As many of you know, I have previously done some research on the Republic P-47 Thunderbolt, focusing primarily on the fuselage, cowling, and empennage. The wings, however, remain in abeyance. Accurate P-47 wing geometry is essential to a faithful understanding of the aircraft’s design, and further research will help ensure that any future work reflects its true form and proportions. Your support will make possible the continued investigation needed to document the wings with greater precision and to advance this project to its next stage.
The exact geometry of the wing, however, remains a mystery; many researchers have developed close approximations of the wing, and even Aircorps resorted to scanning. None, however, have achieved a wing that is dimensionally accurate. For the last few months, I have been reaching out to various groups, organizations, and collectors to source relevant information that will assist me with this quest. Yesterday, I received a valuable list of materials that will finally enable me to do what has evaded others.
As with everything related to historical aviation, there is a cost that is beyond my budget. That is why I am asking for your support to raise the necessary £450 to purchase the materials, and for the first time, I believe that I can produce accurate dimensional datasets for the P-47 wing layout and ribs.
My offer is simple: help me raise funds via the Buy Me A Coffee link or the PayPal link shown in the margin on the main blog page. Every donation over £10 will receive a complimentary copy of the final CAD/Ordinate datasets and layouts for the entire P-47 when completed. The final dataset will not be ready for a few months, depending on how much can be raised and how soon I can purchase the source material.
We are very close to finishing the P-39 Airacobra project at Planes of Fame. Here are the latest pictures from the restoration, along with some related 3D-printing developments for the trim tab, aileron, and rudder dials.
Many of the control panels were built from scratch from the original blueprints and developed in CAD. The Control Panel was derived from a combination of blueprint data in conjunction with the maintenance manual datasets. The actual blueprint for this model does not exist, so the various elements required extensive research to get this right.
The Gunsight mount is one area where 3D printing proved invaluable in the development of these parts. We didn’t have all the variant data for this particular model, so again we researched from known data sources and dimensional information to create an accurate replica. The mounting mechanism and plates were first 3D-printed and then test-assembled before committing to actual machine-shop fabrication.
I mentioned above the Trim Tab dials for the Rudder and Aileron. I actually 3D printed these here in the UK to determine what is viable for 3D printing and what needs to be fabricated. Ideally, a working prototype would be nice but not essential. The first set of images is the breakdown of parts from the CAD model, with the last image showing the 3D composite model.
All 3D-printed parts for this exercise were done with a 0.4mm nozzle and a layer height of 0.12mm. The model’s finish is actually surprisingly good. A smaller nozzle of 0.2 mm would permit a smaller layer height of 0.08 mm, thus yielding even better results.
Some of the smaller parts, like the small gears, were okay but not sufficiently detailed to create a working model. These could, of course, be machined and assembled in conjunction with 3d printed parts, which would, in fact, work well for prototyping. This was a demonstration of what can be accomplished with 3d printing and also outlined the limitations. It will be interesting to see what Planes of Fame decides to do to complete the Trim Tab console.
I shall post new updates as they become available. I can honestly say it has been a privilege and a challenge to work on this P-39 project.
Update 3rd Nov 2026: Control Stick Boot:
The shroud or boot around the base of the Control Stick is now a priority item.
At first, I considered a flat pattern for each section and then assembling them to form the boot, which is doable but will require a mold to align and achieve the finished shape. So why not just use the mold in the first place, place each of the material sections, temporarily sew the seams, and then turn the finished product inside out?
Pending approval from Planes of Fame, this is likely how we will proceed. You can see from the blueprint scrap views that the basic form for the boot is a cone which, when installed, is crumpled to produce the freedom of movement required. In the CAD scrap views, I have highlighted the individual panels in blue, orange, and gray. The mold will be 3d printed, and for reference, there is a small protrusion on one side, which indicates the position of the zip.
The latter position is important because the control stick is not exactly centered in the middle of the opening; the lever mechanism is. So this little protrusion will provide the visual reference to ensure we get this boot in the correct orientation. You may also have noticed a small red wire around the perimeter. The flange material is doubled over this wire to secure the flange to the clamping brackets (not shown), which we have decided to continue around the full perimeter and not just to the central shaft tunnel area. From the images and scrap views from the manuals, it is not clear whether the wire is continuous or not, so it is a choice that actually makes sense, as we do have a clamping plate on each side of the base.
