Thunderbolt

[Ivan]

Hello Aleatorylamp,

History tells us that overall, the higher activity factor "Paddle Blade" Propellers improved relative performance at least from the Pilot's point of view. That sounds like a qualified statement and it is because a full discussion of what the differences were would take a LONG time.
Consider how one would go about simulating the differences accurately with the CFS Propeller Tables and you will get an idea.

Only the Hamilton Standard Hydromatic Propellers had relatively straight leading and trailing edges and those really were not nearly as straight as it would appear from a distance. Compare the pictures attached. Note that there is also a twist along the length of the blade that I do not model at all.
This is one of those cases when I had to choose between following what I could find as the blade planform or making it look correct from different angles.

The prior screenshot was version 1. This screenshot is version 3.
I believe that version 2 is actually closer to correct dimensions but just didn't look quite as nice without the twist that is in the real blade.

- Ivan.

 

[Ivan]

Reality Check

Tonight I decided to do a few checks to see why the Thunderbolts all look just a bit off to me.
First was a check against overall dimensions.

The references don't seem to agree too much in detail:
Wing Span is sometimes listed as
40 feet 9 inches,
40 feet 9.25 inches, or
40 feet 9 5/16 inch.

The last number comes from a 3 view drawing from Republic and I am most inclined to trust it.

Length is usually listed as
36 feet 1 inch, or
36 feet 1.75 inch.

however, the Pilots' Manual lists it as 35 feet 7 inches.

In Aircraft Factory 99 I checked the Wing Span and Overall Length and got
Wing Tips at +- 20.39 feet, so Wing Span is 40.78 feet.
Nose is at +11.97 feet and Tail is at 24.17 feet for an Overall Length of 36.14 feet.

40 feet 9 5/16 inch works out to be 40.8021 feet so Wing Tips should be +- 20.40 feet.
Not much difference, but this means that I probably used a reference that stated Wing Span was 40 feet 9.25 inch or 40.7708 feet.

Length of 36 feet 1.75 inch would work out to be 36.1458 feet which is pretty cloe.

Next, I decided to overlay my model's Wire Frame over a P-47D-40 profile from the Detail & Scale book.
In comparison (The Wire Frame is in Red), the Longitudinal match is pretty good, but the D&S image is noticeably thicker from top to bottom. I wonder if this is worth fixing?
A correction would mean rebuilding the entire Fuselage which does not look like much fun.

- Ivan.

 

[aleatorylamp]

Hello Ivan,
Differences between specifications quoted by different sources are very
annoying, as one never knows if a different variant of the same model
did have different dimensions or not.

Then, it can be worse with drawings. It always amazes me how inaccurate
some artists can be, even if their work is found on reputed internet pages.

Usually the differences are very small and do not not warrant a rebuild, for
example in this case, the entire fuselage - I´d say it´s just too much work!

In the case of your posted drawing: How reliable is it? It also has a fin
extension leading to the cockpit, which is perhaps quite unusual.

Another thing to take into account on photos, is the camera distance-setting
with which the plane was photographed. Wide-angle or tele-objectives, I find
can often distort the image enough to cause differences in appreciation on some
dimensions. Setting for closer-taken photos make things look bigger and fatter.

Lighting can also have some influence. In the case of the comparing photo and
screenshot you posted, perhaps this Thunderbolt looks a bit fatter because of the
extended flaps, that combines with the strong shadow on the fuselage undersurface.
Taking that into account, I can´t see any difference between the two.

Anyway, at the end it depends on one´s personal preferences, and the
patience on can have with repeated corrections, which may or may not even be
necessary on a build.

Cheers,
Aleatorylamp

 

[Ivan]

Hello Aleatorylamp,

The dimensional differences appear to be mostly an issue of rounding.
There is also the possibility that the Length was taken with the Fuselage level in one case and at a three point attitude (12 degree inclined) for another dimension.
The Length while sitting on its wheels is not useful to a modeler, but might be very useful for someone who is trying to determine how much space is necessary to park the aeroplane.

