The Martin XB-51

The XB-51: A Work in Progress

Two “Flying Cigars”, please…

The XB-51 was built by the Martin Aircraft Company for the US Air Force based on its 1947 proposal.  The proposed aircraft would weigh 52,000 pounds, was to be powered by three jet engines, have a variable-incidence wing and tail, a revolving bomb bay door, and bicycle-type landing gear (similar to that which was used for the Boeing B-47, and later on the B-52).  The Air Force ordered two of the aircraft, which became known as the “flying cigar” due to the plane’s elongated fuselage.


Three General Electric J-47 Turbojets were used to power the aircraft. Engines 1 and 3 were housed in separate nacelles, mounted on either side of the fuselage to the rear of the cockpit.  Engine 2 was buried in the tail of the plane.  While the tail was a more conventional engine location, the intake for that engine was unusually placed at the top of the fuselage, and was blended into the root of the vertical stabilizer.

The bicycle landing gear was offset by retractable outriggers, mounted within the underside of the wing tips, which were used to provide balance while on the ground.

Martin XB-51

The rotating bomb bay was a Martin signature.  Rather than the more common bomb bay door configuration, Martin decided upon a rotating door to prevent unwanted turbulence and drag while the bomb bay was open.  In addition, the design and construction of the rotating bay door allowed for bombs to be loaded more-or-less as pallets, which would assist with hastening turn-around times for the bomber between sorties.  It was this type of bomb bay that was implemented during production of the Martin B-57 Canberra, which Martin build under license from English-Electric, in England.


It isn’t clearly visible in most of the photos for this plane, but it was designed to employ eight 20mm cannons.  The cannons were all housed in the nose of the plane and carried 120 rounds each.


Maximum speed at “combat altitude” (around 40,000 feet) was approximately 560 MPH, depending on its combat weight.  Its combat radius was about 460 miles, with a range of around 930 miles depending on its weight, altitude, etc.  There was no provision for in-flight refueling.  Despite its high speed, its ability to carry 10,400 pounds of bombs, and the heavy-hitting octet of 20mm cannons, the plane had limited ability to loiter in a combat area and relatively “short legs” regarding its range.  The Air Force ended-up selecting the B-57 and not the XB-51, despite its many innovative and unique design features.

The 3D Model

Making the 3D model of the XB-51 began with my attempts to find suitable orthographic drawings of the plane.  I ended up using the drawings that were available in the book “Air Force Legends Number 201, The Martin XB-51”, by Scott Libis.  These were the most accurate that I could find.

The first thing that had to be done was a good, high-res scan of the top, side, and front elevations.  Then it was on to modeling the plane, which wasn’t nearly as difficult for me as modeling the Lockheed L-133.  The fuselage, wings, and horizontal stabilizer were relatively straight forward to make…



The fuselage shape was a bit odd.  Study the pictures of the XB-51 and you’ll soon note that, while the front and rear of the plane begin rounded, the middle of the plane is quite a bit more square-shouldered.  The nose is blunt, not pointy like most jet aircraft of that period.  The rounded, blunt shape of the nose blended back toward the more squared-off fuselage, which then tapered back toward the round exhaust of Engine #2.  This took a lot of eye-balling, and comparing images of the plane from different angles with the orthographic drawings.

The cockpit was an extrusion from the fuselage, as was the intake for engine #2 and the blended shape of the vertical stabilizer.  The intake for engine #2 was a negative extrusion into the rear of the fuselage and toward the rear exhaust.  The book’s orthographic drawings were helpful since the side view drawing was also an internal arrangement drawing, which showed the path of the intake toward engine #2.  The engine pods were basic cylinders that were shaped to size, and then given intake and exhaust areas via negative extrusions.  The support pylons were also extrusions from the main body of the engine nacelles.  The nacelles were both rotated until the nacelle struts joined with the fuselage at what appeared to be the correct angle.  Neither of the nacelles were permanently joined to the fuselage because it was easier for me to work on the plane that way.  The horizontal stabilizer was made as a separate piece also.  It could have been (should have been?) an extrusion from the vertical stabilizer, but I thought it would be easier to create separately.  I was also thinking about texture mapping, and thought making it separate would be easier.