This part of the project was only developed in the last 24 hours and may be subject to change following further discussion with Planes of Fame. I shall keep you updated on progress and any subsequent changes to this approach.
I haven’t posted for a while, partly due to getting too involved in the madness that is now embedded within the UK. I won’t go into that right now, as there are a lot of personal implications that are getting overly complex.
Anyway, I just received some images of the P-39 Airacobra restoration from Planes of Fame that I would like to share with you guys. As you know, I have been contributing a lot of technical data and CAD work to assist with this project, some of which is highlighted below.
P-39 Fuel Tank Covers and access ports:
P-39 Instrumentation Panels:
P-39 Gun Shrouds:
In the above image, you can also see the Air Intake, which was one of the first CAD models to be developed for this project:
A few other images for your perusal:
The cabin interiors, controls, and fitments are currently a work in progress. Some of which are being 3D printed, whilst others are being manufactured in aluminium as per original specifications. As I receive more updates, I will post them here, so watch this space.
Whilst I am here, I would like to thank those who have made donations to help me with my work. Small or large, the donation is equally gratefully received. Thank you very much.
It is not common for blueprints to be almost illegible, and without a Parts catalogue, understanding the mechanisms and operations of assemblies like Gearboxes can be challenging. This was the case with the Tail Wheel assembly I built for the P-51 Mustang and, of course, the current work in progress, Landing Gear Hand Crank Gearbox for the Grumman Goose.
I became captivated by this unique gearbox upon discovering its remarkable dual function: it not only raises and lowers the main landing gear but also manages the tail wheel’s movement. However, delving into the blueprints left me with more questions than answers regarding its intricate operation. Intrigued by its complexity, I decided to construct a working model and evaluate its operational characteristics firsthand.
The Gearbox consists of a central shaft featuring an ACME thread along which the Traveler Collar for the tail wheel moves. Additionally, it includes a bevel gear that powers the main landing gear struts, as illustrated. At the base, the ratchet lock offers two positions: one for raising and the other for lowering the landing gear.
I am eager to explore the operational parameters and the criteria for calibrating this gearbox to ensure smooth operation and timing. The available blueprints and installation manuals do not clearly outline how this setup is configured, so I will need to rely on some trial and error.
To successfully complete this assembly, we still need to finalise several crucial details, particularly the assortment of nuts, bolts, and washers. Fortunately, I have access to an extensive library of parametric parts, ensuring that I can efficiently source the exact specifications required for this project.
Developing these assemblies requires a significant investment of time and effort, but I believe this investment is invaluable. Often, manufacturers’ documentation is either unclear, incomplete, or entirely absent, which can create challenges for maintenance and operational staff. By constructing detailed CAD assemblies, we create a visual representation that not only clarifies the intricacies of the components but also serves as a critical resource in the field. This practice can facilitate more efficient troubleshooting, enhance understanding of the system’s functionality, and ultimately improve the overall safety and effectiveness of operations. By proactively addressing these documentation gaps, we ensure that maintenance teams are better equipped to perform their tasks with confidence and precision.
In previous articles, I shared my aspirations to develop a 1/16th scale RC model based on this project. I realised that this gearbox configuration could serve as inspiration for creating a scaled version that would operate using a single servo to raise and lower the model’s main landing gear and tail wheel.
Update: 28th Jan 2026: Spur Gears
The Spur Gears and Splines dimensions are shown as “over pins”, the diameter of which are 0.140 in.
CAD software generally does not facilitate this type of dimensioning for gears, so first we have to determine the important gear parameters using online calculators like this one at Zakgear.com:
The Diametral Pitch is 12 (number of teeth/pitch diameter), which we then input into the CAD gear calculator. To match the calculated diameters from the Zakgear website, we need to adjust the Addendum to 0.800.
By overlaying the CAD data onto the Zakgear data, we achieve a good match. It may only require microdimensional adjustments within stated tolerances to ensure perfect alignment for a correct setup.
Restoration Insights: The Risks of Working from Blueprints
Restoration projects…is working directly from blueprints a good idea?
A company I know is currently restoring a P-40N aircraft, and I came across several posts where they highlighted concerns about the alignment of the fuselage frames. The misalignment was approximately 1/8 inch (3.175 mm), which is quite significant. From their posts, it seems they are working directly from the blueprints.
