Tips and tricks from an idiot savant

[ErickC]

Every once in a while, I come across a piece of information that model builders might find useful. This does me absolutely no good, because I really don't spend much time building 3D models these days, so I thought it might be a good idea to share what little wisdom I might impart somewhere. I figured this is as good a place as any. From time to time, I will try to post little tips and tricks I've managed to come up with, as well as tidbits that will make it easier to understand how some specific airplanes work. The benefit to spending a decade or so in factories, and a great deal of that time as a machine operator, is that I am well-versed in how to read blueprints, and how machines are actually assembled, maintained, and operated. It is my hope that some of that mechanical know-how may be of some benefit to the reader.

My first tip is a simple one. Ronnie Darko had a simple rule when he was president, and it's as potent to the model builder as it is to the commander-in-chief:

Trust, but verify.

Treat all reference materials that are not manufacturer's blueprints with scrutiny. Be especially suspect of anything in operations or maintenance manuals. These diagrams are meant to give the operator a good idea of what things look like. They are not necessarily accurate. The station diagrams found in the 737-200 maintenance manual, for example, provide an excellent resource if you need to know the X or Y values of particular points (i.e. the floor, window centreline, lobe crease, positions of stations along the longitudinal axis, position of the wing, and so on). Don't trust them for the contours of the airplane for a second. The resolution ain't that great. But if you use them in conjunction with the very accurate CAD drawings available on Boeing's website (free!), you'll be on your way to a very accurate model.

On the subject of 737 station diagrams - Boeing, unlike Convair or Douglas, is lazy. If you use the Y-values to build your model, you will end up with a 737-100 unless you were paying attention. Note stations 500, 500A, 500B, 727, 727A, and 727B. Boeing didin't give new coordinates, they just added the new stations in and put an "A" or "B" to denote that they're new. You will have to adjust the Y-values for every body station past 500, and again past station 727. All station values on Boeing, Douglas, Lockheed, and Convair diagrams are in inches. Airbus diagrams will be in metres. The others make it easy for you and just give you the correct values. Beware!

Boeing 737-200 body station diagram (from AMM)​

Oh, yeah, don't trust blueprints, either. You'd be amazed at how many changes that occur in the factory don't show up on blueprints (that was my experience with manufacturing buses, anyway). Double check every source with other sources, such as photographs, measurements, and so on. If you're in doubt, and you're building a 737, pick up a DACO model. They were made from measurements of Sabena 737s and are very accurate - but remember that scaling up from 1/144 will produce a certain degree of inaccuracy in your measurements.

 

[ErickC]

Tidbit #2 - here's a section of the 737 station diagram again. The little symbol with the "C" intersecting an "L" means "centreline." This means that the center of the windows is at WL (water line) 246.1:

Yes, even bus, car, and airplane blueprints have vertical stations denoted as "water lines." Other abbreviations you might find useful are:

LBL (left buttock line) - distance to left of centreline
RBL (right buttock line) - distance to right of centreline
STA (station) - distance back from origin

Note that the tippy tip of the 70237 nose is at STA 130. The origin is at an arbitrary point, as it exists solely for reference. I think the DC980909517 nose starts at station 80, but I could be wrong.

 

[ErickC]

Here's a more substantial one for you poor saps working on the 737. We all know the early 737 wing is about the worst thing you'll ever build, so here are some tips to make it easier.

OUTBOARD FLAP DRIVE

This is the worst part of the process, and most model builders fail to grasp how this all moves, which not only produces inaccurate results, but actually makes the job harder. Let's learn some things!

This is the outboard flap transmission. If you're mechanically inclined, you'll see why almost everyone animates them wrong. If you're not, here's an annotated version:

If there is one thing that I have to impress on you, that you need to have drilled into your head, that I will never forgive you for forgetting because it a.) has been pointed out to you and b.) makes your job easier, it's this:

THE FOREFLAP ONLY EVER TRANSLATES OR ROTATES, IT DOES NOT DO BOTH AT ONCE, AND IT STOPS MOVING LONG BEFORE THE MID AND AFT FLAPS DO.

