Analysing and modifying the AFX file with QBasic.

[Ivan]

Hello Aleatorylamp,

Was that "Futuristic Movie" Soylent Green?
If so, it wasn't "soy substitute".....

The Monogram P-39 is not a rare kit or expensive or even out of production.
I believe you can still find it mail order or in stores for about $15 or so.
They more modern pressings are in a Medium Gray plastic while mine is a very Dark Green.
In fact, I believe I actually have at least two of these kits along with other Airacobra kits in 1:72 scale.
The only Airacobra kit I have ever completed was a Revell 1:72 when I was very young. It was in a Medium Green plastic which is much lower quality than the Monogram material. I believe the two companies are now merged as Revell-Monogram though some kits are obviously of one or the other company's legacy. Revell Germany is a different game but I don't have a large sample of their kits.

I picked the Monogram 1:48 P-39 to look at because by reputation, it is supposed to be one of the more accurately scaled kits.
I did detail painting on my kits but not camouflage because I didn't like the idea of hiding the shape because that was what I was most interested in. I do need to find the large set of model paints I have accumulated over the years.

Regarding the Console and Radio:
My Console is wrong and the Radio may not have actually been there. It isn't shown in the box art for the Monogram kit nor is it shown in the reference drawings I am using.

We just got some unexpected and unannounced visitors, so I will continue this later.

- Ivan.

[Ivan]

Another thing:
I remember that a while ago, we discussed that my CoG positioning was within feasible limits,
and I saw you addressed the subject on your Airacobra thread a couple of days ago.

Your reference to the Datum Line is presumably the one on the diagram in your post:
3 inches (0.25 ft) forward from the Tip of the Spinner - the point your 11.48 ft measurement to the CoG starts from. This would then be 11.23 ft from the tip of the Spinner.

In my case, I have the CoG at 11.47 ft from the Spinner tip - just under 3 inches further back, which is acceptable for my purposes.

Vertically, I have 0.9 ft (10.8 inches) below the engine thrust line. Compared to your estimate of 0.75 ft (9 inches), I´m only 1.8 inches lower than that - so that´s fine too!

Hello Aleatorylamp,
Perhaps I should be responding to this in the Airacobra thread, but there would not be a context for a response.
Something is very strange here. I don't think I have ever used 11.48 feet from the Datum or from the Spinner Tip as a CoG estimate.
The mistake I had made earlier was in reading a diagram as indicating the Length of 30 feet 2 inches was from Spinner Tip to Rudder.
Other diagrams indicate that the 30 feet 2 inches is actually from the Datum to Rudder.
Since the Datum appears to be 3 inches Ahead of the Spinner, that really makes the Length from Spinner to Rudder 29 feet 11 inches.

My original estimate for CoG was 11.25 feet behind the Datum which I assumed to be the Spinner Tip.
It was first revised to 11.40 feet and then to 11.35 feet from the Datum which is the number I will use for my own project.
As for Vertical CoG, my original estimate was 9 inches (0.75 feet) below the Thrust Line.
That has not changed.

The Vertical CoG and even the Horizontal will vary a bit depending on if the Landing Gear is Up or Down, but we can't really simulate that very well in CFS. You can of course use whatever numbers you like. I can only tell you my best guess at the moment.

By the way, I did notice the little spur that appears on the Pitot Tube in your more recent screenshots.

Time to go do some more editing on the Eric Johnson model. This time the edit looks to be pretty extensive.
It is a difficult repair to do without either fixing the Gull Wing shape that isn't supposed to be there or leaving the weird shape in place and just connecting the dots in a better fashion.

- Ivan.

[aleatorylamp]

Hello Ivan,
Thanks for your answer! My brain is doing quirky things.
I should have answered on the other thread, and also, you are quite right,
you never mentioned 11.48 ft, but 11.35. I should have quoted:

"Longitudinal CoG for the model will be
Fuselage Station 136.2 or 11.35 feet from the Datum line."

...then I wouldn´t have made yet another mistake. Actually, at the end it´s
better that the mistake-mess is kept here on this thread!

With this new calculation, the CoG should be 11.1 ft from the spinner tip.
With my 11.47 ft, it would make the difference in my case a bit greater:

I´d be 0.37 ft or 4.44 inches out, so I will most probably be letting my
Qbasic Modifier Program have another go at this!
That should get a little adrenalin flowing again!

About the scene in the film - now I remember! It WAS Soylent Green, and
of course, it WASN´T soy substitute! Uffff...

