Need for Research

[Ivan]

Hello All,

Warning: This is going to be a bit long winded.

I released a Birdcage Corsair a few months ago. It had been sitting for a couple years. I originally started it because someone wanted a Night Fighter Corsair with a Radome and the NF Corsairs had the framed Canopies.

The Corsair was a Project I started MANY years ago, back in 2001. This was my second completed aircraft design project. The ones before were a N1K2-J Shiden KAI which was lost entirely with a disk crash and a P-51A Mustang which was completed.

At the time, my view of CFS aircraft was that they were generally very crude. I was also trying to learn how to use Aircraft Factory 99. Each project at the time had a well defined goal. The goal for this project was to figure out how difficult it was to build an aircraft with a cranked wing such as a Stuka or a Corsair. I had found a few Corsairs available for download but other than the Alain L'Homme F4U-5, none of them were very good. Many folks repainted that post-war Corsair to be wartime versions. To me, this didn't seem appropriate because the -5 version had cheek pouches and a 4 bladed propeller unlike the wartime -1 Corsairs.

The idea of proper research didn't come up because I figured that very few of the aircraft for CFS actually looked like the real thing and also because the greatest concern was problems that would come up in actual building.

I spent probably not more than an hour or so in finding a drawing that looked like a Corsair. The next few days were spent with a dial caliper in measuring dimensions and using a calculator to translate into AF99. Correctly translating the drawing was important. Never mind that the drawing itself was rather poor.

The result was a "Airplane" that didn't have any serious bleeds and kind of resembled a Corsair. There were serious shape problems which I didn't notice because it was a "CFS Corsair". The first revision corrected some minor shape problems. The second revision added a pilot. The third revision happened several years later when it finally occurred to me that the bends for the wings were way too far out (about 1 foot). During this rebuild, I found that there were LOTS of other shape problems. The Wings and Flaps got a total rebuild but were altered a bit to make them fit the original Fuselage.

That is where this project stands today. I have been debating for a while whether or not to re-release this project because the shapes aren't very good. I probably will anyway because most of the work is already done and it only needs a couple hours to complete little stuff.

The moral of this story is to do your research properly to prevent problems WAY down the road.

- Ivan.

[Ivan]

Thanks to Hubbabubba and Alain Breton before him, the idea of locating the Virtual Cockpit view and Padlock view together is not a difficult task. This shows a successful confirmation of the VC and Padlock Views being in the same location or at least VERY close.

- Ivan.

[Ivan]

In doing some further research, there appears to be several distinct canopy types for the Corsair.
The F4U-1 had a Birdcage Canopy
Later -1 aircraft had a small bulge in the first bay to hold a rear-view mirror.

The F4U-1A through F4U-1D had a Windshield Frame that came to a rounded point at the top.
The Windshield Braces at the front came together a bit as they went up.
The Canopy Braces disappeared between the -1A and the -1D models but the shape didn't change much if at all.

The F4U-4 had a flat panel on the front Windshield Pane.
The Windshield Braces appear from the front to be parallel as they go up.
The top of the Windshield frame was flat to match the front Pane.
The Canopy was noticeably more bulged.
The later versions lost the semi-circular part at the rear of the Canopy Frame.

The F4U-5 added a Dorsal Fairing on the Fuselage at the Aft end of the Canopy.
The Canopy was bulged even more and reshaped to match the Fairing.

This is a fairly good example of a F4U-4 even though the author called it a F4U-1D.
The -4 had a Chin Scoop while earlier versions did not.
Note that it even has the 4 cannon typical of a F4U-4B?

http://simhq.com/forum/ubbthreads.ph...lacement_.html

Research is a good and necessary thing.

- Ivan.

[Ivan]

There are some problems with these drawings, but I really wish I had knew these existed before I started my project way back. From these drawings, you can see what the profile should really look like.

The problems I have seen with these drawings are the following:
1. The Pitot Tube is too short.
2. The Front Braces for the Windshield don't look quite right. I believe the bottoms are too close together.
3. The Landing Gear Doors are not correctly shaped.

If I had gotten it this close, I would have been happy.

- Ivan.

