperhead guided artillery shell program. The original .050 diameter rod was coiled up on a bobbin at the base of the shell. The outer case would shed after firing and the .050 rod would spring out and become a trailing wire antenna. Later versions of these rods were made in various diameters, the most popular being .125" diameter. This is because the bend radius is 19 inches and  means that they can be coiled for inexpensive shipping. The .125" rods were used by Bell Helicopter in the V-22 wing skins as pre-cured unidirectional reinforcement for the wing skin hat stiffeners. They later abandoned that process, because it was cheaper for Bell to lay graphite tape by robotic tape laying machines. Nepco went out of business or sold the Pultrusion business. It has changed hands a few times, but  is being manufactured by Goodwinds, a high performance kite company. The kite fliers and model builders were early adapters of the .050 product when it hit the surplus market.
Pultrusions are similar to extrusions, but as the saying go's "You can't push a rope" They make pultrusions in 5000' lengths at a time. This is the amount of material on a creel bobbin. The fibers and gathered up in redirection combs until they come together at the opening of a die. Prior to entering this die, they are impregnated with resin. Once in the die the resin is cured by either progressive heating or microwaves. The product that exits the die is one tough piece of structure. Pultrusions have the highest ratio of resin to fiber than any other composite. Typically around 85%. There is a great deal of tension applied to these fibers as they are pulled through the die and this tension is trapped into the fibers. Typically this is around 15,000 psi. This means that for all applications where the pultrusion is in a beam in bending, the pultrusions placed in compression, actually see "less tension" for the first portion of the loading. This makes them very stable in compression. The Robin uses these pultrusions for the spar cap, and for the first 1 G of positive loading, there is no compression in the upper spar cap.

This is a bit of a digression, but the reasoning is exactly the same. One of the greatest military secrets of the British Navy in the days of sail, was the design of their masts. The British ships were well known for their ability to take tremendous punishment when the rigging of their masts were shot out. French and Dutch ships masts tended to break in similar circumstances. The reason was that the British insisted on using whole logs for their masts. The Dutch and the French would laminate up their masts from smaller cut sections. This is actually the reason Britain colonized Australia, to gain access to their virgin timber. When wood grows the outer rings are the growth rings and the core is the old wood, or compression wood. When a mast is fully rigged, the main loading in the mast is in compression. Once the rigging has been shot out, the masts would bend. The British always stepped the base of their masts such that they would resist bending. When a whole log bends, the outer fibers which contain trapped tension, can deflect at much greater angles because they see "less tension" before they see compression. Cut wood on the other hand exposes wood already in compression and the result is a compression failure. And that is the reason Britannia used to rule the waves.

Back to landing gear. The reason I am using flat springs in a molded beam is because of a phenomenon called creep. All plastics creep. The resin matrix of a composite structure is plastic, and it will creep. Creep is the slippage of the resin molecules relative to each other when a constant small load is applied. If a curved beam is molded, the outer fibers of that beam under a 1 g normal loading have a distributed tension component all along the outside of the curve. This tension load in the fibers is normal to the fibers direction and the vector is pointing away from the surface.  This small tension load will cause the landing gear to splay apart. When you use a flat beam, the normal 1 g loading causes the gear to deflect in a concave direction. The distributed compression load along the outer surface is now pointing inward, or into the laminate. Plastics do not creep generally in compression.


Pultrusions being bonded under vacuum

 An added benefit of using pultrusions is the relatively high modulus of elasticity, it is however not as high as Aluminum or Steel, so careful attention must be paid to the amount of deflection the gear has, especially sitting on the ground at 1 G or during taxi. It is desirable that the axles be parallel with the ground in this condition for easy ground handling. This leads to a gear that has a great deal of camber when unloaded.