If you don't believe me, have a video.

If you still don't believe me, here's another one.

This is actually necessary for the midflap to translate correctly. In neon green, I have highlighted a linkage that runs between the foreflap carriage (red) and the midflap carriage (dark blue). As the midflap carriage runs back on its track, the track forces the back end down, while the linkage rotates up and back to allow the leading edge of the carriage to move back and up. This gives the midflap a slightly greater angle. As the midflap rotates down, it pulls on the foreflap guide rail, which causes the foreflap to rotate downward. Notice that the gap between the foreflap and the midflap is huge compared to the midflap and the aftflap. Treat the foreflap and mid/aft flap as two separate, but dependent assemblies. This makes your job easier because all the foreflap has to do is move back, then rotate down. It can be completely separate from the rest of the assembly in your model's hierarchy.

The aftflap, by weight of comparison, begins to move along its track immediately. This is because the midflap carriage exerts force upon the linkages outlined in pink, which are connected to the bellcrank (brown), flap track fairing, and aftflap pushrod (dark green). This causes the bottom of the bellcrank to roll along the bellcrank cam track (yellow), which forces the flap track fairing down, and the combination of forces pushes on the aftflap pushrod, thereby making the aftflap move along its track. You don't have to model all this, because it's all invisible anyway, and, so long as you understand how it moves, you can approximate it. But now you understand how it works, and can model the motion faithfully, even if you don't model all of the parts.

So what makes this all move? Hydraulics? Yes. Sort of. The flaps are powered by hydraulic motors located in the main gear bay. These motors transmit their rotational force via shafts that run along the aft main spar, just below the spoilers. At each flap track fairing is a gearbox which transmits the motion to a jackscrew that is connected to a small mount on the midflap, outboard of the carriage. You can see it very clearly in the first video, as that 737 has white paint on the stop at the end of the screw. The system can also be electrically actuated in an emergency. What does this system look like? Like this:

But this mechanism must be at some odd angle, right? Otherwise the whole system would bind up, right? Wrong. The flap tracks are parallel to the fuselage, set at a slight downward angle to allow the flaps to move down as they move back. The reason the whole thing works is that the midflap is connected to its carriage by a spherical bearing that allows it to rotate. I suppose there's a way to model this with bones, maybe, but I just linked the midflap to the outboard carriage and made sure to rotate it to keep it in line with the inboard one as the flaps moved.

Note that the leaf spring is what holds the trailing edge of the flap up. The carriage wheels are on the inside of the track.

INBOARD FLAP DRIVE

The inboard flap drive is largely similar to the outboard drive, but the tracks are located in the bullet fairing (inboard) and engine pylon (outboard). This is why the New Generation/Classic 737 kept the Classic/Original assembly, even though the change to the CFM56 negated the need for a big, long pylon. They didn't change this until the Next Generation (I call the 737-8 the -800 Advanced, but we could probably call the whole family Deep Space 9, I think?). On that note, marketing people are annoying. I side with Bill Hicks and the exhaust pipe on that one.

The inboard flaps use cables to move the aftflap, but this assembly is largely hidden, so I won't really bother with it. Animate them the same as the outboard flaps and you're golden. Don't pull a Captain Sim and forget the vane flap that covers the inboard track on the bullet fairing when the flaps are retracted. It connects to the midflap and is an obvious part.

SLATS

Here's another thing I need to drill into your heads:

AT FULL EXTENSION, THERE WILL ALWAYS BE A MISMATCH BETWEEN THE OUTBOARD SLATS AND THE REST, UNLESS YOU ARE BUILDING AN NG OR MAX SERIES AIRCRAFT.