Considering how closely the longitudinal dimensions match, I am pretty sure I had some kind of reference drawing, but I am also pretty sure there was some element of eyeball measurements which is why there are so many discrepancies. There is also the possibility that the drawing that I was using as a reference was just not good, but I have not yet located the book where I believe the drawing originally came from.
I also know that sometime there are issues with the Detail & Scale drawings, but I am pretty sure they are closer to the real thing than my model is.

I have a lot more reference data now than I did back in 2003. The next step is to check dimensions against the Station Diagram attached below.

Another item worth mentioning is that the Hamilton Standard Propeller Hub is noted as being shorter than the Curtiss Electric Hub though I have not found a source that lists the exact difference. I am finding both 36 feet 1 inch and 1.75 inch as pretty common measurements for overall length and will interpret the difference as due to the difference in Propellers.

Regarding the Fin Fillet extension to the rear of the Canopy:
The loss of keel area from going to a razor back to bubble top resulted in some loss of directional stability.
This was addressed in the last of the D series (D-40) by adding a Fin Fillet.
This Fillet was often retro fitted to earlier Thunderbolts just as Propellers were.
The Fin Fillet was changed in shape with the M and N series Thunderbolts.

Some sources also list the M and N as being about 2-3 inches longer and some do not, so that will need to be resolved if those models are to be huilt.

Although the shape is a bit incorrect, it isn't too far off and is a pretty clean build, so I will most likely just continue with the series without major modification to the Fuselage.

- Ivan.

 

[aleatorylamp]

Hello Ivan,
Interesting information, indeed!
Station diagram are not easy to come by, and would obviously be more reliable.
Cheers,
Aleatorylamp

 

[Ivan]

Hello Aleatorylamp,

So far, this project is following the regular pattern.
What looks like a simple and easy update turns out to be much more complicated than expected.

From the "Conspicuous by Their Absence" Thread, I will quote a rather arrogant and foolish statement:

Hello Aleatorylamp,
.......

Even if you do not edit the MDL for FS 98, the mapping of the textures will not change when I update the model,
so even with the current MDL, there should be no limitations on re-texturing even if it is an issue for a release.

- Ivan.

Some things just don't hold up to close scrutiny.

Upon further examination, I am finding that the texture mapping has some wacky effects.
There are places where some of the mapping does not make sense and others where there are textures that bleed from one piece to another. The only reason that they don't seem to affect the appearance in the simulator is because this is basically a single colour paint scheme. There are other minor problems with laying out symmetrical textures which hopefully will also get addressed.

I will need to add at least one new texture file anyway, so it makes sense to move things around now to correct the minor bleeds and highlights (white background bleeding onto the piece). I remember now that I knew this was happening and that is why the empty spaces in the texture files are all white. The effect was supposed to resemble reflected light.

- Ivan.

 

[aleatorylamp]

Hello Ivan,
Not to worry. Obviously, the comment about textures being unaffected would
only be applicable to minor changes, and the situation is different now,
so I wouldn´t say it was either arrogant or foolish.

My Dad used to say:
"Firstly, things happen differently, and secondly, than what one expects."

The updated model will be a different one, and once it is ready, should Udo want
to make updated Dottie Mae textures for it, it can be done separately. For the
moment he is very busy workwise, and I have no news as to his progress. He also
wanted to prepare two unarmed Ju52´s for FS98, so I don´t know how he´s doing.

Anyway, we´ll take it as it comes - don´t let it disturb you!
Cheers,
Aleatorylamp

 

[Ivan]

Hello Aleatorylamp,

Here is a little history....
Back in 2002, my second child was born.
When the second child (my Son) came along, I had to give up my hobby of plastic modelling.
With two children and two parents, all hobbies had to be instantly interruptible.

A conversation might go like this:

Michael is crying, can you check his diaper! (Note that this is NOT a question.)