So the Gull-wing aspect of the trailing edge would be erroneous. Interesting.
Perhaps the upward slope to the fuselage from the trailing edge should be flatter.
I´ll also have a look there.

Update: It seems the fuselage component parts in the aft wing-fillet area, condition
the way the fillet is built. The contour of the fuselage-joint with the wing-fillet on the
model is too horizontal in the rear, and should curve downwards somewhat.
That way the trailing edge at the joint would be lower and also a bit tucked in.
The lower aft-fuselage on the model is too straight, and should be a little rounded.
Adjusting the lower parts of the different fuselage-slices, would pull down the vertices
to which the aft wing-fillet vertices are connected, and might well do the trick!

Cheers,
Aleatorylamp

[aleatorylamp]

Hello Ivan,

The CoG shift of 4.44 inches forward and 0.15 inches upwards went well.
Nothing broke, although the tachycardia still lasts!

Now I´ll shift the textures into place and redo the animations, but that will be quite fine.
Then I´ll see how it goes with the lower aft fuselage to see how the wing-fillet shape improves.
Thanks for the tip!

Update: The AF99 texture-mapping went quite quickly (quicker than writing a
Texture-shifting QBasic routine!), and the animation was no big deal.
Now the CoG as located as per your calculations, thank you very much!
Let´s see what happens with the wing-fillet now.

Cheers,
Aleatorylamp

[Ivan]

Hello Aleatorylamp,

Be careful about taking the numbers that I am using as factual. They are ESTIMATES.
I just realised that although I told you about the Vertical CoG, I didn't really mention the Longitudinal CoG except to say that what was there was pretty close.
I continuously revise my numbers as I work.
The Vertical CoG estimate of 0.75 feet below the Thrust Line has not changed since I started but that is not because I really did a lot more than very quick observations with diagrams that I have already posted in the Airacobra thread.

Be CAREFUL with those drawings and diagrams because I can tell you that they are contradictory.
I just had to pick one that I believed more than the others or combine pieces of data from multiple sources.

So far, my original Longitudinal CoG has gone from
11.25 feet to
11.40 feet to
11.35 feet.

The Datum point has changed though!
To be honest, these numbers don't really matter all that much as long as the other locations of weights are in agreement.
I believe there are other factors that are much more important and they are mostly in the AIR file.....

- Ivan.

[aleatorylamp]

Hello Ivan,
Don´t worry, I know the vertical figures you mentioned are estimates.
I suppose it wasn´t absolutely necessary to make the changes, but I thought it would
be a good idea to update my CoG, as it was 1.8 inches too low and 4.44 inches too far aft.

As I said, there weren´t any complications with the model shift or texturing re-mapping.
Although I can always scrap the changes and revert to the previous build, there doesn´t
seem to be much point because it´s more accurate now than before.

P.S. If the Datum line stands at 3 inches ahead of the spinner, and the 30 ft 2 inches is the length of the plane, it would be totally absurd to suppose that the plane length goes from a place in the air infront of the spinner, to the rudder-end at the back.
...Unless the Datum line is where it is for a specific reason: Would that be because it includes the short barrel of the 37 mm cannon? Holy Cow, this is going round in circles!

There is a .pdf document about how this plane was built,
http://legendsintheirowntime.com/LiT.../P39_draft.pdf
originally published in the May, 1943 issue, Volume 42, number5, of Aviation magazine, published by McGraw Hill. It has a diagram on page 4 where the length includes the short barrel.
Well... once we know what´s right, we can always shrink the whole model again, improvements and all included - so it´s not really worrying.

Cheers,
Aleatorylamp

[aleatorylamp]

Hello Ivan,
I checked on some of the drawings and saw that the lower aft belly contour is hardly as straight as I´d had it. This was actually inherited from the original model - that´s why your wing-fillet trailing edge is a bit high high too.

I measured it at about 0.3 ft. and lowered that point on the fillet end and on the fuselage-belly, adjusting all adjacent pieces, to see what would happen.

Here are some pictures and a blueprint - Perhaps it could even be made a bit more pronounced, but I decided to be conservative.

The texture isn´t corrected yet, but one can see that the trailing edge descends more than before. The nook between the two triangular sections that conform the concavity at the rear wingroot, is also proportionally lower. Putting in an extra triangle in that area, it bould go a bit lower still, perhaps.