This mismatch depends on the series. Let's examine the LED indicator on the overhead for some insight:

Note that the slats 2, 3, 4, and 5 only have one position on the -100 and -200, but two on the -200 Advanced, -300, -400, and -500. This is your clue as to the nature of the mismatch. Slats 1 and 6 always move to the exact same angle, regardless of whether it's a 737-130 or a 737-436. What changed is that on the -100 and -200, the inboard slats don't move as far as the outboard slats, and on the rest, the inboard slats move just a little bit further. It's not a small mismatch - from here on out, I guarantee you'll always see it. There are photos of -3/4/500 series airplanes where the mismatch appears absent. This is because it appears that it's absent at the first position on those aircraft. If this is how the wing came from the factory (be mindful of the available retrofit jobs for these aircraft that re-sequence the high lift devices), this is different from the -200 Advanced (the mismatch is proportional to the percentage of extension, so it's hard to see at the first position, but still there). If you have a good photo of the airplane with the slats fully extended, and it's not an NG or Max, I guarantee there will be a mismatch. Note also that the 737-3/4/500 slats are longer in chord.

SLAT DRIVE

Each slat has at least these things:

-A drive piston
-A telescoping anti-ice duct
-Two curved guide rails (outboard from slat center)
-Two straight-ish auxiliary guide rails (inboard from slat center)

The #1 and #6 slats have an additional auxiliary rail, and it's shaped differently from the others (it has a wee kink in the middle!). The covers that go over the rails and the anti-ice ducts are part of the slat on the -100, -200, and -200 Advanced. On the -3/4/500, the anti-ice duct has a door, just like the actuator piston. It's the same size and shape of the covering on the others, but it's attached to the wing and hinged at the back. This diagram should illustrate most of what I'm on about:

These insights should prove helpful for anyone attacking this Sisyphean task.

Last, but not least: for anyone who might argue that these are nitpicky details, I would argue that they're at least as important as opening radomes or engine cowlings. Use your polys wisely!

 

[Heretic]

Not sure when this will come in handy, but thanks Erick.

 

[Ronald]

Great material and tips EricC. Thanks for sharing your wisdom.

 

[gapeters]

Have always loved the added details of Ericks models and this is a great insight into how he tackled them. Your right though, eye candy sells more than canoe flap tracks and correct flap modeling.. Looking forward to the insight on modeling cascade style engine reversers!

Greg

 

[ErickC]

You're more than welcome. Many of these things apply to the 727 as well, as its drive mechanisms are largely similar.

Another couple of things:

-The last 737-200 was airframe #279. Airframe #280, delivered to ANA (as JA8412), was the first -200 Advanced.
-In 1973, the blow-in doors were eliminated. This means that, depending on the timeframe and the airframe, they may have been present on any airplane, including the -200 Advanced. My research indicates that the first airframe without them from the factory was #324. Few aircraft retained them, but I have seen recent photos of an Iraqi -200 Advanced that still has them. Check your photos!
-The aircraft often cited as the -200 prototype (with the red and white scheme and the N-number that doesn't actually exist - N4560V I think) is not even a -200. It's a -200 Advanced, and nobody seems to have any clue as to what airframe it actually is. At any rate, the 737-200 prototype was N9001U, and came out of the factory in United colours (albeit with a photo-reference mark on the CG like the first 737-100, a big "5" on sides of the radome, a circle on the front/bottom of the radome [presumably red or black], and "EXPERIMENTAL" titles below the cheatline and aft of the main entry door ).
-All 737s built before sometime in 1969 had short nacelles and the 727 thrust reverser system. All of these aircraft got the new HPTR (adapted by Rohr from the DC-9) at no cost.

 

[ErickC]

Have always loved the added details of Ericks models and this is a great insight into how he tackled them. Your right though, eye candy sells more than canoe flap tracks and correct flap modeling.. Looking forward to the insight on modeling cascade style engine reversers!

Greg

Cascade reversers aren't bad if you remember how they work. The translation ring creates an opening to expose the vanes, and pulls the ring segments out via drag links:

This photo is of Delta's first 767 (preserved at their museum in Atlanta) and shows it fairly well (or at least my intent was to show it fairly well). There are some good videos on YouTube (a search for CFM reverser ought to get you going).

Another tip: Concourse D at KATL has both a Chipotle and a Five Guys, and this is awesome. Also, the $12.50 it takes to get in the Delta Flight Museum is well worth it (they use 757 nose gear assemblies as building columns, ain't that great?). Just check in with security at their headquarters (on Delta drive, no less), take a left at the parking lot, and walk around to the other side of the big white hangar.