[Seagull V]

Some info available on the net :-

http://www.voughtaircraft.com/heritage/html/down.html
http://www.alternatewars.com/SAC/SAC.htm
http://membres.multimania.fr/wings2/

Regards, SG

[Ivan]

Thanks Seagull V.

Those are pretty good references.

Now, the next question is how badly do I want to fix my old Corsair before I release it.

- Ivan.

[Ivan]

The F4U-4 had a flat panel on the front Windshield Pane.
The Windshield Braces appear from the front to be parallel as they go up.
The top of the Windshield frame was flat to match the front Pane.
The Canopy was noticeably more bulged.
The later versions lost the semi-circular part at the rear of the Canopy Frame.
.

Early versions (WW2) of the F4U-4 had Canopies and Frames visually identical to that of the F4U-1D.

- Ivan.

[Ivan]

It is amazing how relatively small changes in shape make the whole aircraft look wrong. The Purple is MY Corsair. The Black is Paul Matt's drawing. His looks correct. Mine does not. This overlay drawing surprised me a bit with how small the differences actually were. The Aft Fuselage on mine is actually not as fat as I thought. Paul Matt's version is more fat.

I'll let you all decide. I had to seriously reduce the image to upload here. Anyone who wants the full sized version (about 1.8 Meg) can email me for the full sized version. It is set at 1 pixel == 0.01 foot which makes it about 3500 pixels wide.

Comments?
- Ivan.

[Ivan]

A proper rework of this project requires some tools that I haven't built yet. I figure that the Cowl should be moved down a bit as should the Center of Gravity of the entire model. This project currently stands at Version 3 which has significant improvements in wing shape over the last version.

Perhaps this aircraft should be released before further rework?

- Ivan.

[smilo]

A proper rework of this project requires some tools that I haven't built yet. I figure that the Cowl should be moved down a bit as should the Center of Gravity of the entire model. This project currently stands at Version 3 which has significant improvements in wing shape over the last version.

Perhaps this aircraft should be released before further rework?

- Ivan.

as always, it's your call.
if you're happy with the current version,
then, release it. if not, wait until you are.

i've always seen the corsair
as long and lean, proportionally.
the short, fat models
just don't look right to me.
but, maybe, that's just my perception.
or maybe, it's my old screen resolution.

i can't help but wonder,
why the discrepancy in drawings?
are Paul's measurements off?
i can't that imagine your's are.

[Ivan]

Hello Smilo,

Actually your confidence is misplaced. I KNOW the drawings I used way back are not accurate. I didn't do my research back then. AF99 was the greatest concern at the time. These days it is different. AF99 is easy. The shapes are tedious.

I found out when I tried to use another of these drawings for the N1K Shiden and the Cowl Scoops didn't even come close to matching photographs in the same chapter.

I believe that for the most part, Paul Matt's drawings are pretty good. That is good for my Corsair also because as you can see from the overlayed drawing, there isn't very much difference between the Black Paul Matt Drawing and my Purple 3D Model. The only thing I did to create the two drawings to overlay was to resize mine to 3333 pixels long (33.33 feet) and the Paul Matt drawing to 3334 pixels (33.34 feet). I used 33 feet 4 inches for dimensions. The Paul Matt drawing specifies 33 feet 4.125 inches.

My perception of my model being overly fat isn't supported by the comparison with dimensional drawings. Even the Engine Cowl isn't too fat. It is just a couple inches too high which throws the shape off. The back side of the Rudder is VERY close. The Fin is fairly far off except at the base. The Canopy is too high and narrow but you can't see it from the drawing overlays. Almost everything lines up on the Longitudinal axis. If you look at the the Landing Gear Doors, there are some significan differences, but I believe that is where Paul Matt's drawings are NOT accurate.

Either way, I figured the only missing things were a "ReadMe" file and couple Screenshots. With the rebuild, the AIR file, Checklist, DP, and probably textures would all change.

- Ivan.

[Ivan]

This is an overlay against another profile from a different source. It actually looks closer than the first one. This one shows a slight difference in the Engine Cowl Diameter.

So Which one is the CORRECT one????

- Ivan.

[smilo]

who knows?
got any production drawings?

i would bet that if you found
5 different sets of drawings,
you would find 5 different values.

it seems to me that, at this point,
it's up to your artistic license.