The amount of deflection of a landing gear determines its ability to absorb energy. These types of gears do not actually absorb energy, like a rubber bungee gear would. Rather the energy is stored and added back to the airframe in the form of a rebound load. Actual energy will be absorbed by the lateral scuffing of the tires when contact with the ground is made on landing. On the Robin I decided to isolate the airframe from the bending loads of the gear. This would have required a much heavier support frame to react the bending loads. I have connected the two gear halves together with a welded rectangular tube section. The center joint is capable of transmitting moment from one gear to another. By using this method, I also get higher gear deflection and soften the landing loads.





center joint and clamp blocks


This center joint is bonded to the gears and has a large overlap so as to allow the bending loads to transfer by a wide couple. The ends of the tubes are also chamfered at a 45 degree angle to reduce the stress concentration. The gear itself is clamped up in two fittings and pin jointed to the frame. There is a compression/tension load that is induced into the frame due to the fact that both sides are clamped and not allowed to slide or translate. This loading was accounted for in the design of the frame. The ends of the







Landing gear being rigged in position


gear are designed similar to the ferrule of a chisel. They are bonded to the ends of the gear and have the axle and brake supports added. This area has undergone a little trial and error. As has the actual gear itself.

I have revised the end fitting design and the gear height. I just finished building a brand new gear because I was lacking enough ground clearance for the new GSC prop I ordered. I had estimated that I would use a 48" prop, optimised for





original gear drawing







soaring, but I decided to use the 56" prop because I am in love with rate of climb (see Wren page!!!) I have verified that the calculated deflections exactly match the actual deflections. The earlier gear was test loaded by sand bagging the pilots seat to 2.5's. a 1/2 g above the design ultimate load.

original gear at 3 point

The tail spring has undergone a design evolution also. originally I designed and sized a tail spring using 4130 Steel. I did not want to pay for heat treating so I sized it for the normalized condition. Eventually this proved to be inadequate. All of my hanger flying started to creep the gear such that I noticed the tail wheel caster angle starting to increase. My first try at a composite gear was to use the same Graphlite flats that I used in the main gear combined with a sandwich core of NEMA grade C fiber Glass plate. This not only proved to be heavy, but I completely neglected the torsional loading that would occur with the tail wheel kicked over.

This design would actually make a damn good torsion spring. So I went back to the Goodwinds catalog and found a pultruded graphite tube of 3/4" diameter x .125 wall. I bought this tube and the purchased some 7 oz knitted carbon sock and bonded it over the outside . The tube would probably have been able to resist the torsion loading, but would have done so by differential bending. This is an inexact load path and could have had eventual delamination issues. I sized some 4130 steel tubing to slip over the outside and act as a ferrule. The tube passes through a solid plywood bulkhead where 1/2 the bending is reacted. The center of the gear has a sleeve tube of 4130 welded to  a support fitting. The tube is bonded into the 4130 tube, after appropriate surface prep, and this mount reacts not only torsion, but the also couples with the bulkhead and reacts the bending



center tube support and fitting



Engine installation




 I am using the MZ-34 from Leon Massa at Compact Radial Engines in Surrey BC Canada. This is a single cylinder 28 hp engine. This engine was originally designed in Italy by Zanzottera as a back pack powered Parachute engine. Leon has up dated the engine with Nikasil coatings and absolutely exquisite machining and internal compression releases.The engine comes with a pull starter and an electric starter as an option. I have eliminated both of these features on my prototype to save weight. I will need to redesign the wings to an all wood design to save enough weight to allow for an electric starter, more on that later!! 

The main reason I bought the MZ-34 is because of the mounting system. It uses 4 radial shear type mounts placed around the center line of the crank. This is similar to the mounting system of a Radial engine, but not quite as refined. It is however greatly improved over a flat bed mount like the Rotax 277.
 Years ago I worked for Ford Motor Company in the NVH lab (Noise, Vibration and Harshness) our job among other things was to isolate vibration and eliminate noise. One of the basic principals of vibration isolation is to convert the vibration into heat. This is accomplished in a rubber mount system by absorbing the vibration by shearing the rubber. In order for a rubber mount to work, the fuselage side of that mount must be held fixed in all 6 degrees of freedom. Any movement in any axis will result in vibration transmission. So far this is all theory on the Robin, because I have not yet started the engine. You can all drop by in a few months to see if I have to eat some Humble Crow Pie!!! But you have to start with a theory!!