 

[ErickC]

Here's a big one inspired by another thread: how to make cockpit windows the easy way.

This method is best suited to older, manly airplanes with windows made of (mostly) flat panes of glass. I call it "HILARIOUSLY OVERSIZED POLYGON." It will create perfectly flat panels with a minimum of time, effort, and resources. Let's go!

WARNING - BIG PICTURE ALERT!

I'm just freelancing this one, but it's based roughly on the DC-9. I started with a simple cube.

I deleted all but one polygon and slanted it to the angle I wanted for the center pane:

Select this polygon. Clock "move." Hold shift. Drag it to a convenient position to clone at the sub-object level:

Snap those bad boys together in vertex mode. You can weld if you want, I didn't bother (lots of welding will follow later, so it's lazier to do it in one final step).

Move the outer vertices to the new position. This is where the oversized bit comes in - you want the polygon to cover the entire area of the window, and it will need to be cut down in the next step.

The minimum to do this accurately is two out of three: side, top, front. I based this on the DC-9, which means the rear of this panel slopes outward viewed from the front, and up and aft from the side. I always cut in editable mesh mode. E-poly is a whiner about edge cutting.

Delete the chaff.

Select the aftmost vertices. Clone at object level again. Move to cover aft extremities of the next pane.

Clone the four vertices at the edges, snap them to the bottom edge.

We now have the making of yet another HILARIOUSLY OVERSIZED POLYGON.

Continued in part two!

 

[ErickC]

PART TWO.

This hilariously oversized polygon only used four vertices to create an angly pane of glass, but depending on your needs, you might want to use five and make a nice, big square one. At any rate, let's cut it down:

Repeat process for aftmost frame:

I didn't make a shot of the removal of the unnecessary parts. Let's move on to making the window. I started by extruding a hilariously large amount, just to get the extruded polygon out of the way:

Snap it back from whence it came. Do not weld.

Uniform scale it to size:

Extrude inward to taste and delete the (now unnecessary) polygon:

Convert back to E-poly since E-mesh is whiny about chamfering. Chamfer edges as required.

Clean your work. Apply smoothing to taste.

 

[ErickC]

Now create a new polygon and detach it. I'd normally do all the openings first, then create the glass last, and detach as one assembly, but I only bothered with one window for this tutorial. I applied some basic materials to make it all pretty-like.

It should be noted that, these days, I start with the cockpit windows and build the nose around that.

EDIT: It should also be noted that if this were a DC-9 the #2 window would need a curved bottom edge. You'll have to do some gefroingling to do curved surfaces, and that's a process of trial, error, and planning where you want your edges.

 

[ErickC]

REFERENCES

Your model is only going to be as accurate as your data. With a little bit of attention to detail and some careful interpolation, a reasonably-accurate model of even a rare subject can be made. Let's think a bit about how to best use the different kinds of data available to us.

Photographs

Photographs are useful primarily for figuring out how things fit together during different phases of operation (such as the bizarre intermeshing trailing-edge flaps on the Falcon 20, or the equally bizarre intermeshing slats on the Mercure). They are useful for providing position data for probes and whatnot, so long as certain criteria are met. For example, your station diagrams or 3-views might show some panel lines, and you might be able to determine the position of a probe based on the location of that particular panel line with a reasonable degree of accuracy.

The main drawback to photographs is perspective distortion. Those of us who have had ample education in the visual arts know how to see through it, but if you don't, then I suggest learning how vanishing points work. A working knowledge of how perspective distortion changes the appearance of things can allow you to draw lines to vanishing points in photographs to determine relative locations of things. You can even guesstimate the position of things downrange with enough practice, but this is something that can't really be taught without using up a lot of space, so I won't elaborate on it, just know that practice makes perfect.

Your ideal photograph, if you need to figure out a shape, is one that is taken from very far away with a telephoto lens. This minimizes the effects of perspective distortion greatly, giving us the closest we can get to a real-world orthogonal projection. Good luck getting a top view, though - unless Clay Lacy is a good friend of yours.

