An Engineer's Close-up of the Curtiss-Wright Transport

by George A Page
Chief Engineer St Louis Airplane Division of Curtiss-Wright Corporation
As Told To Jay P Au'Werter
Assistant Editor, Aviation
In the design of the fuselage structure of the Curtiss-Wright Transport, two fundamental design considerations predominated. These were
  1. the development of a shape which would allow a large cargo capacity and would locate the disposable load as near to the center of gravity as possible, and
  2. the development of a shape and structure which would allow pressuratition of the cabin for high-altitude operations.
On the basis of these considerations, the Curtiss-Wright engineers developed a basic cross-section composed of two intersecting circles joined together by a common chord through the points of intersection. The use of this cross-section presents several advantages allowing as it does a more efficient utilization of the enclosed space as compared with a similar circular cross-section, while the use of circular arcs above and below causes the loads on the fuselage wall due to the pressurization to be taken out as tension loads in the structure. This shape is employed in the region of the cabin, at the widest portion of the fuselage. Aft of the cabin the shape is somewhat modified, tapering to a tail cone.

The fuselage is of the semi-monocoque type of construction, having a shell made up of reinforced 24ST Alclad sheet, with stringers and forming rings of the same material. Forward of the bulkhead which is located just aft the pilots' compartment, the skin is flush-riveted, while aft of this point modified brazier head rivets are used.

Fifty-one fuselage rings are employed. These are formed channel sections of 24ST Alclad sheet, with the exception of those rings to which the center section of the wing and the tail wheel are attached. The rings are riveted directly to the skin.

The skin is reinforced by stringers of 24ST Alclad sheet rolled into a Zee section, this particular cross- section having been chosen only after extensive tests of various types and shapes had been made. Formed sheet material was used for the stringers in preference to extruded shapes because of its lightness and the ease of rolling it to the desired cross-section, and because it may be obtained as Alclad stock, that is, coated with a corrosion-resistant coating of pure aluminum. Also the use of stringer material of the same specification as the skin allows a more homogenous structure than would be possible with the use of extruded shapes, the physical properties of which are somewhat different from sheet.

A feature of the fuselage shell lies in the pattern and layout of the stringers. A constant circumferential spacing is maintained between any two stringers throughout the length of the airplane, individual stringers dropping out by merging with a central keel member as the diameter of the fuselage decreases. This eliminates stress discontinuities caused by abruptly-terminated stringers. The stringers, which are riveted to the skin, pass through cutouts in the rings, clips being employed to fasten the strmgers to the rings and also serving the purpese of stabilizing the rings themselves, allowing them to develop a higher allowable stress.

A pressure rib is installed in the wing center section, on each side, in approximately the region which would be occupied by the fuselage wall. The function of this rib is to form a pressure seal for the wing for pressurization of the fuselage, and at the same time to act as a shear beam in transferring some of the loads between the spars of the center section at the fuselage atthmentnt. The lower portion of the fuselage structure in this region is attached to the lower surface of the center section by means of a splice angle, the skin and outer portions of each ring in this area being attached to the splice angle, while the inner edges of the rings are clipped to the skin of the center section for stability.

The floor of the airplane, which is the connecting member between the upper and lower arcs, becomes a structural member of the fuselage when pressurization is installed, taking out as tension the loads arising from the pressure loads on the fuselage walls. The floor is constructed of 24ST Alclad sheet set upon transverse channels of the same material, rolled to shape. These in turn are supported on longitudinal girders which are fastened to the fuselage structure at the forward and aft ends of the floor and to the wing center section. The transverse channels are bolted to an extrusion which runs forward from the pressure bulkhead to the nose of the airplane at the intersectnon of the upper and lower arcs on both sides of the fuselage. This connection serves to transfer the floor loads directly to the fuselage structure.

The tail wheel assembly is mounted in the rear of, the fuselage and retracts into it. The tail wheel ground and retraction loads are transferred to the fuselage structure through special reinforced bulkheads and attachment fittings.

When the Curtiss-Wright Corporation's St Louis Division engineers began designing the new 36-passenger Transport which has just recently been completed in that factory, they were not planning to produce the first twin-engine airliner ever designed for substratosphere flying. They were thinking rather of turning out an airplane that could perform a specific function efficiently; they thought in terms of operating economy, high payload ratios, low maintenance costs, and high performance characteristics. How the engineering details necessary to produce these characteristics were carried out, has furnished fuel for many a drafting room conversation throughout the industry. It is the purpose of this article to give an engineer's eye-view of how a few of the transport's salient features were worked out.