[Ivan]

Artistic License is fine, but it is also good to know what one is "Sposta" loo like. The two things that seem to vary a bit even in photographs are the Cowl and the Fin.

I have been looking around for a good CFS Corsair lately. They seem to be quite rare.....
Anyone have a link to one?

- Ivan.

[hubbabubba]

Hello Ivan

I can't tell you what to do, but I can tell you what I would do in similar circumstances. First, I would not hesitate to make a "collage" of different blueprints, taking what I think was more detailed in one and only sketched in another. Second, I would try to find the "right look" in studying photographic evidence. With your keen sense of observation, this is the kind of game you're better at than anyone else. Third, I would recognize that there's no such thing as a "perfect Corsair".

In a book called "Whistling Death - The Test Pilot's Story Of The F4U Corsair", the author says that modifications were so often made on the production line that someone having worked there, with a good memory, could tell you which month/week/day such airframe was rolled-out. So, if you see two different rivets' pattern on Corsairs of the same era, don't make too much of a fuss about it.

[Seagull V]

Accuracy of Drawings.

Many years ago I spent many hours in my then favourite hobby of plastic modelling.

A very well respected member of that community did an article on the P-51D Mustang comparing different models and different drawings.
There were quite a few differences between them, now you would think that aircraft like the Corsairs and Mustangs would be very well documented,
but it appears that different people use different sources.

People who do colour side profiles often start with borrowed drawings so unless you know that they are meticulous with their own research for that
profile be careful about using it. Also I have seen different colour profiles of the same aircraft with different colours.

As a general rule I look for 3 different sources and compare them, if there is still no consensus of agreement between them I look at as many photos
of the period aircraft as possible to compare the drawings with. Photos in the old days were often airbrushed so that is always something to consider as well.

Finally, you have to be happy with what you are doing, as well as remembering the learning curve that you started with.
My first plastic model aircraft was a Grumman Avenger and from memory the wings were on back to front with more glue on the transparent canopy then there was plastic, but hey I was proud of my work back then.

Enjoy your previous work - if you want to do a rebuild/new one then get the resouces to a point that you happy with and build away.

Regards, SG

[Ivan]

Hello Hubbabubba,

Thanks for the vote of confidence. The problem here is that I can't find ANY set of drawings that I consider "reliable" any more. They are all starting to look bad. I know there are problems with the shape of "My Corsair" and will be addressing at some point. I believe there IS such a thing as the "Perfect Corsair" in that it would match the manufacturer's blueprints. I know I won't get there because with AF99, you really can't within the resource limits. Still, I would like to know what one is supposed to look like. I had thought that mine was overly fat, but the more I look, the more reasonable it appears. There are still a few things wrong, but I haven't quite figured out what those things are yet.

I read the book "Whistling Death" by Boone Guyton quite a few years ago. It is excellent. The differences he describes though are mostly in detail and installed equipment. We don't have enough detail in a CFS aeroplane to represent that kind of stuff.


Hello Seagull V,

I also used to do a fair amount of plastic modelling. I still do with GI Joes, but don't build much in the way of airplane kits. Most of what I built was either 1:72 or 1:48. In the Corsair line, I still have a Monogram F4U-4 in 1:48 and a Ertl 1:72 F4U-1A or -1D. The Monogram seems to have a decent shape. The Ertl plane is too long and thin and it is a nice looking plane despite my work. It does however have a lot of incorrect details for a Corsair. I still have a few more kits that I have not built. There are a couple Hasegawa 1:72, an Arii 1:48, a Revell 1:32, several Monogram 1:48 and probably a few more I haven't seen in a while.

Regarding looking at photographs, I have looked at several hundred within the last few weeks and gone through at least 4 books on my shelf: Squadron Walkaround, Squadron In Action, Motorbooks, WarbirdTech and a couple I can't remember offhand. All I can say is that this critter is TOUGH to pin down.

I am tending more toward a rework of a few critical places on my CFS aeroplane rather than a complete build from scratch these days. It seems like the more I look, the less I see really wrong with the general shape.

Check out this overlay comparison with an actual photograph. It is difficult to find a good profile photograph, but this one isn't that far off. Note that the differences between my 3D Model and this photograph shows differences in different areas than the last two Overlays???

- Ivan.