I need to hold these pieces together for a couple minutes while the glue sets!
or
I need to clean my brushes first! (Exclamations because it is difficult to be heard over a crying baby if you don't yell.)
Can you go check him?

No, I am giving your Daughter a bath!

......

Two or three hours later, after figuring out why there was a general baby fault, I come back to a couple dried paint brushes and paint jars or a part that sagged and isn't on straight.
After a few of those events, I realised that I could not continue that hobby while the kids were small and needed attention.

A little while after that is when I started building aeroplanes for Combat Flight Simulator.
This Thunderbolt was done about a year after I started building and was probably the fourth or fifth project and at this point the lack of experience shows up in a few places. I am finding many of those issues now with the project as I attempt to update it.

- Ivan.

 

[aleatorylamp]

Hello Ivan,
Well, thank God that with AF99 et al (SCASM, AA, etc.) you don´t
get glue or paint drying up when you´re interrupted - although you
do loose concentration...

So now with the older kids there´s more time, and also, with the
greater amount of more exact information available I ´m sure that
re-taking the Thunderbolt for further attention will be an enjoyable
and well worth-while challenge.

Cheers,
Aleatorylamp

 

[Ivan]

Remapping Textures

Hello Folks,
The remapping of textures is going fairly slowly but it is progressing.
Most of the "necessary" changes have been made at this point.
There are only a couple more items that need to be moved around between texture files.
The number of texture files has gone from 11 to 15.

The appearance of the model is almost exactly as it was before I started.....

So....
One might wonder why I would go through this effort to remap textures when the overall appearance remains the same.
The main reason is that I want to be able to easily develop more versions of the Thunderbolt and the current mapping only worked for the current versions.

The first indication that I would need at least one extra texture file was when the Propeller Hub and Fairing did not fit into any of the other texture files without colliding with something else. In the Razorback's textures, there simply was no room at all.

Next, this rework is intended to represent the P-47D-25.
Early versions of the Thunderbolt mostly had disk wheel covers as shown on my current models.
Later versions tended to leave off the wheel covers and show spoked wheels.
The current version uses just one pair of textures to represent both Main Wheels, so the inboard and outboard textures need to be identical.
There was no room for another Main Wheel texture.

Current build standards also require an internal Canopy Frame which needs its own texture.

Just as in the case of the Main Wheels, the Tail Gear Doors shared textures.
This meant that the inboard side of each door had to be the same as the outboard side.
This prevents applying Zinc Chromate as a corrosion inhibitor to the interior of the doors.
There is a much worse problem though.
On a monochrome all gray aeroplane, everything looks fine but if someone wanted to apply a contrasting colour to the underside of the tail, the interior of the doors would also get coloured the same way.
Stripes on the inside of the Tail Gear Doors would look pretty silly.

The P-47D-30 had Dive Recovery Flaps installed. Additional texture space for these Dive Flaps would have been easy even with the exiting texture files.

Starting with the P-47D-40, a Fin Fillet was added to increase directional stability. This was often retrofitted to older machines.
The Fin Fillet on some versions appears to extend from the Fin to the rear of the Canopy which means that there needs to be texture space for an object that is up to 10.5 feet long.

The last version of the Thunderbolt, the P-47N, had a much longer Wing Center section to accommodate additional internal fuel tanks.
While the texture files for the Inner Wing would allow for the stretch, the Flaps were mapped in such a way that there was no additional room and any stretching would need a remap anyway because the outboard edge was on the edge of the texture file.

If I intend to build more versions of the Thunderbolt, my choice would have been to make changes as I needed them at which point there would be substantial differences in mapping between each version or to make all the anticipated changes at this point and only adjust textures for different versions. There will always be some differences in mapping between the Razorback and Bubble Top versions that are unavoidable because the pieces are so different between the two versions.

The attached screenshot shows a mostly successful update.
I seem to have forgotten to move the mapping for the Aileron Trim Tab.

- Ivan.

 

[aleatorylamp]

Hello Ivan,
So it´s getting even better than what we thought before.
Sounds excellent! ...even though it seems rather more complicated
than initially expected.