Cheers,
Aleatorylamp

[Ivan]

P.S. If the Datum line stands at 3 inches ahead of the spinner, and the 30 ft 2 inches is the length of the plane, it would be totally absurd to suppose that the plane length goes from a place in the air infront of the spinner, to the rudder-end at the back.

...Unless the Datum line is where it is for a specific reason: Would that be because it includes the short barrel of the 37 mm cannon? Holy Cow, this is going round in circles!

There is a .pdf document about how this plane was built,
http://legendsintheirowntime.com/LiT.../P39_draft.pdf
originally published in the May, 1943 issue, Volume 42, number5, of Aviation magazine, published by McGraw Hill. It has a diagram on page 4 where the length includes the short barrel.
Well... once we know what´s right, we can always shrink the whole model again, improvements and all included - so it´s not really worrying.

Hello Aleatorylamp,
Actually for a Datum Line, this is may be absurd, but that is how things are done.
Note that the VERTICAL Reference is 40 inches ABOVE the Thrust Line.
There is Nothing there at all, but that was the chosen reference.
I don't know who picks these references, but they own the design and I do not.

In the case of the Airacobra, the Longitudinal Datum is not hanging out in space in MOST aeroplanes
It is part way up the Cannon Barrel.
Of course if no Cannon was mounted (and plenty of the early aircraft did not actually have a Cannon installed), then nothing was ahead of the Spinner Tip.
You also Paul Matt's Drawing that show the Overall Length including the Cannon as 30 feet 3.4 inches which is a touch longer than the 30 feet 2 inches normally quoted. I believe this difference is because it includes the full length of the Cannon Barrel. This is consistent with another description I have seen that mentions that the extra length is 1.5 inches.

I have read the article from Legends in Their Own Time, but I have also read the article that it is based on.
There are actually some loss in precision because of rounding in both of these articles at least compared to some of the other drawings I have already posted.

Perhaps it would be a good idea to post this response also in the Airacobra thread. It may clarify a few things for other folks as well.
I figure if you are working with the model and didn't understand what I was trying to express, others probably also didn't understand.

- Ivan.

[Ivan]

Oops, I almost forgot:

The little "Belly Bump" is clearly visible in the Monogram P-39 model.

I also did quite a bit of work on the Eric Johnson P-39D last night.
I can tell you my idea for the Nose Gear / Propeller bleed now that I know it works:
Put the Nose Gear in the Nose Group.
The way to avoid bleeds is to Start the Glue Sequence with the Nose Wheel.
Sequence (simplified) would look something like this:
Nose Wheel
Fork
Strut
Brace
Nose Gear Well
Nose Gear Doors
Nose
Nose Guns
Spinner Back
Propeller
Spinner Front
Cannon

As you can see from the First Screenshot, it seems to work.

I also fixed the Wing Root Fairings at both Front and Rear.
The main problems here are that:
1. The Fuselage is not really at the correct angle which should no longer be a problem with your model.
2. The Wing Root Station at the outboard side of the Root Fairing is shaped very poorly and the Trailing Edge is too low relative to the Fuselage. The choice here would be to have the Fillet rise as it meets the Fuselage which gives a Gull Wing appearance or to make it horizontal (my choice) and do blending closer in and also adjust the Fuselage.
3. At the Front End of the Fillet, the intersection with a very angular 8 sided Fuselage section so there isn't an easy match.

This is also shown in the Second Screenshot.

I reworked the Flaps the best that I could but the problem is that the underside of the Wing is not Planar and the Flap is not planar.
To fix it would require a full rebuild of the Wing which is way out of the scope of this project. I will try my ideas of how things should work when I build my own project.
The general appearance is shown in the last two Screenshots.

The result still isn't too bad in my opinion. There are a couple momentary bleeds especially when looking at the Flaps and Flap Well from outboard. This assembly method works well enough that I will use it for my project.

- Ivan.

[aleatorylamp]

Hello Ivan,
Thanks for clearing up the measurements and the info on the Datum Line
(strange, where they put it, but OK!). So the cannon itself could well be
4.5 inches long, apart from the restly 30 ft 2 inches length.

And... Wow! Congratulations on figuring out a good, clean way of fixing the
other conflicting areas on the build - The nose gear fix looks really great,
both the results and the sequence. Looking at it, it looks totally logical!

I had also been thinking whether raising the point on the inner wing leading
edge, would straighten things out a bit more, but hadn´t got as far yet.
It sounds useful, so I´ll do it!