Manufacturer's blueprints

These are the gold standard - if you have the plans they used to build the real thing, you're going to be in good shape. Unfortunately, these are exceptionally difficult to get your hands on. They will read a lot like station diagrams, with all dimensions from an arbitrary reference datum. American airplanes will generally be in inches. Airplanes from other places will generally be in metres.

Station diagrams

Station diagrams, as I have mentioned before, are really great for numbers, but can be a little iffy for contours. These are going to be the easiest sources from the manufacturers to find, as they are readily available in maintenance manuals (and can sometimes even be found with the assistance of our good friend Google).

3-view drawings

3-views are my absolute last resort - I will make a model from photographs and numbers first. This is because they're only as good as the artist who made them. Note that I said "artist," not "engineer." Some 3-views are made from engineering drawings and are very accurate. Most are not. The best ones come from Russia. Be very wary of 3-views.

WHERE TO GATHER REFERENCE MATERIAL

The internet is a fantastic resource, depending on how common your airplane is. For example - the 737 is a challenging airplane to build a model of, because it's very complex, but it's the hardest of all airplanes ever made to screw up. Why? Because it's absurdly common, and good resources are easy to come by. Any Boeing aircraft can be reasonably built using the CAD drawings that Boeing supplies on their website. They're free, available to anyone, and can be imported directly into GMax. You literally cannot screw up the 737-200 without putting in extra effort. But the 737 model builder also has another invaluable reference: Chris Brady's website. Consider the entire website required reading if you plan on building a 737.

But there are other internet resources that you might not have even thought of. Google patents is one. You would be amazed at what you might be able to find. This patent is for a modification to the 737s trailing edge flaps. It has detailed diagrams of the transmission. You could use this in conjunction with the videos I posted above to build a very accurate model of the 737s trailing edge flaps - and you also can modify them to represent this somewhat common modification. That's a lot of good information, and you didn't have to pay for any of it. Look for academic papers. You would be amazed what you can dig up. I managed to find pure gold for the DC-9-10 wing in one - airfoil profiles as a percentage of chord, thickness to chord ratio as a percentage of chord, things like that. I dislike the overuse of the phrase "think outside the box," but it carries a bit of truth.

Incidentally, if you plan on making a DC-9 - the wing is not straight in frontal profile at all, like the presently-available models depict (and I am equally guilty here). It has a major increase in thickness inboard of the vortilon, and has an S-curve in profile to the tip. A thickness-to-chord chart of the DC-9 wing looks like a profile drawing of a ride cymbal.

Scanned materials aren't too hard to come by these days, either. You'll generally pay for them, but they're generally cheap. This guy sells high-res scans of all kinds of great stuff - including station diagrams and drawings for display models that were obviously made by the manufacturer, and obviously made from blueprints. FCOMS and AMMs for older airplanes can be had for ridiculously cheap in scanned form - think in the $20 range.

See this?

Ebay. 20 bucks. It sits in my nightstand. Ebay is also useful for photographs of parts. Yes, there are people selling commercial airplane parts on Ebay. And they will generally have photographs of them from many angles. And you can save those photographs. You could buy the parts, too, I guess, but they're usually not cheap...

And last, but not least - museums. If you can stand to fly Spirit, and can stand to fly on funky days, you can get to museums relatively cheaply (I paid $40 to get from Minneapolis to Atlanta and $30 to get back home recently) and view the genuine article in many cases. Want to photograph a CV-880? Go to Graceland. Want to touch a 767? Go to Atlanta. There's a 727 in St Paul, and the first 737 is in Seattle. The museum might even have technical drawings if you ask, and might be willing to make copies for a nominal fee.

You can build a good model really cheap these days.

 

[Heretic]

I want/need to rebuild the engine-related gauges (and probably .air file tables along with drag tables) of Sky Simulations' DC-9 models* at one point. You don't happen to have some performance charts per chance?
There's Schaufele's aerodynamic analysis, but it only deals with the 10 series.

And do you know if SGA's FDE guy is still around?
I did a 727 FDE and used a few of the custom .air file tables and check if he's cool with it.