[Seagull V]

Side profile & mesh looks pretty good, a smidgin out in a few places - but I would be happy with that.
Wings look a bit out, if you move those a bit in AF99 you should have a nice model.

Will have a look at my references that I can dig out of my boxes, will let you know if I find anything usefull.

Regards, SG

[Ivan]

Hello Seagull V,

I just tried yet another reference: Squadron's Detail & Scale book on the early Corsair. This is the result of yet another overlay. The wings don't look quite as far off on this one. The claim for these drawings is that they were done with reference to factory dimensions and blueprints.

I personally am not so sure any more. I believe the photographs are the definitive reference, but even they are not consistent. Note that there are Corsairs that appear to have a slight hump about where the Aft Antenna Mast meets the Spine. Note also that the -1D Corsair at Udvar Hazy does NOT have this spine.

Am I really the first person to compare these drawings with actual photographs???

(Yes, I posted another image. Brave, aren't I?)
- Ivan.

[Ivan]

From an Eyeball check of my Corsair, the cowl opening apears too large. The relatively small appearance of the cowl opening is what struck me first when I first saw the F4U-1D Corsair out at NASM Udvar-Hazy.

A couple weeks ago (starting before my vacation), I had started to look for photographs with a front view of the Corsair cowl. I downloaded a few and did a pixel count to compare the diameter of the cowl overall in comparison to the diameter of the opening.

Last night, I did a comparison of these numbers with the cowl radius of MY Corsair. Here are the rather surprising results:

The Cowl of my Corsair has a radius of 2.48 feet at the back and 2.41 feet at the front before the rounded section to the opening. The opening itself is 1.90 feet in diameter.

Because I often can't tell if the pixel count of the photograph is of the large diameter or the smal diameter, I used both for calculations.

The range appears to be from 1.83 feet minimum to 1.96 feet maximum. The numbers (High on one side and Low on the other) appear to show 1.89 feet and slightly over 1.90 feet which is dead on for what I had built.

Now if this is correct, then why does it look wrong?

- Ivan.

[Ivan]

This overlay shows the greatest differences.

- Ivan.

[Ivan]

From the overlay above, the size looks generally pretty good. Note that there are differences between the Propeller Blur in my 3D Model and the Propeller Arc in the colour picture. There are differences in the Propeller diameter between different models of the Corsair. The Wing Tip was missing in the Colour Drawing and I pasted a reflection of the opposite Wing Tip to correct that issue as well as rotated it by a couple degrees so that the Wings were level.

The "Big Difference" is that the Wing on my model is almost exactly 3 inches too high and there are notable differences in the Wing Root area which I may continue to ignore in an updated model.

I believe I have also figured out the "Hump Back" profile that is shown in some photographs and drawings. It exists because the Fuselage Spine has a slight parallel section on the sides though not the top so that the Canopy can fit over it when it slides back. If the photograph is taken from above or significantly below the aircraft's horizontal center line, it shows up. In a strict Profile shot, it doesn't show up.

- Ivan.

[Ivan]

In order to shift the whole AF99 project a bit, I have done a fair amount of programming.
The single AF99 Parts have been addressed.
The program I wrote for this purpose years ago is called "MoveIt".
Recently I wrote another program called "StructMoveIt" that shifts a AF99 Structure Fore and Aft.
This is all that is really necessary for the AFS file. The Templates get moved around using "MoveIt".

After all the pieces are moved around, the textures will not line up.
The latest program I wrote is called "MoveTexture" and edits the AFA project file to shift the values used to overlay textures.

Screenshots attached show that currently the textures don't line up. That isn't very pretty, is it?

Note that the AIR file will need modification as well because things like Gear contact points and Cockpit view points need to agree between Model and AIR file.

- Ivan.

[Ivan]

A little more tweaking of the program and I got the Top and Side Textures lined up.
Correcting a very silly bug and the result is a program that will move ALL the textures to their appropriate places as can be seen in the last screenshot.

To Summarise:
At this point, I have a set of utilities that can shift the CoG of a AF99 project in a matter of a few minutes.
This allows for very easy correction of a project at its basics rather than using the Offset method.
I am sure you folks know of a few projects out there that don't look all that bad but just have their CoG incorrectly located.

Good Evening.
- Ivan.