Right now it´s rather late, but tomorrow I´ll get my hands on the job again
and put all this into practice. I´ll also see if I can fix the wing with some
extra triangles - although it´s not terribly bad.

I thought the belly bump would be visible on your Monogram model! Very good.

This model´s going to come out great!
Cheers,
Aleatorylamp

[Ivan]

Thanks for clearing up the measurements and the info on the Datum Line
(strange, where they put it, but OK!). So the cannon itself could well be
4.5 inches long, apart from the restly 30 ft 2 inches length.

No, that is not what I meant!

The Total Overall Length is 30 feet 2 inches Plus 1.4 inches more of Cannon Barrel if it is a 37 mm Gun.
OR
The Total Overall Length is 29 feet 11 inches Plus 4.4 inches of Cannon Barrel sticking out past the Propeller Spinner.

....So the length of the Spinner to Rudder is 29 feet 11 inches by that calculation.
Of course I could simply be wrong, but that is how things look at the moment.

Don't get too hung up on the exact length of the Cannon Barrel sticking out of the Spinner.
The barrel moves a lot under recoil.

The Wing Leading Edge isn't a real problem. The Wing TRAILING Edge is a problem.
There is also a twist in one of the Wing panels that causes it to disappear from shallow angles.
Many of the Wing Fillet Parts are also non planar which also cause them to fade at certain angles.

I wanted to post a few screenshots to show the issues with the Flaps (which I do not consider severe) and the compromise with the Nose Gear Doors, but the Game Machine does not have any Graphics Utilities at the moment besides MS Paint.

- Ivan.

[aleatorylamp]

Hello Ivan,
I´m so sorry, I did mean to say trailing edge, not leading edge!
Regarding the Overall Length issue:
Very confusing... but OK, I´ll ignore it for the moment.
Cheers,
Aleatorylamp

[Ivan]

Hello Aleatorylamp,

You are of course free to do as you wish.
You asked about the Length and I gave you my opinion based upon what documentation I have seen.
It almost fits together in my mind but there are still a matters I am unsure of.

Last night I did a minor rework of the Console area. That was not hard.
I also reworked the Trailing Edge of the Wing Fairings yet again.
That was difficult, but mostly because of the naming convention of the Parts and not because the task itself was very difficult.
I could see the polygon I wanted to change but it would often take a few tries before I could figure out the name of the file I wanted.

I believe it was an improvement overall, but I am still not satisfied.
This part of the rework is important not because I believe the model is really good (in fact, I am finding more things that are wrong), but because the techniques here will be transferable to the real Airacobra project.

I would post some screenshots, but I can't figure out where I left my Flash Drive.
Image files are sometimes not easy to find because the clock on my "new" Development Computer does not work.
I believe the battery is original so it would be a bit over 15 years old by now.

- Ivan.

[Ivan]

Here are a couple of the screenshots I was describing earlier.

The first screenshot shows the newly constructed Wing Fillet ends.
The shape isn't as good as it could be because the Fuselage isn't shaped correctly there and reworking an 8 sided Fuselage is not part of the plan.

The second and third screenshots show how the Flap Wells sometimes vanish. This is something I can accept because the alternatives seem to be worse.

The last screenshot shows the Nose Component in the Assembly Blueprint.
Note that the Nose Gear Doors are also cut to match the underside of the Component.
The rear sections of the Nose Gear Doors are in the Gear Center Group.

- Ivan.

[aleatorylamp]

Hello Ivan,
A dead CMOS battery is of course a bit off, but at least the computer works Modern ones won´t even start with a flat battery.

I had written a lengthy reply explaining that I had understood just the opposite of what you meant about the overall length, and then started arguing against the uselessness of seemingly accurate information from so many sources being so misleading.

What should be a straight forward measurement is based on some obscure Datum Line. Imagine reading a specification in sheet:
- Overall length, from 3 inches infront of the spinner to the end of the rudder: 30 ft 2 inches. (!!!??)
But OK, that´s the way designer engineers do it. They want to be mysterious, hoping to increase the awe others feel towards them.

Then I scrubbed the comment, and said I would ignore the overall length issue for the time being.

Originally Posted by Ivan
You are of course free to do as you wish.
You asked about the Length and I gave you my opinion based upon what documentation I have seen.
It almost fits together in my mind but there are still a matters I am unsure of.