*Don't look at them too closely as they got the subtle, yet distinct differences between the models wrong, but it's the only VC-equipped complete DC-9 range out there that I know of.

 

[ErickC]

I haven't heard from Fraser in ages, last I heard he left FS for good, but didn't mind if we built on his existing work for free use (at one point we actually uploaded all of the special SGA XML animations, possibly to Freeflight, I think). I actually have some very good data tables from Douglas, not much engine data, but lots of killer graphs with all the things you'd imagine would be in them. Send me an email. Same address it's always been (you've probably got it in a few readme files!). The JT8D actually is an exceptionally easy engine to guesstimate N1/N2 relationships with, if you have recordings from the last few rows of a DC-9 or 727. This is because the N2-related bearing noise is obtusely loud, and the fan can actually be heard all the way down to idle as a low-frequency fundamental.

Start here. Depending on the engine you are dealing with, the maximum N1/N2 values will be different. Notice there are no idle N1 values given. This is where guesstimates come in. I have a takeoff clip at a known RPM setting - 96.6% N1 - and by analyzing audio clips for that low-frequency blip in the spectrum at idle, I can divide that frequency into the fan rotor frequency at 96.6% and calculate the approximate idle value. Because I know the maximum RPM of both rotors for the JT8D-15/17 (102.4% N1, 100% N2), I can calculate the N2 value at 96.6% N1 based on a simple linear relationship. I can then use that as a reference because the bearing noise varies precisely with N2 RPM. In other words, if you have several clips at different RPM levels, you can analyze the sound and look for the two easily-found rotor noises and divide them into the frequency that you know is max RPM to determine what RPM each spool is running at. Then you can make engine tables as you like. It sounds harder than it actually is. I have some JT8D-17 clips from 25A in a DC-9-51 if you'd like them.

On the subject of DC-9 model differences - if you pay close attention to my last series of DC-9s, you'll notice the -30 is only accurate for a -33 or -34, the -50 ought to have angled reversers, the airfoil/chord on the -10 is way off, and all kinds of things. I may have to make a new DC-9 someday to atone for my screwups.

By the way - if you want to both maintain accuracy and have a relatively easy path to a full DC-9 family, the most efficient order is 10, 20, 31/32, 33/34, 40, 50. This is because the -20 has all the wing changes from the -30, plus the anhedral tail (I think), the -30 just needs a fuselage stretch (be careful about the half-station aft of the pressure bulkhead, it does change engine position relative to the windows), the 33/34 just needs angled reversers, a wing incidence change, and a new wing/body fairing, and the -40 and -50 are just stretches from a -33/34 (mind that the -40 did not come from the factory with angled reversers).

Oh, yeah, for sound developers - harmonics. Learn about them. Understand them. Jet engines are all about harmonics and perceived frequencies that aren't actually there. The JT8D actually produces some weird resonant noises that can make it sound like two engines just above idle. There's a distinct N1 idle harmonic at around 1600/3200Hz (most noticeable in the -7, but audible in all of them) that resonates somewhere in the engine, but maintains a constant pitch. When they increase power to just above idle, you'll hear the actual noise from the engine pitch up, but the resonant note will remain constant until it fades out. It's like Louise Post's vocals on Everlong. Once you hear it once, you'll always hear it forever.

 

[hairyspin]

The only thing with really good, manufacturer-derived, drawings is they've often already been used for existing models ... you know, the aircraft every sim has available in umpteen skins and rivet-counter detail. You gotta really love an aircraft to build a properly good model, but there are already so many Mustangs. And Spitfires. Not to mention Corsairs and Cessnas, Pipers and P-47s.

 

[ErickC]

Actually, on that site I mentioned, there are some station diagrams for the Convair jets, and good blueprint-derived drawings for the DC-10... let's face it, the last freeware DC-10 was (derivatives included - I know it's been reworked for FSX) mine, and it's severely lacking in overall accuracy. I originally only intended to build the NWA-spec -40, but everyone wanted a -30, and the JAL -40, and a -15, and a -10 - I didn't know enough about all the differences between series to make an accurate model back then. It shows; they're all chimeras in a weird kind of way.