Throughout my experience in the industry, I have encountered occasional dimensional errors in the blueprints of nearly every project I have been involved in. This recurring issue fuels my passion for my work. I strongly believe that dedicating time to meticulously developing these designs in CAD is essential for uncovering any anomalies before fabrication begins. This proactive approach not only enhances the accuracy of the final product but also ensures a smoother assembly process. However, I recognise that this level of diligence may not always be feasible due to various constraints.
For example, if you are building the fuselage frames and one of those is 3mm out of alignment, you naturally assume that it is incorrect. That may not always be the case because, as the assembly progresses, there may be factors that are as yet unclear that influence this misalignment, or it could simply be a mistake. You won’t know for sure until all the parts are assembled.
Consider for a moment the following example from the Grumman Goose Tail Wheel blueprints.
I have intentionally highlighted the revision box to indicate Revision H. This revision specifically documents the change in dimension from 6.5 inches to 6.25 inches. If we examine the other dimensions, the blueprint specifies that the centre axis for the fork should be set at a 45-degree angle. Additionally, the key setting out dimension is 5.25 inches, measured horizontally to the intersection of the vertical axis and the centre of a 1.25-inch radius.
This immediately rings an alarm bell…to achieve a 45 degree fork with the dimensions shown, you would expect that 6.5 inches is in fact correct and that in this case the 6.25 inch is not. But yet it was the only purpose in this revision to record a change to 6.25 inches.
The tilde “~” indicates that this dimension is approximate, but for this to be a revision would suggest that the actual dimension is closer to 6.25 inches than it is to 6.5 inches.
To ensure all key dimensions align with the blueprint, particularly noting that the 6.25-inch measurement is approximate, the setout for the Tailwheel Fork should follow the above depiction. However, we now have a concern: the vertical post is meant to extend to the diagonal intersection and be welded to the curved plate’s interior. As shown in Detail B, the edge of the posts is too close to the fork’s edge, while the blueprint indicates they should be positioned further inward. Additionally, the actual component, seen in the following screenshot, reveals that the heel of the fork is more bulbous than the blueprints suggest.
There was a reason for the 6.25-inch revision, though we do not know it at this time. Therefore, in order for this to be correct and meet all criteria, something other than the 6.25-inch dimension should change.
Honestly, I’m not sure what the correct answer is here. Unless I can physically get my hands on the real thing, this will likely remain a conundrum. I will retain the CAD design as it is for now, which serves my intended purpose to demonstrate the deployment parameters of the Tail Wheel and provide clarity on the assembly configuration.
I recognize that the dimensions in most blueprints are generally accurate, with only a few exceptions. When budgets and schedules are tight, it may not be practical to explore entire assemblies in CAD before fabrication. However, in cases where discrepancies are identified, I recommend examining all relevant assembly components in CAD. This will help in identifying the correct solution and understanding all influencing factors before making any changes.
For quick renderings that are perfect for blog posts, I typically prefer KeyShot. It provides an intuitive workflow and a large library of environments and materials. However, the trial version has some limitations: you cannot save projects or export a rendered image, except as a screenshot. When I was recently asked to produce high-quality renders of the Republic JB2 for a museum display, I was uncertain about how to accomplish this.
These products are very expensive and far exceed my budget, so I urgently needed to find a solution. That’s when I discovered Autodesk VRED. I downloaded the software along with the accompanying asset library, and to my surprise, the trial version is fully functional. It allows me to save projects and create high-resolution renders, and it runs for 30 days.
Autodesk Vred retails at around $14000, which is extraordinarily expensive, but it is aimed primarily at the Automotive industry. Consequently, the product is packed full of features and limitless options on environments, materials, lighting and camera setups. It truly is a comprehensive and, to some degree, rather complex product, so there is a steep learning curve.
Undeterred, I set to work by reviewing tutorials, YouTube videos, and various online resources. Over the course of six days, I gained a deeper understanding of the nuances of VRED rendering. While I’m not an expert yet, the test renders started to come together, culminating in the images showcased below.
These images are not final, as I still need to work on the texture mapping and apply materials to some internal components. However, they demonstrate that it is possible to achieve satisfactory results in a relatively short time. Although the product has a steep learning curve, it encourages you to deepen your understanding of materials, textures, and lighting, which ultimately enhances your grasp of rendering processes.