It was not meant as a rebuke, only as an explanation of what I was going to do, because you had not yet reached a definite conclusion, and I doubt whether it would be a good idea to multiply the model by 0.0982 to get what we had before, redo all the textures and animations again, before the present length of 29 ft 11 inches is confirmed as definite.

Anyway, your aft wing fillet and trailing edge look quite correct now. Mine has come out looking very similar. The re-shaping here also improved the shape of the flaps, that are more planar now.

I also triangulated two of the undersurface wing-panels, and this stopped the disappearances.

OK on the tip of having the nosegear doors with a rear part in Gear Centre. Thanks! It seems logical, as they go back so far.

I haven´t started on the Nose section yet, but hope to do so tomorrow.
Cheers,
Aleatorylamp

[Ivan]

Hello Aleatorylamp,

Perhaps I misunderstood what you meant. I did mean what I stated though.
Even when I finish a model and everything falls into place, things are seldom entirely certain.
In this case with so much contradictory evidence, one has to go by a best guess unless you happen to have a full scale Airacobra or a set of blueprints to work from.

Considering where we both started from with the EJ P-39D, there is NO WAY we will arrive at a pa
rticularly accurate model unless we basically throw the original away and start from scratch and even then we have the AF99 limitations to work with.

Regarding the strange Longitudinal Datum:
This discussion will be a little bit of fact but mostly theory and a few guesses based on evidence.

Picking a Datum Point is a difficult thing.
Everything literally revolves around it and changing it at a later date is difficult in Aircraft Factory 99 with a thousand or so simple polygons.
Changing the Datum Point for a REAL aircraft would be pretty near impossible. Imagine how many drawings exist and for how many pieces.
This aeroplane was not designed in the age of computers.

The P-38 Lightning that I worked on a while back had its Datum (Fuselage Station 0) at the Tip of the Nacelle.
For the most part, the Shape and Dimensions of the Lightning did not change during its life. There were exceptions though.
The "Droop Snoot" versions I believe had a longer Nacelle to house a second crew member in place of the guns.
In the case of those planes, there would have been Negative Fuselage Stations.

Now here comes the Theory part:
In the case of the P-39, although it had been designed from the start for a cannon through the Propeller Hub, the prototypes did not have a cannon there. They had a Spinner that came to a Sharp Tip. My guess is that the Datum was established with the original design for a Spinner or perhaps it extended to the end of the Cannon Barrel as it was at that time?
When the Spinner was made shorter, the Datum would not have been adjusted.

Take a look at the Wire Frame on my EJ P-39D. What do you think of the general shape and appearance?
It is basically the same as yours except that yours has a 12 sided Fuselage and mine has a shortened Spinner and a shorter and thicker Cannon Barrel. That was enough to correct the length dimensions.
CpG os still at 11.25 feet from the Datum and I am not about to correct it because I would need to do the Animations again.
As for the other dimensions, this model is full of errors and shape problems, but we already knew that.

- Ivan.

[aleatorylamp]

Hello Ivan,
I just had a thought: Wingspan is doubtlessly 34 ft.
On your plastic Monogram model, the physical wing measures a certain number of millimeters.
Now, how many millimietres does the fuselage measure, spinner-tip to rudder-end?

That will give a proportion - who needs more?
I gather Revell-Monogram have a respectable reputation for accuracy,
so that could possibly be a help in our situation.

Cheers,
Aleatorylamp

[Ivan]

Hello Aleatorylamp,

I actually don't know how much I would trust a measurement of a plastic model.
If you check around, you will find that they tend to vary in accuracy and usually a millimeter or two doesn't seem to bother people.
I also don't have the tools to measure to that level of precision with something that big.
My Dial Calipers measure to 0.001 inch but have about a 6 inch limit.
My Micrometer can measure to 0.0001 inch but only has only a 1 inch limit.

That reminds me of a comment made by an old friend of mine. He was a Colonel in the US Air Force.
"A Ten-Thousandth (of an inch) is an O-PIN-ion....."

Consider what we might actually get for precision.
If we are accurate to 1/2 millimeter in measuring a 1:48 scale model, then the accuracy when scaled to full size is 48 times 1/2 mm or 24 mm or just under 1 inch.... The level of accuracy we are actually going for is about 1/8 inch.

In the past, Monogram has done well with some kits and poorly with others. They have a Razorback and a Bubble Top Thunderbolt. One is excellent according to reviews while the other is not so accurate. Both kits look good to my untrained Eyeball and I forget which is the good and which is poor.