Safety features have been designed and built into the ship both structurayly and aerodynamically. One of these safety assets is the provision against wing tip stall. The means used has been the gradual alteration in aerofoil section from root to tip, coupled with use of moderate washout of angle of attack. This is not an untried design but is rather a modification of previous designs used on Curtiss-Wright ships. From experieece, records show that structurally the landing gear has been the weakest point in the past. To reduce to a minimum possibilities of landing gear failure, an extra length oleo stroke has been used that will withstand a descent of 800 ft/min, far greater than that normally used or required. In addition a locking device has been installed that will not permit the gear to be raised once in the down position with the weight of the ship on it. Another example of the thought of safety that was carried throughout the design is the location chosen for the fuel tanks. These are in the outer panel of the wings, at a distance of near 30' from the cabins, reducing the hazard of fire in case of an emergency. As a safeguard against pilot error the Curtiss-Wright "Tell Tale" instrument was devised as a visual indicating unit showing by signal lights when various control of the operations have been omitted or improperly performed. (See Aviation, March, 1939, p 54).


Turning first to the problem of pressurization to meet high-altitude flight, giving increased passenger comfort for normal flight operations, several special stumbling blocks were uncovered. In the general desngn procedure for weight economy, cabin pressure must be retained by circular elements wherever possible. The complete circle is, of course, ideal from a pressure standpoint but is not always the most suitable for space distribution required. One method of obtaining considerably greater depth than width is by means of a cross section consisting of two non-concentric arcs with the floor serving as the stay member at the points of interction. This type of construction was described in an article on the transport by T P Wright, Vice-President, Curtiss-Wright Corp, in the August 1938 Aviation (p 29). Once this type of construction has been chosen there arises the problem of the necessary strength requirements to withstand all of the various loads imposed upon the fuselage. Consolidating these loads into three groups we have: (a) Those due to wing, tail and landing gear reactions; (b) those due to useful load or dead weight items located within the fuselage structure, and, (c) those due to the differentials of pressure between the inside and the outside of the cabin.

With these loads in mind the problem develops into one giving a structure which would satisfy five different conditions, namely, that:

  1. The maximum imposed load likely to occur shall not impair or permanently deform the structure;
  2. A sufficient reserve of ultimate strength should be provided to cover exceptionally severe loading conditions. This requirement is intended to provide the highest possible degree of ultimate safety, so that if some damage, such as permanent distortion of structure should occur, it is of no special concern under these conditions.
  3. The structure should have a margin of strength sufficient to cover a reasonable amount of deterioration during its service life. This includes such items as wear, fatigue, corrosion, and reduced strength due to normal repair work.
  4. A reasonable allowance should be made for normal variations in materials of construction and fabrication processes.
  5. Allowance should be made for uncertainties of analysis methods and criteria.

These five considerations constitute the basis for the establishment of a design load factor for practically any condition where strength is of basic importance. Following along with the problem it appears that a practical design might be accomplished by use of the following design load factors:

  1. For a steady flight condition superimpose upon the loads due to normal unaccelerated flight those due to the maximum probable pressure differential, the latter to be multiplied by a constant which would provide for fatigue, creep of the material, and the excess pressure possible before the safety valves would open (the factor is between 1.25 and 1.50).
  2. For momentary accelerations the pressure loads without the factor for fatigue should be superimposed upon the critical limit flight loads. It was with this analogy that the pressurized cabin was designed.

Along with the solution of this problem many smaller but very imrtant t problems present themselves. One of these is the riveted seams. This problem arises from the desire to use a minimum number of standardided rivet patterns, reducing productnon costs while at the same time obining g a joint that would have required strength under all conditions. The result became a multiple-row pattern with generous spacing of outer row rivets. The rivet rows were placed close together since by this arrangement the difference of strength between the sheets is reduced to a minimum, and the unit loading of the outermost rivets is not appreciably higher than the average for the entire pattern.