I highly recommend that anyone interested in creating renders try Autodesk VRED. It offers the full functionality of a high-end rendering product, including the ability to save your projects and export high-resolution renders. The availability of a 30-day trial version is exceptional—Keyshot, take note!
I want to clarify that I have no affiliation with Autodesk, but when it comes to the accessibility of professional products, Autodesk is unparalleled. I have no problems recommending worthwhile products, like this one.
The CAD/Ordinate datasets are designed to offer detailed documentation of the dimensional information pertaining to the core profiles of various aircraft components. This includes elements such as fuselage bulkheads, cowls, vertical stabilizers, horizontal stabilizers, wings, rudders, flaps, ailerons, and elevators. Essentially, these datasets provide all the dimensional information needed to develop the main profiles for aircraft construction.
The research studies were conducted to fill in important gaps in information and to clarify unclear details. Often, data on blueprints can be difficult to read, making it necessary to record and analyze the bulkhead or rib profiles in CAD. This process helps accurately determine the correct dimensions.
The examples of ordinate dimensions above are not necessarily the worst; in fact, there are truly poor examples that exist. To tackle these issues, we should start by recording the known dimensions in Excel and making educated guesses about the worst examples. Next, we can create each profile in CAD. This CAD profile will give us a clear visual representation of any anomalies in the curvature, which can be further analyzed through curvature analysis to identify low and high spots. This process is done for every rib and bulkhead profile where we have ordinate dimensions.
The spreadsheets above are typical examples of CAD/Ordinate datasets. The first spreadsheet contains the Ordinate record for the P-38, while the second one features the Aileron sheet for the FM2. You may notice a Linear Regression analysis table included in the FM2 sheet. Initially, determining the individual profiles of the ribs or bulkheads is just the first step; we now need to assess the assembly of all these components and check for proper alignment.
Each drawn sketch profile in CAD will serve as the border for containing a surface patch.
There are two primary reasons for doing this. First, it provides us with a plane that can be converted into a working surface, which can be utilized in any CAD product. Secondly, it provides us with a tangible element that we use to check assembly cross sections at key locations for alignment checks.
For example, consider the wing of the FM2. The wing assembly has been converted into a part file, and cross-section sketches were created at various chord locations: 30%, 60%, 70%, and 80%. Each sketch utilized the “Project Cut Edges” function to generate a cross-section of each rib. As shown in the second image, the array of lines representing the rib cross-sections provides a visual aid to identify high and low spots on the wing assembly. By creating a surface plane for each rib, we were able to generate these cross sections effectively. There were a few high and low points, which were double-checked and rectified.
If we require additional verification and strive for precision, we could use Excel’s Linear Regression to generate the coordinates for a Best Fit Line and make adjustments as needed. However, this approach may be excessive since our primary goal is to clarify the original blueprint data and apply it to identify appropriate rib and bulkhead profiles within acceptable parameters.
We can also use Linear Regression to give us an overview of how the ordinate profiles align with one another and to identify any discrepancies. Typically, acceptable parameters are within +/- 0.01 inches (or 0.254 mm), as specified by the dimensions on the blueprints, which usually only provide accuracy to two decimal places. Sometimes, as was the case with the P-51 and P-38, we had key design parameters that allowed us to calculate the exact profiles for each wing.
Validating dimensional data is crucial because the actual wing construction may not always match the accepted specifications. The design specifications for the FM2 call for a NACA 23015 airfoil at the root and a NACA 23009 airfoil at the tip. You might be surprised to learn that the NACA 23009 is a modified version of the standard 23009. Nothing is therefore assumed or taken for granted.
The CAD/Ordinate datasets are the result of extensive and thorough research and analysis, often taking many months of work, sometimes around the clock. These spreadsheets include every known ordinate dimension for various aircraft, gathered not only from blueprints but also from manuals, reports, and even correspondence. The CAD/Ordinate packages also include various 3D CAD models in various formats, including 3D DWG and fully dimensioned 2D DWG. All documents provided are fully editable so you can adapt the information to your work processes.
For more details on using the Ordinate spreadsheet data for your own CAD systems, see my earlier post here: Ordinate Overview
With over 45 years of experience in structural and mechanical engineering, my expertise influences everything I do.