If you are looking for that kind of confirmation, I have already done that a couple times with the Monografie Lotnicze drawings.
When I scaled the plan (top) view down (yes, Down....) so that its Wing Span was 3400 pixels, the Length was 2992 pixels from Spinner to Rudder. Perhaps it was a pixel off, but it was very very close and considering each line is about 4-5 pixels wide even scaled down from an image that started off at about 5500 x 7500 pixels, that is pretty good.

- Ivan.

[aleatorylamp]

Hello Ivan,
OK, thanks a lot. The proportion comes out to 0.88, i.e. 29 ft 11 inches.
The highest resolution Paul Matt drawing I had got hold of, gives quite a
few inches less, as does the measurement on the original E.J. AFX.

So much for the 30 ft 2 inches then. I´ll run the Modifier again. Thanks again!

Update: Now there´s the supposition: As the model will come out 3 inches shorter, the distance from the Datum Line to the CoG will be somewhat shorter, so the CoG will have to be re-located so as to be at 11.35 feet from the Datum line.
Update 2: OK! With a multiplier of 0.991932, and shifting 0.08 ft, it´s done, done, done:
Spinner tip is at 11.10 ft from CoG, (Datum Line is 3 inches, or 0.25 ft, infront of Spinner-tip).
Rudder-end is at 18.82 ft.
Thus, if my math is the same as yours (as a work-mate of mine used to say),
overall length is 11.10+18.82=29.92 ft (29 ft 11.04 inches)

Well, that should take care of accurate dimensions.
Cheers,
Aleatorylamp

[Ivan]

Hello Aleatorylamp,

There still seems to be a serious communication problem here.
Either I am expressing myself poorly or you are not interpreting what I am stating correctly or both.
I know that at times I have no idea of the message you are trying to convey.

Here is one of those times:

OK, thanks a lot. The proportion comes out to 0.88, i.e. 29 ft 11 inches.
The highest resolution Paul Matt drawing I had got hold of, gives quite a
few inches less, as does the measurement on the original E.J. AFX.

WHERE did the 0.88 number come from? 29.92 feet / 30.16 feet = 0.992.

Update 2: OK! With a multiplier of 0.991932, and shifting 0.08 ft, it´s done, done, done:

How does the Paul Matt drawing give a smaller dimension? Are you working from a different Paul Matt drawing?
Mine shows Overall Length as 30 feet 3.4 inches to the Cannon Muzzle and that is about it for length dimensions.
That number is consistent with the discussion thus far.
The original dimensions of my drawing are 2120 x 1632 pixels so it is not a very small drawing.

Regarding the EJ AFX, I simply would ignore any dimensions implied by the model.
A lot of other dimensions such as Wing Span, Propeller Dimensions, probably Wheel Track are wrong.
Proportions sort of look like an Airacobra but are pretty far off if you start looking at specific details as I have been doing.

Here is another case of a serious miscommunication.
This is what I stated back at Post 191 "Theory and Conjecture" which is only about two posts back:

Take a look at the Wire Frame on my EJ P-39D. What do you think of the general shape and appearance?
It is basically the same as yours except that yours has a 12 sided Fuselage and mine has a shortened Spinner and a shorter and thicker Cannon Barrel. That was enough to correct the length dimensions.
CpG os still at 11.25 feet from the Datum and I am not about to correct it because I would need to do the Animations again.
As for the other dimensions, this model is full of errors and shape problems, but we already knew that.

What I was trying to tell you was that all I did on this model was to shorten the Spinner to correct the Length.
I didn't re-scale anything!
Heck, I didn't even bother moving the CoG because it was only 0.10 feet off from where I wanted it to be.
There are errors of more than 1.2 inches in other places in the model and the direction of this error brings it closer to a conventional configuration.

- Ivan.

[aleatorylamp]

Hello Ivan,
I have no idea what it is, that is so confusing now.

The proportion 0.88 is directly related to your post prior to mine, and to mine before that, about the proportion between wingspan and fuselage length, when I asked about your plastic model to this respect.

You had written about the Polish drawings, that you felt were the most reliable:
"...so that its Wing Span was 3400 pixels, the Length was 2992 pixels from Spinner to Rudder. "

Dividing 2992 by 3400 I came up with the proportion of the fuselage being 0.88 of the wingspan.