For the new metallic texture issue, there´s absolutely no hurry anyway.
Udo is under such work pressure, that it´ll be a while yet before he can
start work on the new Dottie Mae textures.

Cheers,
Aleatorylamp

 

[Ivan]

Hello Aleatorylamp,

Actually I really wasn't concerned about Udo's "Dottie Mae".
He has his canvas and that is really the end of my involvement on that project.

As I commented in my prior post, there actually will be no really significant visual changes other than Zinc Chromate on the Tail Gear Doors and possibly a set of spoked Main Wheels.
The changes are mostly for greater ease in building future Thunderbolt projects.
These are the kinds of things I do on current projects in the planning stage, but back in 2003, these were not issues I was thinking about.

There are a bunch more small changes that also probably won't be noticeable unless I point them out but as the plastic modelers say:
"I'll know they are there!"

- Ivan.

 

[aleatorylamp]

Hello Ivan,
Not to worry! I know the changes are more for future variants
you are planning, not for Udo´s planned new textures.
I only mentioned it just in case it was causing some hassle.
Anyway, I´ll upload the existing model with new textures once
they are done, whenever that may be!!
Cheers,
Aleatorylamp

 

[Ivan]

As mentioned before, the visible differences will be minor.
Here is a comparison between the original texture mapping and the new texture mapping.
The two difference are the Tail Gear Doors and the Main Wheels.

The Wheel textures actually took two tries. I made up a set last night and was planning to overlay them onto the appropriate files this evening. Unfortunately, I managed to do something stupid and overwrote them on the flash drive before I could use them.
They were not that hard to redo but the original version was slightly better.

In the earlier version, if a spoke texture was applied to the inboard side of the Port Wheel, it would also appear on the outboard side of the Starboard Wheel. This would not work well because the inboard and outboard sides are different when the hubs are not covered with a flat plate.

Many little things still left to do before starting on the revised Flight Model.

- Ivan.

 

[Ivan]

Hello All,

Last night, after poking around in another project for a while, I decided to take the D-25 Thunderbolt up for a bit of "Fun Flying".

As some folks may already know, the old Development Computer I had been using died recently with what appears to be a failed Motherboard. It had been having intermittent errors and Video and I/O lockups for a while before it finally died.
One of the big problems with working on the D-25 version of the Thunderbolt was that the Aircraft Factory 99 / Aircraft Animator combination would not always generate a useable model and Combat Flight Simulator would often crash or lockup for no apparent reason.

Until last night, I thought I had a clean build for the D-25 but the first Screenshot shows otherwise.
The rebuild cycle was actually quite fast with the updated Development Computer and things look a it better on the second Screenshot.

I also made a minor adjustment to the AIR file, but that subject is worthy of a new Thread.

- Ivan.

 

[Ivan]

Smilo's Hammer

understood...
i'm still happy to have the visual update
and will wait for the air file improvements.
as you said, it's been a long time.
so long in fact, "smilo's hammer"
has left my memory banks.
sorry, i don't remember it.

Hello Smilo,

Last night, I was working on my version of Eric Johnson's P-39D and was looking for the some files.
Ivan's Workshop is a very large place and there are many places that don't get visited very often.
In a remote storage area of the Workshop, I found a crate with an interesting label on it.
It was too small to contain a complete Aeroplane or Project but looked like a bunch of pieces to a Razorback Thunderbolt.
The pieces were to a very old (probably initial release) of the P-47D-23 Thunderbolt which has been out of production for a very long time.
This was around 2:00 AM, so my Engineers and Technicians were all rather annoyed when I called them up to see what could be done with what we had.
Although that particular D-23 was out of production, the team looked to see if those old pieces could be completed with spares from the current (soon to be replaced) P-47D-23.

The current production airframe parts actually fit without any great problems and the result is what can be seen in the screenshots.