In summary, the purpose of the CAD/ordinate datasets is the result of intensive work and research to provide the user with correct usable data that can be utilized in any CAD system.
When you buy CAD/Ordinate datasets and Blueprint collections from me, you support my ongoing research to provide the most comprehensive and probably the most accurate dimensional information about various aircraft. This blog and my research work would not be possible without your support.
The Grumman Goose project is both challenging and frustrating; it is definitely not a straightforward aircraft to work on. I have primarily focused on updating the empennage, which includes the vertical stabilizer, horizontal stabilizer, rudder, and elevator. During the development of the ordinate study, I observed discrepancies in the documented locations of various components. Let me explain what I mean.
Upon reviewing the CAD drawings on the left and comparing them with the Maintenance Manual diagram, I noticed that the level of the ribs varies by 1/16 of an inch. This discrepancy caused me immediate concern, and I began to wonder where I might have misinterpreted the Grumman drawing data. Therefore, I felt it was necessary to review and verify the information.
Initially, we do not have any reference location information on the Rudder Layout drawing. Normally, you would expect reference dimensions to the fuselage centerline or a fuselage station reference, but there are none. We do, though, have locations of the Hinges on other drawings for the Station bulkheads and Fin layout which in turn will help derive location information for the Rudder.
The first image above is the bulkhead layout at Station 36, which specifies the centre of the hinges 1, 2, 3, and 4 relative to the Fuselage Ref Line.. The second image is the bulkhead at Station 33, which shows the dimension of 65 13/16″ to the top of the Lower Rib on the Vertical stabilizer Fin.
I am looking to verify the dimensions and locations of the rudder ribs and hinges in relation to the Fuselage Reference Line. To accomplish this, we will start with the information we have and determine what additional information we need. The first image confirms that the CAD drawings for the rudder accurately depict the positions of the hinges. The second drawing further supports this; the “Top of Rib” location refers to the lower rib of the fin which includes the locations of the hinge centers. At this point, we have established the correct locations of the rudder hinges from two different sources.
Having determined the hinge locations, we know that the ribs for the rudder are offset by 5/8″ on either side of those locations, which allows us to derive the final levels noted on the Rudder Layout CAD drawing. Does this mean that the Grumman drawings, and therefore the CAD drawings, are correct while the manuals are incorrect? Yes and No…let me explain…
The first image is the Lines Diagram for the Vertical Stabilizer Fin Ribs. In the Table of Offsets, you will notice a list of dimensions from the “Root,” with the first rib specified at 10 7/8 inches. If we overlay these dimensions onto the CAD drawing, we observe a 1/16-inch discrepancy to the top of the first rib. However, all other sources, including those mentioned above and additional references not listed, such as the fuselage Lines layout, indicate that the top of the rib is correctly positioned in the CAD model (second image), contradicting the information provided in this Table of Offsets.
So what is going on?
We should take into account the revision history of the Grumman Goose development. If you examine their drawings, you’ll notice that they have made numerous revisions, some of which are labeled with letters as late in the alphabet as “R.” That indicates a significant number of changes.
I believe that various details have changed over the year, with the more prominent aspects being updated while the less prominent drawings remain unchanged. Regarding the manuals, it seems they were created early in the project, and it may have been considered too labor-intensive to update the level references. This aircraft is quite complex, and I can only imagine the effort involved in both its development and the ongoing updates to its design.
Whenever a small anomaly becomes apparent, I will make an effort to gather information from other drawings to verify the final result. This is one reason why these Odinate studies take so much time; it is crucial to ensure that the final study represents the most accurate dataset possible. If I were building a Grumman Goose replica, I would be using my datasets.
Progress Update 18th March:
A few screen shots showing the latest updates to the JRF Goose. The wing has been completely rebuilt with all dimensions verified.
Working on the controls and instruments for the P-39 spawned a plethora of questions about how these controls actually worked. So I endeavored to incorporate the inner workings in the Trim Tab Control CAD models. This was specifically to get a better understanding of how they work. This was not a mandated requirement. The initial work scope was replicating the external components for a static display P-39 restoration.
Often enough in museums and private collections, we only see the external controls. For many, this is all they want to see. But what if we also see the internal gears, pulleys, shafts, and bearings to understand how they operate? This is exactly where I now want to go with my future projects.