Once the rivet pattern had been decided upon it is necessary to find some method of sealing the joint so that the pressure will be retained inside of the cabin. After a considerable amount of research with test specimens, it was found that one of the most reliable and at the same time economical methods of sealing was merely to paint the inside of the surface of the seams with a suitable compound. The latter should have moderately heavy body and be of permanently gummy nature which does not become brittle hard with age or at low temperatures of high-altitude flying. Where gaps are encountered as in the intersection of stringer with a pressure rib, a strip of cloth can be used as a base for the fill or sealing compound. Sealing the door is done quite differently, however, because neither the door nor the frame can be considered to be especially rigid. A flat tube is inserted in the space between the door and the frame and is inflated, after the door is closed. The de-icer pumps furnish a convenient pressure supply for this purpose. To offset its apparently greater complexity, this system has the advantage of creating an air-tight and therefore a more sound tight seal even when flying without pressurization.


Because of the added weight due to pressurizing the cabin, it was of utmost importance that the remaining parts of the ship combine a great strength with as low weight as possible. With this in view, the design chosen for the wings was of the stressed-skin type, in which the axial loads due to the bending of the wings under load, are carried in the stiffened skin. The stiffened shear webs have extruded flanges next to the skin. In the center section of the wing three such shear webs are used, but in the outer panel only two are employed.

In addition to the shear webs, support is given to the skin by rolled-sheet stringers which are laid in a "Herringbone" pattern largely termating at their outer ends into the main beam, thus avoiding the stress discontinuities which arise from abruptly-terminated stringers. These stiffeners, which are of a special ha-shaped design selected by test to give the maximum strength-weight ratio, are so spaced that their support is practically uniform over the entire surface, thereby eliminating the tendency of the wing skin to wrinkle under normal loads. This minimizes a source of parasitic drag which has heretofore been objectionable in large stressed-skin wings. As a further aerodynamic refinement, the stringers are attached to the skin by means of flush rivets, presenting a smooth, low-drag surface. The stringers were fabricaed on special rolls from material of the same specification as the skin itself (24ST aluminum alloy) thus taking advantage of the condition of identical physical properties in both the skin sheet and the stiffeners.

The ribs are of sheet metal construction, of both the web and truss types. Web-type ribs are used in th outer panels, near the inner ends where they support the fuel tanks. Light, stamped sheetmetal ribs are employed in the vicinity of the wing tips. The two tank support ribs in each wing are of heavy, reinforced sheet, with padded openings to accomodate the fuel tanks. The truss-type ribs are constructed from rolled sheetmetal hat-sections, and are employed throughout the center section of the wing and in the outer panel outboard of the tank support ribs. In all cases, the ribs are not attached directly to the skin, but are fastened to the stringrs by means of clips. Due to the use of deep stringers, the ribs are spaced more widely apart than is common in similar wings.

The center section is continuous completely through the fuselage, supporting the engine nacelles and the landing gear attachments. The outer panels are bolted to the center section just outboard of the nacelles, each outer panel being attached by a number of special high-strength bolts. Easily accessible splice bolts are recessed within the wing contour thus having no external protuberances and reducing to a minimum the structural eccentricity.

Servicing, moderate repair work and inspections are facilitated by the removability and interchangeability of the leading edges and wing tips of the outer panels. In both cases these attachments are completely flush with the surface of the wing. Each outer panel is also jig-built to assure interchangeability, so that in the event of damage to one panel, another may be substituted for it with no special fitting operations required.

Wing Flaps

The flaps are of the full trailing edge, rearward-moving, slotted type, possessing advantages not found on the conventional type of split flaps now in use. Hydraulically operated, with a follow-up system to regulate their travel, they may be controlled to any position by a single movement of a lever on the pilot's control pedestal. Movement of the control lever to the required flap position opens a valve on the actuating cylinder, cauinsg it to move the flaps to the predetermined position, at which point the follow-up system closes the valve and stops the motion, maintaining the flaps indefinitely at that position until the control lever is readjusted.

The flaps are given their motion by the parallel linkages to which they are hinged. These linkages cause the flaps to move first rearward and then rearward and downward so that the overall effect obtained is to increase the wing area and camber. Thus small flap movements rearward will give a high gain in lift with little increase in drag — a condition desirable for improving takeoff performance. The last part of the flap in downward motion gives little change in lift, but produces a large increase in the drag — this combination being ideal for landing conditions. An appreciable gain in lift is obtained by means of the slot which opens between the leading edge of the flap and the wing when the flap motion commences. This slot smooths out the air flow over the top of the flap and improves the wing lift curve at high angles of attack.

The flaps are of metal construction, fabric-covered, and extend from the fuselage to the ailerons. The motion of the hydraulic actuating cylinder is transmitted through cables to the flap linkage mechanism, which is contained within the wing. The forward edges of the slots are approximately normal to the lower surface of the wing, allowing free air flow with the flaps extended, but providing no possilityty of ice formation even under the most severe conditions, which might impede the flap operation.