Not all the correct pieces have been located yet.
No one wanted to hang around for more than a general "Proof of Concept".
Armament and wiring for the Instrument Panel are not yet complete.
The Factory Test Pilots are also complainng that this Aeroplane does not have a correct standard Canopy Interior or the current standard "Alpha-Glass" Canopy for better visibility.

This really is a beautiful model and all I can really claim is that I provided the canvas for the artwork.

- Ivan.

 

[Ivan]

Changing Standards

One of the details that is visible in the textures for "Smilo's Hammer" is that the textures for the inboard and outboard sides of the Main Wheels are the same. On my original model, the inboard side was covered by a disc and looked about the same as the outboard. On this model (and on the P-47D-25), the inboard side of the wheels do not have a cover and the spokes are exposed. The outboard side of the main wheels does not show spokes in any version.

This was one of the reasons the textures needed to be rearranged. The original wheel texture was shared for both the inboard and outboard sides.

One of the other things that was missing was the Interior of the Canopy Frame.
The first screenshot shows the modified version which is dimensioned slightly different from the exterior Canopy Frame.
It needs to be different because sometimes polygons that line up well in an Exterior view do not line up as well on an Interior view.
Once the displayed Component looks good from the inside, it is switched from "Regular" to "Smooth".
The switch to "Smooth" causes the interior of each Part to disappear and only the outside Parts remain visible.
The next step is to run a program to reverse the facing of each Part so that it face inward instead of outward.

The second screenshot shows the interior view after the flipping of the Component is done.

The third screenshot shows the exterior view of the Component that now has no outward facing polygons.
This is always an odd thing to look at. It appears to be facing in a direction that it is not. Not many things in the physical world have an inside but not an outside.

- Ivan.

 

[Ivan]

Canopy Frame Installed

Installing the Canopy Frame was fairly easy.
The result should be usable even if the textures are remapped at some later date.
The colour switch to "Interior Green" or Zinc Chromate was just a matter of mapping to a different file from the External Canopy Frame; The offsets are actually the same which makes testing easier.

Next come some Panel, Checklist, and Flight Model changes.

- Ivan.

 

[Ivan]

Pulls to the Right....

One of the odd things about my P-47D-23 is that it has (now had) a tendency to bank and pull to the Right.
Banking and pullng to the Left is expected as a result of Engine Torque effects.
I had noted that this was happening years ago when I last worked on this aeroplane and never did find a solution to what had to be a broken AIR file.

A couple days ago, I was waiting for something and had nothing better to do, so I decided to list the possibilities as to what could possibly cause this situation and how to go about testing and finding the problem. I am not claiming any great knowledge of what can be done in CFS, but here are the things that occurred to me:

1. There are some Aerodynamic effects in the AIR file that are asymmetrical that should not be that way.
Years ago, when I first began working with AIR files, there was a very involved discussion about what could be done in CFS to simulate Propellers that rotated in a direction opposite to what we have come to expect.
It is even mentioned on the Game Box that this issue was addressed though I had never found an example of a Flight Model that actually appeared to have a reverse rotation Proeller.
Many thing were suggested. None of them actually worked. Sometimes the resulting changes were so damaging that the simulator would not load them.

There was the possibility that there are some side force parameters that are incorrect
The Engine location may have been specified incorrectly.

2. The Weight of something on the Aeroplane was offset to the Right.
Perhaps a Fuel Tank had a lateral offset. Both tanks on the Thunderbolt should be on the center line.
I could not think of any other weights that could be adjusted in the AIR file.
Later it occurred to me that the AIR file was not the only place where weights could be configured.
The DP file might have a Gun Position set incorrectly or be defaulting to a Bomb carried off center.

3. Also in the AIR file are many Mach effects tables. By themselves it was unlikely that they would be specifying a directional force to one side but perhaps there was a factor that was being amplified by another asymmetrical force such as Engine Torque or Rolling moment.