  So here is a side view of the engine installation



Notice the 4 struts that connect from the longerons to the fire wall.  The fire wall is a structural foam and glass sandwich of 1/4" 4.5 lbs PVC foam and 120 style fiber glass. This will eventually be covered by a layer of Quartz Fiberfrax for a heat shield. The fire wall itself has a great deal of in plane stiffness in the inboard/outboard and up/down directions, this effectively accounts for 4 of the six degrees of freedom. The struts account for the stiffness in the fore/aft direction. There will be an out of plane vibration mode due to the angle of the struts, but that is again reacted by the in plane stiffness of the fire wall.



mounting locations


The engine was mounted in space per the layout of  the plans and the location of the 4 attach points were located. A hole saw was used to cut through the fwd face of the fire wall only. Glass inserts  using 1/4" commercial NEMA grade fiberglass plate were fabricated and bonded into place. A doubler of two ply's of 120 glass were laid up over the top of the inserts. 




core removed prior to insets installation

The engine was temporarily installed after this.



MZ 34 Engine  installed



Fuselage Construction

This is an area where I have spent the largest portion of my time in both design and construction. I have redesigned the fuselage 3 times. My original concept was similar to my original ultralight the "Wren" , a fuselage built from 4 sheets of 1/2" Styrofoam. The foam would have been covered by a light weight glass. I couldn't make the weight on this design, I had tried two iterations before I threw it out and discovered Natures composite, Spruce wood!!!  The following is a picture of the original fuselage truss I designed.

  Now this may seem minor, but I could not show this design "Good" or structurally viable. The main loading for the aft portion of this fuselage is a downward acting tail load. It is always a percentage of the main wing loading. Usually around 10%. An airplane actually carries 110% of its design gross weight in the max pull up design condition. This is the balancing tail load that reacts the wing pitching moment. So in this design, for positive loading all of the diagonal members of this truss will react this load as a shear component. The angle of the member and the direction it faces will determine the actual load (vector sum of the load) and whether this load is in tension or compression. To simplify this design I chose to build the truss with 1/2" x 1/2" spruce sections. The diagonal members are longer than the vertical members, this means that they become a much longer column when placed in compression, I could not show most of these members "good" based on the compression loading. The solution was to reverse the members such that they are all in tension for the positive loading. Remember, negative loading which WOULD place these members in compression  is 1/2 that of the positive load. Its a simple trick, but it saves almost a pound!! All of the longerons are doubled up 1/2" x 1/2" sections. The pitch of the vertical members was determined by the critical length of the longerons in compression. The design of the gussets by the way is very straight forward and conservative. The assumption I made on all gussets is that they should be able to transmit the net tension load of the attaching member. Since all attaching members are 1/2" x 1/2" , I assumed a tension load equal to the max allowable of the spruce and added a safety factor. This is conservative, because actual loading is less than this design stress. Most members are actually sized for compression. Here is an Isometric view of the final fuselage design. 




Fuselage Isometric everything in this picture weighs 26 lbs


Construction of the fuselage started by building a flat level work table 17 feet long x 4 feet wide.

   





Fuselage jig table

  








I used particle board and painted the surface white. The base had sufficient cross members to keep the surface from sagging. I used the old wing assembly jig as a table base. The next step in the construction is to lay out the full size jig from the fuselage plans.






marking jig



I made a pretty useful jig tool here. This is used to locate the bevel or end trim of a member. The untrimmed spruce member is set on top of the existing trimmed pieces. The jig is set on top of this untrimmed piece with the two blocks aligned to the trimmed member. The bevel angle is then drawn on the untrimmed member. I usually cut to within a 1/16th of an inch and then carefully sand the rest. I will do this numerous times until the fit is exactly line to line. Not too tight to where the adhesive is forced out, or too loose where it could actually flow out.