A noticeable difference between the cowling on the Curtiss-Wright transport and the conventional cowling is that while the conventional cowl discharges the exit air evenly around the periphery at the trailing edge, the Curtiss-Wright cowl is completely sealed around the upper portion at the trailing edge and all the cooling air escapes throug an opemng in the lower portion. This "tunnel type" cowling, as it is called, affords several distinct advantages. By eliminatin the cowl "skirt" opening in the region of the upper surface of the wing interference between the air flowing out of the cowling and the air flowing over the wing is avoided. This results in a noticeable decrease in drag and a favorable effect on the stalling characteristics of the wing.

Cooling is controlled by means of an hydraulically-controlled retractable flap which partially covers the exit opening in the bottom of the cowling and serves to regulate the passage of the air. This results in efficient cooling of the engine in all flight conditions, especially those of taxiing and climb, and regulates the flow of air to the extent that the engine may be operated at its most efficient temperature at all times, with lowest drag coefficient.

Structurally, the cowl is fabricate from 24ST aluminum alloy, flush riveted throughout. It is of the stressed-skin type of construction employing sheet skin reinforced with internal stiffening ribs of formed rolled hat-sections.

The cowlings are mounted on the nacelle structure independently of the engines, so that they have no motion with the engines at any time. This reduces the likelihood of fatigue failures due to engine vibration, allowing longer serviceable life, and presents more pleasing appearance at low engine speeds when engine oscillation amplitudes are larger.

The cowlings are of three-piece construction, each section occupying approximately one-third of the cowl circumference. Two of the sections are hinged at the top along the center line of the cowl, and open upward and outward. The third section is equipped with transverse hinge, just forward of the firewall, and opens downward. In this manner, the entire engine is accessible for servicing without the necessity of completely removing the cowling. If it is desired to remove the cowling, this can be done in a period of about five minutes.

Power Boost Controls

The problem of "pilot fatigue" has come to be recognized as a vital factor influencing the design of such items as cockpit controls, instrument installations, warning signals, etc. An important aspect of "pilot fatigue" has been found to be the actual physical fatigue arising from the work that the pilot must do to operate the controls against the heavy surface loads that are encountered in modern large aircraft.

The design of the "power boost" system, relieving these loads, is such that small movements of the pilot's controls open servo valves, which allow pressure immediately to build up in the Sperry Pilot hydraulic actuating cylinders, causing the control surfaces to move to a position determined by the amount of movement of the pilot's controls in the cockpit. At his point the valves are automatically closed. Any desired amount of control surface movement may be obtained by moving the cockpit controls a proportionate amount.

When reverse motion of the controls is desired, as in returning the controls to neutral, the servo valve admits pressure to the opposite side of the servo cylinder, allowing the oil in the first side to drain back to the hydraulic reservoir.

Aside from the small travel necessary to operate the servo valves, the pilot's controls are directly connected to the surfaces through the same cable and push-pull tube systems which are actuated by the power cylinders themselves, making direct manual control available in the remote possibility of failure of the power controls.

The system is so designed that as soon as the motion of the control column has stopped, the servo valve is closed, thereby locking the pressure in the servo cylinder and stopping the motion of the controls. The stop on the control column is so adjusted as to be contacted at the same instant as the stop on the control surface, thus halting the entire control system when the surfaces contact their correct stops, and thereby preventing excessive loads from building up on the surfaces from the power control.

The power cylinders which for the basis for the "power boost" are also part of the automatic pilot installation, and are connected directly to it. Three power cylinders are employed for the aileron, rudder an elevator systems in the airplane, on for each system, with direct connection by means of cables and push-pull tubes to the control surfaces.

The use of the "power boost" provides a system which is automatically irreversible at the control surfaces. This property is quite advantageous inasmuch as it eliminates the necessity of having external locks provided for the control surfaces to prevent the controls from being blown about by wind while the airplane is parked on the ground. This, however, presents no hindrance to the manual operation of the controls by the pilot from the cockpit.

This article was originally published in the March, 1940, issue of Aviation magazine, vol 39, no 3, pp 46-50, 53.
The original article includes a small photo of one of the authors with a colleague, 6 detail photos, and 4 drawings.
Photos are not credited, but are probably from Curtiss-Wright.