The P-47D-23 Razorback AIR file is almost identical to that of the P-47D-27 Bubbletop.
The first test was to see how the D-27 behaved.
It did the typical "Roll to the Left" thing.
That told me that if the issue was in the AIR file, it wasn't going to be something elaborate but a simple fat-fingered parameter somewhere. There are SOME differences between the two versions, but this was a pretty good indication that the problem was probably not in the AIR file.

The next test didn't even require leaving the simulator.
If the problem was in the ammunition load, firing it all off should correct it.
Thunderbolts carry 400 rounds per gun so a continuous burst at the invisible Gremlins lasted a bit over 30 seconds.
When the ammunition was gone, the D-23 slowly Rolled to the Left.

Next step was simple: Figure out whether the Right side was loaded with super heavy ammunition or whether the Left side had super light ammunition.
The result was that the ammunition in the Left Wing had Zero Weight.
It sure is amazing stuff! Must be made of Unobtanium.

- Ivan.

 

[Ivan]

This was Supposed to be SIMPLE!

Hello All,

It has been an amazingly busy month!
After correcting the ammunition weight in the DP file, I thought I was close to done.
The basic handling of the D-23 Razorback is actually fairly good, so I figured that updates to the AIR file would be fairly simple.
These turned out to be famous last words yet again.

This was one of those cases that I found that my knowledge of AIR files is somewhat inadequate to make the adjustments that I wanted. I found out that some of the values that I was "Tuning" really had no effect at all. The Wing Center of Lift appears to be a FS98 vestige and while the later simulators use Record 1534 for the same purpose, none of the stock CFS AIR files have this nor do they respond when such a record is added. There are obviously other ways that I can attempt to create the same effect, but those are a bit more complicated and would require some experiments and creating a spreadsheet for the calculations.

The DP file also turned out not to be quite as simple as I had thought.
Many sources list the P-47 Thunderbolt as carrying 8 x .50 Cal M2 Machine Guns with 425 rounds per gun.
I was finding that the weight totals in my spreadsheet were coming up way too heavy and traced it down to ammunition load.
As it turns out, the "standard" ammunition load tended to vary depending on the aircraft model but to match the maximum ammunition weight (which also included ammunition considered to be Alternative Load) as listed in America's Hundred Thousand, each gun would only be loaded with 275 rounds.

The DP file was modified for 275 rounds as default with 425 rounds optional.
This default ammunition load was used for performance testing.

Tuning the Roll Rate also turned out to be unexpectedly difficult.
While I could easily adjust the Roll Rate, I had a rather tough time getting the Rates to match the Airspeeds as specified.
It got bad enough that I spent an evening working on a spreadsheet to predict the roll rates at each speed.
With the numbers I got from that spreadsheet, I found that I had to plug in numbers that were entirely non-intuitive in order to get the performance that I was looking for. (The numbers were really not exact but more an indication of where to go.)

The next thing I found was that the DP file weights for Rockets and Bombs were also quite a bit different from what they should be.

In order to address some of the issues I was finding in the AIR file, I started adding a bunch more record descriptions to the FDE Control File. That ended about a week ago when FDE complained that the control file was too complicated and barfed. I had to revert to the last saved FDE Control File and now some of the more obscure records are being edited with AirEd which doesn't do well with conversion factors.

At this point, many of the issues have been addressed if not resolved and performance and handling are about 90% of where I want them to be. The technicians from Ivan's Propeller Shop have been a bit busy elsewhere, so there won't be any work on a new Propeller for any of the Thunderbolts for a while.
Propellers are tedious to build, but the process is pretty well defined. I believe time is better spent in trying to pin down things that are not quite so well defined.

Specifications and Performance thus far (for the P-47D-23):
Empty Weight ----- 9956.7 pounds
Basic Weight ------ 10545.1 pounds
Gross Weight ----- 13582.1 pounds

Zero Fuel Weight - 11003 pounds

Speed @ 500 feet ------ 328 MPH (Military Power)
Speed @ 25,500 feet -- 433 MPH (Military Power)

Service Ceiling ---------- 41,600 feet

The performance will most likely change slightly if a new Propeller is installed.

- Ivan.