jig being used

 

Wood can only be ordered economically in 109" lengths. Because of this the longerons need to be spliced. The type of splice I used is a 15:1 scarf splice. I used a scarfing jig on my table saw to cut the exact same angle in both pieces of the longeron. The longerons are then splice "insitu" or during the build up of the side truss. The 12:1 to 15:1 taper is a standard aircraft wood splice angle. The idea here is to have more bond strength than the ultimate tension of the piece being spliced. A proper splice will fail in the wood before the glue joint. Its important not to sand these edges, sanding overturns the surface wood fibers and does not allow the epoxy to flow into the exposed wood cells. All cutting was done on the table saw using a hollow ground planer blade. The long locating blocks on the assembly jig board are screwed into place and are used as clamping blocks for the splice. Excessive clamp force must be avoided to keep from forcing out all of the adhesive. Remember, epoxies only require contact, not pressure for a successful bond. The jig board was coated with 3 coats of carnuba paste wax  to ensure everything would release.


Scarfing jig



The fuselage sides are identical for the L/H and R/H sides. The joint between all members is made with 1/16th" plywood gussets. Gussets are used on both the inner and outer surface. So when a pattern is made for one gusset, it is duplicated 3 more times. There is no real drawing of a gusset, just minimum areas that need to be maintained. The idea here is that the gusset will be sufficient in tension area to develop the full ultimate tensile load of the member. This simplified my analysis because I only had to worry about a single master gusset. In the design of a truss, the idea is to have all of the members intersect at the centroid of their sections. This eliminates any induced out of plane loading. from a practical stand point, its pretty tough to do, so in the plans I approximate it. Attach the gussets with T-88 adhesive and 3/8” staples. Shoot the staples through a piece of heavy Dacron fabric or Peel ply. This make removal a snap after the bond cures.

The following picture is the completed fuselage side and my dog Boo (Aussie Sheppard). Both sides are built up in the truss jig and gussets are attached to both surfaces.

 


side truss and "Boo" my Australian Shepard

 

Once the side truss is completed, the builder has the option to either make a new top view template or clean up the side truss template and lay out the top view template. Since the Robin was fully modeled in a 3D CAD system, it was possible to lay out an exact developed flat pattern of the side truss. In reality the flat pattern is slightly longer than the actual side view of the truss. This becomes apparent when the sides are brought together on the plan view template.



side truss jig


The procedure for building the completed truss is similar to building the side trusses. The fuselage is constructed up side down because the upper longeron is flat. When cutting both the horizontal members and the gussets a duplicate is also made for the lower longeron. When fitting the horizontal members, use the jig earlier described and cut an accurate bevel. I usually cut a piece long and then sand the parts on the belt sander to a perfect fit. When attaching the lower longeron horizontal members, be sure that the pieces are parallel to the lower longeron surface where the gussets are to sit. Make some large 90 degree triangles out of particle board to act at squares to keep both sides square and parallel.



Truss flat pattern



After the side trusses are attached, let the fuselage sit in the jig. At this time you will build up the wing box carry through and the landing gear attach frames. These can be built up ahead of time and installed when the two sides are jigged together. Both frames are built up separately and then jigged into position before the additional box members are added. The main wing attach pins are used to locate the wing frames parallel to each other. The landing gear frames have two large bushed holes; these holes are used in conjunction with a maple dowel pin to align the frames. After the frames are bonded into position, the dowel is removed and the bushings are replaced. 



completed basic fuselage truss. 21 lbs


Here are some more pictures of the completed truss.



after the truss is finished, build up starts for the  Roll over cage and turtle deck. The Vert fin is atached and the  landing gears can be assembled  and Whoopee!!! A milestone is reached!!  "Weigh on Wheels" Normally in the business, everyone gets to go to the main assembly area and they give you an ice cream bar. Because I'm a little smaller than than the "big boys" I just cracked a Bottle of Merlot and a Neighbor and I killed it!!!

Weight on Wheels