I definitely see your point about the abundance of Spits and Mustangs, though. I'd like to see a Cessna 421B. I knew a model builder who was making one some time ago, and he even got recruited into SGA, but he left a little on the "screw you, I quit model building forever" side. And I don't blame him, but that's a long-dead personal issue, a topic that I've addressed before, and definitely not in the scope of this thread.

 

[Heretic]

I haven't heard from Fraser in ages, last I heard he left FS for good, but didn't mind if we built on his existing work for free use (at one point we actually uploaded all of the special SGA XML animations, possibly to Freeflight, I think). I actually have some very good data tables from Douglas, not much engine data, but lots of killer graphs with all the things you'd imagine would be in them. Send me an email.

If it's c_d - c_l graphs and (maybe) other stability derivatives, I'm game.


- Edit:
Eh, sent you a mail anyway.

...

The engine tables are already present as I've created them for the 727 FDE, although there was some extra-/interpolation involved. I have some cruise charts for the -15 on the -200, so at least this one cruises pretty much on the numbers. The rest are...plausible.

On the subject of DC-9 model differences - if you pay close attention to my last series of DC-9s, you'll notice the -30 is only accurate for a -33 or -34, the -50 ought to have angled reversers, the airfoil/chord on the -10 is way off, and all kinds of things. I may have to make a new DC-9 someday to atone for my screwups.

I'd help out. Not sure with what as I've grown to dislike any dev work that's not related to XML, but the thought counts, right?

(For an easy way back in: I have your 732 in 3ds Max with a complete VC model in FSX standard. If I remember your post in the HJG forums correctly, it needs some detail upgrades like angled inlets, but I just can't be arsed to tackle it. Wanna have a go?)

By the way - if you want to both maintain accuracy and have a relatively easy path to a full DC-9 family, the most efficient order is 10, 20, 31/32, 33/34, 40, 50. This is because the -20 has all the wing changes from the -30, plus the anhedral tail (I think), the -30 just needs a fuselage stretch (be careful about the half-station aft of the pressure bulkhead, it does change engine position relative to the windows), the 33/34 just needs angled reversers, a wing incidence change, and a new wing/body fairing, and the -40 and -50 are just stretches from a -33/34 (mind that the -40 did not come from the factory with angled reversers).

My observations about the SkySims model couldn't have been made without prior study of this very handy guide to Douglas' greatest success/mistake:
http://www.airlinercafe.com/page.php?id=396

 

[lionheart]

Very cool tutorial, EricC! Love it. It will help out a LOT of souls.

 

[antaris]

Hello lads

@Erick:
Dear Sr., this thread is impresive and very interesting. In advance, thank you for taking time to write it and share it. At least for me, is pure gold.

All the best,
Sergio.

 

[ErickC]

UVW MAPPING AND TEXTURE LAYOUTS - MY CONCEPTION OF BEST PRACTICES

Rule number one - befriend some repainters and listen to every word they say.

I am convinced that the only reason any of my models had any real popularity was that they were made efficiently (If I can confidently say one positive thing about my otherwise average abilities, it's that I will match or beat anyone on drawcells and texture vertex counts any day of the week) and easy to paint. That wasn't an accident. Before I built models, I repainted other people's, and I got a real taste for what I didn't like (I remember the PSS Flanker being kind of a real pain). I also spent a lot of time communicating with people like Steve Drabek, Jim Campisi, and Mike Baumann about what kind of texture layouts they'd like to deal with. Here's what we concluded:

-Precise alignment is a must
-Consistency is good
-Simplicity is good

I've added a few extra ones with subsequent generations of models. These include:

-Big maps with lots of things crammed in them are tidier and easier to manage overall in terms of the number of files to keep track of. If you were repainting a wing, would you rather have five texture files open, or one? I'd rather have one, so I can edit everything at the same time. It turned out, by accident, that this was a huge benefit to overall performance.
-Big, continuous surfaces are good. This is my #1 goal when setting up a basic layout. Easier to manage, easier for the painter to visualize where things fit, better on the texture vertex count - you take a hit on resolution versus what can be done when smaller sections are mapped to an equal size bitmap, but with 4096 x 4096 pixel images these days, it ain't no biggie.
-Map all major parts to the same resolution. Map comparable aircraft (as in two similarly-sized narrowbodies)to the same resolution - no reason why someone making both a DC-9 and a 727 can't have 6" on one equal 6" on the other. It just takes some planning (bear in mind that smaller aircraft will have some wasted space. This is going to be true if you have a DC-9-15 and a DC-9-51 using the same basic layout, so it's really no biggie). Then your SkyTeam logo can be the exact same image at the exact same resolution on both airplanes. Consistency is good. Mind you, I think in terms of fleets, most developers don't. To each their own - but it's a good practice.
-Corollary: If you're making a DC-9-15 and a DC-9-51, use the same basic layout. The short fuselage on the -15 will lead to some wasted space, but the consistency that you gain will make painting entire fleets a simple task. I stole this basic practice from PFG and their DC-9s.
-The big thing I used to do with FS2004 was have the tail and the fuselage one piece so you could easily go from one to the other. I decided a couple years back that I was sick of the polar distortion on fuselage tops and bottoms, so the Captain-Sim style fuselage became the way to go, so long as precise alignment in the Y-axis (long ways) is maintained. This is so that you never have a pixel on one surface in the middle of a pixel on another one. Common sense.

So how do we get really precise alignment? Match marks:

How do we line them up nice and precise? Unwrap UVW and zoom. I move things around until they're close, then I turn off the map in Unwrap UVW (it will slow down as you zoom in - even this i7 loads-a-RAM good-vid-card machine doesn't like zooming in as far as we need to go with a map displayed), zoom on a fair amount, and zoom in on a mazimized window, get it closer, and repeat until we can't zoom no more. I calculated once that my precision ended up being around 1/1000 of a pixel with this method. It's important to use orthogonal views only when doing this, and it's also important that this view allows you to see what you're doing. If I'm matching an aileron to a wing, for example, I'm going to be looking from the top or bottom (depending on what side I'm aligning). I use plus-shaped alignment marks to align the top and bottom of things. Since what I build usually has swept wings, this means I can align it left/right from the front, and fore/aft from the side. Here's a horizontal stab on a simple AI model:

Here's the real trick: Mapping is the last thing I do, but I usually build the stuff on the right side (I build left to right) anyway, because it's nicer to look at (and I can have a fully animated model tested and QAd before I start throwing images on it). I don't map from scratch, though, on the right side, ever. I invert the colours on this match mark for the ones on the opposite side, then clone and mirror the part, move the original mesh out of the way (subobject mode - don't move the whole thing), attach the mirrored part, and delete the old mesh. Then I can add a new unwrap UVW modifier, and align the top to the old match mark, and the bottom to the top when this is done. Now your mapping is the exact same - albeit mirrored - on all wing surfaces. All tolerances are the same, and you can even be absolutely certain that a panel line on one side will be in the same spot as a panel line on the other. Consistency is paramount in my world.

Left side match marks are green, right side are purple. I'll never accidentally mix up sides!

Why do I map last? One, so I can have the model thoroughly checked for flaws before the picture is muddied up by textures (it's harder to see a wing ever-so-slightly out of alignment with the associated hole in a wing-body fairing when you don't have a nice dull grey surface to look at). Secondly, because I've built 19 737s across four versions of MSFS, quite possibly more than anyone else in the MSFS community. That's not bragging, that's me telling you I chose a masochistic subject best left to someone else 19 times. Sometimes, when you're animating a part, you find it doesn't fit right when you animate it, so you need to finangle it a bit. And this happens very often when you build 737s. And now you have to modify the texture. And now, more often than not, you have to play with the mapping - especially if you prefer to apply the UVW map once and do your dirty work in unwrap UVW, like I do. Unless you map last. I always lay out and map with finalized parts.

On unswept wings, by the way, I put my fore/aft alignment marks on the tips.