Multiplying 34 ft by 0.88, and if my math is the same as yours... we get 29.92 ft, which is 29 ft 11.04 inches... basically identical to the 29 ft 11 inches you had commented as being the most reliable data for the length from Spinner tip to rudder end.
The wingspan on my model is exactly 34 ft, so only the fuselage length is wrong.

I had originally multiplied the whole E.J. AFX model by 1.0191, as you had re-scaled your model, and it came out to what we had talked about before, 30 ft 2 inches from spinner tip to rudder end. This had been discussed almost at the beginning of the modifications and improvements needed for this model.

To achieve the now more correct 29 ft 11 inch overall length, I had to multiply my current model this time by 0.991932, and to move it so that the CoG was correctly placed again.
Obviously, I was not going to use the original E.J. Model for this, and re-work all that another time.

Incidentally, there is another difference between your model and mine apart from the 8 or 12 sided fuselage, namely the fatter belly.

My mentioning the original E.J. AFX and the Paul Matt drawings was only to comment that the fuselage compared to the wingspan, came out quite a few inches shorter - about 4 and 4.5 inches,
the proportion being 0.86 or 0.87 or something, I think it was, but I´m not about to measure it again.
It was only meant as an example, to illustrate how inaccurate THEY had been.

I will study your comment about the spinner length now.
I had no idea that the difference in length on the model was solved only by altering the spinner. The two comparative superimposed drawings you posted right at the beginning, refer to a change in the overal length, not just a spinner adjustment.

Anyway, while I was correcting the model, lowering the tail and the underbelly, putting in the transparent canopy, I placed all the key elements in the correct positions as per Paul Matt´s drawings, so canopy, air intake, fin, spinner, propeller, spinner tip, and everything else is shaped and dimensioned proportionally.

Here´s a blueprint-screenshot showing the current model, with its correctly dimensioned fuselage.
I can´t screen-shot the airplane yet because it needs texture-adjusting first.

Cheers,
Aleatorylamp

[Ivan]

Hello Aleatorylamp,

I understand now.... Thanks for the explanation.
I believe you also see why I chose to use drawings other than the ones by Paul Matt.
The Monografie Lotnicze drawings are also far from perfect (The Landing Gear doesn't look right), so it is a matter of piecing together something that can be used.

I am still wondering why you are doing a re-scale of the model after you re-constructed the 12-sided Fuselage with reference to the Paul Matt drawing. When you do that, all that work would be wasted.
If the dimensions of your model are the same as where mine was when I started, then all that is really necessary is an adjustment to the Spinner which is the wrong shape anyway.

..........

Last Night, I did nothing on the actual 3D model. Instead, I did quite a bit of editing of the Flight Model.
Most of it was edits to the stability and control graphs and the end result was that the general handling was much improved.
In fact, in some ways it is behaving better than it should, so I need to undo some of those changes.

The net effect was that it now handles much better than the stock P51D that it was derived from, but isn't quite as nice as the P-40E.
I also found that the Sea Level speed seems to have dropped slightly (about 1 MPH) and the speed at altitude (12,500 feet) has dropped by 6 MPH. (Really!) I believe it might be time to remove the Backfire Screens, but not before I replace the Propeller.
Otherwise I would have to do the tuning twice more instead of just once more.

Attached are some screenshots.
The original flight model was difficult to fly at low altitude but this has been greatly improved.
The Landing Flare shows multiple shots because I was testing Elevator control authority which appears to be too low.
The first Flare image shows an airspeed that converts to 92 MPH which is barely above where the stall is supposed to be at 90 MPH.
More testing and some adjustments are required there.
Rudder control authority also appears too low but that is hard to show in a screenshot.

- Ivan.

[aleatorylamp]

Hello Ivan,
OK. Well, at least there´s no confusion now, and the airplane is more correctly dimensioned.

Thanks for your concern, but I believe the situation does not warrant so much worry!

If Paul Matt´s drawings are not accurate anyway, and the pixel proportion to the wings on it
is 770/887 = 0.868, as opposed to the 0.88 proportion as on the Polish drawings, I don´t think
there could be significant discrepancies after having applied a 0.991932 multiplier to make the
fuselage length 3 inches shorter.

Everything will have exactly the same proportions as before, only along an overall length
of 3 inches shorter, so I really don´t see the problem.

The Spinner measures 1.38 ft from base to tip now. Perhaps you could check on the Polish
drawings and tell me the measurement there. Before my re-adjustment with the 0.991932
multiplier, the spinner measured 1.39 ft.