The Trim Tab controls for the Elevator, Rudder and Aileron are already modelled for the P-39 including the internal components. These dials and controls are currently being manufactured for the restoration project. The decision has now been made to incorporate the working mechanisms as functional replicas. This is great and will actually have some form of function, however, the mystery of operation still eludes the operator. I want to take this a step further and produce desktop models with Clearview casings so that the internals are visible. The exact method is still under review. It will mainly comprise 3D printing techniques for the main components attached to perspex casings.
The dials for all 3 controls are similar with the Rudder and Aileron dials operated by a control knob (not shown) and the Elevator Tab controlled by a wheel as shown. At the base of each control dial there is a sprocket for a short Roller Chain which in turn is attached to operating cables. Out of curiosity I decided to have a look at other aircraft to see how alternative mechanisms were developed for the P-51 and the FM2.
For the P-51 the Trim tab controls are comparable in their operation with the internal gearing arrangements but differ slightly in design.
The dials for the Aileron, Elevator and Rudder are all similar to the CAD model shown. The Elevator and Rudder have cable drums attached to a long shaft for direct cable operation whilst the Aileron has a chain sprocket similar to the P-39 Trim Tab controls.
The plan for the P-51 is to fully model all the components in the assembly shown, complete with cables and chains to simulate operation.
A small point of interest; the various aircraft designed by the same manufacturer often share common parts; for example the NAA drawings for the B-25 share the same Trim Tab control knobs as the P-51 and listed accordingly. For some reason, the P-51 drawings do not reciprocate.
If you can’t find drawings for a particular part, check collections for other aircraft by the same manufacturer. Occasionally, this can be worthwhile. Similarly, with Grumman, many parts were shared with the FM2 and the Grumman Goose.
The above model is the FM2 Elevator Trim tab control, the main body of which is typical for the Aileron and Rudder on Grumman drawing 13690. The Grumman Goose has similar controls shown on the Grumman Drawing 13693. Shared components across the various aircraft are listed on the Grumman FM2 drawings.
This Trim Tab control for the FM2 is probably the most complex I have studied so far…requiring very fine manufacturing tolerances. I am not entirely sure yet how this works as there is a complex array of tabbed washers that act as stops for the dial in both directions; it is unclear at this stage how they should be configured…I will get it worked out in due course.
A lot of work to do on these projects which will definitely keep me busy through 2025.
During the development cycle of any aircraft the manufacturing standards tend to evolve and commonly change content, description and name. Keeping track of those changes is key to ensuring the defined parts are correct for the assembly of components.
Alongside the P-39 Restoration project, I still develop several aircraft components for other aircraft. This includes my old favourite, the P-51 Mustang. This example refers to the Quadrant Assembly for Engine Control; drawing #102-43005 for the early P-51B.
Many of the standard parts called up on this assembly use the early standard references. These are typically prefixed with “B”, like B1009, B1135, etc. These standards were later updated, and a new series of standards replaced them. I have correlated these in a spreadsheet, as shown below.
The spreadsheet only lists the standards that are available in the blueprint archive and may not be a complete record. For a more comprehensive listing, refer to the Erection and Maintenance instructions for the P-51A series (T.O. No. 01-60JC-2). Check the Conversion lists from pages 404 onward.
The CAD development of the Quadrant Engine Control is still a work in progress…created to the exact original dimensions.
My plan for the P-51 is to further develop the instrumentation for the early and later versions.
P-39 Update and Standards:
Over the last few months, the P-39N Restoration project has been my primary focus. We are close to completing the CAD work for the Cockpit instrumentation.
By contrast to the P-51 Mustang, we don’t have the same collection of Bell Standard part blueprints. We only have what is available within the manuals. However, the notations for the parts are similar for the industry as a whole. For example, a Spacer Part noted as Q065-6-20 shares the same notation for the dash numbers as the P-51 (standards 4s3 and 4S4). This in turn will define the spacer size…the first dash number sequence indicates the bolt size and the second is the length. In this case, it would be #6 bolt size. The length is defined in 1/32nd inch, making it 20/32″ (5/8″) long. This table derived from the iPart feature explains the designations in more detail.
The next time you come across a part reference you are unsure about, cross-reference it with other aircraft-known standards. Also, consult the comprehensive collection of AN and MS standards on Everyspec.com.