XB-51 (© John Matthews)



The canopy framing was cut, and then the canopy was removed from the fuselage making that a separate piece of the model…



This also left a nifty hole in the fuselage, which I closed and extruded into the fuselage to create a cockpit area where, at some point, our daring pilot figure (if I ever finish one) will sit.

Interior 1

The Libis book had a TON of great photos of the cockpit panels, switch and knob placements, markings, etc.

Main Cockpit Control Panel Diffuse

Here’s the UV map for the front panel in the cockpit.  The instruments are actual photos of (mostly) period-correct gauges.  I tried to keep it as accurate as I could.  However, at any distance where you can see the whole plane, these won’t really be very visible anyway.  Neither will the side panels, with all their switches and knobs and various markings.

The ejection seat isn’t correct for this plane.  I’m using the ejection seat from my L-133 model for now.  Martin actually built their own ejection seat for use with the XB-51.  Once I complete modeling that, I’ll substitute it for the “placeholder” seat that’s there now.  As it is though, it fits and looks like it belongs there, despite not being “correct” for the plane.

I’m also going to have to make a gun sight.  I’ve seen no photos of any gun sight mounted in the cockpit of this ship.  So, I’m planning on using the gun sight that was mounted in the F-86, which would at least be “period correct”, if not an item they might have actually ended-up using.

The wings were just rectangular slabs that I tapered and further shaped until they looked correct.  No, I didn’t consult NACA airfoil shape cross sections to get a perfect shape.  Ha!  This is correct in wing sweep, general shape, and dihedral angle.  Sometimes things have to be arrived at more artistically than anything else using the TLAR method (“That Looks About Right”).  The leading edges and wing tips required numerous cuts in polygons and creation of new edges to “pull” the shapes the way I wanted them.  The wing tips, and the tips of the horizontal stabilizers, aren’t straight or angular, but curved.  That made getting their appearance correct (or close to correct) a bit of a challenge.


It was right around this time I started getting concerned about clouds.  Making them, to be specific.  Sidetracking myself from the plane, I started getting distracted by wanting to learn how to model an environment and I began researching methods of making clouds without using plugins or buying additional software.  My results were not generally awful…



XB-51 (© John Matthews)

I rather like the image of the two-ship formation.  I decided to get back to work on the plane, and skip atmospheric modeling for the time being.  Experimenting with making clouds was tedious and time consuming, I have to admit.  I’m glad I began learning how to make them, however.  I’ll revisit making them soon, but I need to first focus on the plane.  Or, I may bit the bullet and buy additional software (like “Vue”) for environment modeling.  Why re-invent the wheel?

This plane was never painted, and was left in natural metal.  That I’m aware of, there was no camouflage paint scheme, and one can only guess as to tail markings and other insignia that may have eventually appeared on the plane had it gone into production and been flown by the Air Force.  So, bare metal…

Bare metal, if you look at it, appears to have a “grain” to it.  It’s an an-isotropic quality that I needed to learn to duplicate or replicate in the plane’s texture mapping.  I did a LOT of searches, read a LOT of web forum posts, looked at a LOT of tutorials, and then experimented.  A LOT!  I know I drove a good friend berserk with my almost non-stop talking and obsessing over it.  Ha!  I also had to learn how to use alpha channels to mask-off certain bits of the plane where markings or insignia would be painted.  Kind of basic stuff, but nothing I had used when I made the L-133 model.


I started with the wings, trying to give different parts of the wings different values of shininess/gloss and reflection.  The control surfaces were cut from the wings and made into separate objects (in the event I want to have them “move” in future images, or an animation).  I made the mistake of placing both wings on the same UV map, not doing them separately.  Why is that a mistake?  Because the anisotropic value in the shader I use only runs one way, not in two different directions.  The “grain” effect, therefore, had to be simulated using a combination of specular and gloss level maps.  I could go back and re-map each wing, but I’m leaving it as-is for the time being.  I’m actually fairly satisfied with the result right now, after twiddling with it so much.


The engine nacelles and struts were easier, despite having to re-do the UV map once.  And, after toning-down the sky a bit, the metal finish began looking more natural.  Probably still too reflective, but better.


XB-51 (© John Matthews)


I added some specular color to the engine nacelles at this point.  May have to tone that down a bit, too.  You can see the control surfaces on the horizontal stabilizers were separate from the remainder of the stabilizer.  The rudder was also cut away from the vertical stabilizer and made a separate object.  I also cut out the top window for the LORAN operator, who sat behind and below the pilot.



Well, I think that’s it for now.  I’m now working on panel lines and specular level maps.  The panel lines take time, since the fuselage is in a dozen pieces on the UV map and everything has to line-up.  So, the next up date may be awhile in coming.  Ok, so not as long as it took to publish this update, but awhile.  Until then…

Aaaaand We’re Back!

It’s definitely been a long time since I’ve posted anything here.  After the last post my schedule at work and home went a bit crazy.  Kids playing baseball, summer vacations, construction inside the house, and a constantly fluctuating work schedule made for less time working on this project and more time doing other things.  So, now work has settled down somewhat, the kids are finished playing ball for the year and are back in school, and I’ve found time to focus on this project again.

I’ve been tinkering with the metallic textures on the plane of late.  You would have thought I’d tinkered with it plenty, but after being away from it for awhile it really needed something.  There are more shader settings than Carters has little green pills (yeah, it’s an old expression… Google it… Ha!), as some of you are no doubt aware.  It’s confusing.  I guess if it were easy there wouldn’t be any art to it.

I started by ditching the grainy texture I was using in the reflection glossiness and anisotropic slots of the shader.  I played with the gloss, reflection, and anisotropic values and began to get something I was happier with…


The reflections are quite a bit more subdued and blurred now, not as detailed as they were previously.  I think that’s more appropriate for the bare metal finish of this plane than the previous, more highly reflective finish.  I still have work to do though, since there are parts of the plane I think would benefit from being glossy and chrome-like, rather than having a more brushed appearance.  All a matter of which texture map gets used where.


Yes, the starboard wing is gray (except for the flaps and spoiler/aileron).  When I’m ready, I’m going to clone and mirror the port wing, then change some of the markings and insignia, then it’ll be all metally-looking, too.  Previously the wings had been done as two wings, but one UVW map.  I wanted to apply different anisotropy direction settings for each wing (which I couldn’t really do the way I had it before).  So, I separated them and re-UVW mapped the starboard wing for its texture map.

You’ve probably noticed part of the horizontal stabilizer and the rudder are black.  That’s because all of their shader values need to be changed after I started using a not-quite-yet-dialed-in photometric lighting set-up.


I was also making it far too hard on myself trying to put markings and insignias on the plane.  Drove myself crazy with nested mix-maps and .tga file alpha masks.  Couldn’t figure out why the alpha channel in my .tga image files would cause light gray to appear in places where only black lettering should have been.  Switched over to working with .png files when there had to be an alpha mask for anything (like, all of the stenciled caution/warning items on the plane, insignias, etc.)  and viola (!), no more problems.


As you can see, there are a lot of them.  Over eight hours, over the past couple days, were spent making rivets for a bump map.  Like the panel lines, the rivets all have to line-up where the UVW maps join.  Yes, that’s a pain.  Ha!  Also, they can’t be too big, nor can the bump map settings make them appear too “deep”, or they just look goofy.  Put them too close together and they look like another panel line.  So, there’s a lot of experimentation until “happy” is reached.  I don’t think I’m quite there yet.  Here’s a tip: make panel lines and rivets in two different layers in whatever drawing app you’re using.  It’s easier that way to rid yourself of rivets while saving your panel lines.  You can always merge the two later.  Also, I can say that all of the rivets you think you may want to put into a texture will likely never be seen.  You may only want to make panel lines, and just skip the rivets entirely.  Why don’t I do that?  Too late!  I started with them and I’m going to finish with them.  Ha!


(Here’s one with a black canopy frame.  This was rendered before the canopy frame texture had been finished.)

The areas where darker toned metal is seen is the result of a map with those tones tossed into the reflection color map slot.  I’ve turned up the map output a bit, but not outrageously.


I have many markings left to place on the fuselage.  More rivets.  More tinkering with the metal shader and the lighting rig.  At some point it’ll all come together.  Thanks to all who have stopped by and looked around!  More updates to follow…