I´m glad the .air file is turning out fine. My criteria is less reliable, but putting in data from
other models into certain parameters, the behaviour looks OK to me - which obviously
doesn´t mean it will be OK to anyone else!

Well, now I´ll adjust the textures and the animations again, and that is just as
easily said as it is done!!

The model IS coming out very well!

P.S.
Regarding your fixed P39: At the very beginning, I had already commented on the
great improvement on the primitive model after your simple fixes. Even if you say the
model as such is still inexact, with your subsequent improvements and the transparent,
inhabited cabin, it has come to look rather impressive!

Cheers,
Aleatorylamp

[Ivan]

Hello Aleatorylamp,

I try to warn you if I believe you are making an error. I have done that.
Sometimes my opinions are correct. Sometimes they are not.
It just happens to be the best information I have at a particular time.
I figure if your dimensions were correct with the model you WERE working on, then they are no longer correct.

As I mentioned before, the Monografie drawings have their inaccuracies as well.
I can see some of the mistakes when comparing to photographs.
The length of the Spinner from one of these drawings is 1.32 feet.
It might actually be 1.31 or 1.33; The drawings are not that precise.
In theory the Station Diagram should be useful, but it is not. That is one of the areas that I believe is VERY wrong.

The actual Overall Length I am using is 29.91 feet -- +11.10 feet to -18.81 feet.

The AIR file is hopefully getting better, but there are at least three or four things that I need to learn how to do.
My deal is trying to develop by a repeatable process with minimal backtracking and be able to justify everything I am changing.
That is why I was discouraging fine tuning of performance until the very end because there is no point in doing it over and over again.

As for developing AIR files, do whatever works for you. This is why I have not been discussing specifics very much. I want to see if we both come up with something similar while taking different paths.
You will notice that I am much less reluctant to discuss specifics about the AF99 model.

I was looking at the model in the screenshots as I took them and although the features don't look bad, the proportion of the model don't look quite right. (At least mine doesn't.) The Canopy definitely looks too large.

Time for me to do more work on the new project, but first I need to have the AIR file in a reasonable state and finish up the Pilot.

- Ivan.

[aleatorylamp]

Hello Ivan,
Thank you for your concern, but in this case I think there is no need to worry.

I used Paul Matt´s drawings for the basic layout of the airplane, as they looked like having more correctproportions than other drawings of "slimmer" P39´s that the E.J. AFX seem to have been based on.

The resolution was the highest I could find, and they also seemed more correct than the Wikipedia diagrams that looked so clean. The Spinner in the Paul Matt drawings is 1.38 ft long.

Incidentally, similar to your Polish drawings, the spinner length on the original E.J. AFX is 1.31 ft, but the spinner-tip/rudder-end length is only 29.6 ft, falling short by 0.31 or 0.32 ft.

I don´t think I want to change over to any other drawings, because that would involve re-building everything manually, and who knows how accurate other drawings will be in that respect.

Ideally, would be a set of exactly 90-degree 3-view photographs, taken with a camera lens setting that won´t give perspective distorsion, but that´s too much to ask.

Initially I (we) scaled the length to 30.16 ft from spinner to tip, and that has now changed to 29.92, or 29.91 ft. In my case I have the spinner tip at the same position as you do, and the rearmost point at just 0.01 ft more.

After re-scaling the model with the QBasic Modifier, if one were to re-scale the drawings, all the elements of the plane would by logic perfectly fit the new drawings, so the re-building work I have done is not wasted by any means. I don´t know how this could be different.

Incidentally, as vertical and horizontal scales on the drawings are the same, the height also has to be scaled down, otherwise shapes distort. I hadn´t got round to that yet, but I´ve just done it, and the wheels are perfectly round, which is a good indicator. The topmost canopy height went down by 0.03 ft, so it was necessary to do this.

Thankfully I hadn´t repaired the textures or the animations yet, as I had a 2-day migraine after eating about 14 fresh yellow plums on Monday from my brother-in-law´s plum trees, which I mistakenly expected not to have been treated with insecticides... Today it´s gone though.

As regards the E.J. Canopy proportions, I remember I had to make some adjustments there after I´d corrected the tail-empenage height. The sill was a reliable reference point, but that was before the vertical CoG correcting shift. At my moment, my Canopy "box" coordinates in side view are: 3.72 up, 2.06 ft down, 3.5 fore, and 4.24 aft.

My .air file is no big deal, but seems to work OK, although I have no way of measuring how good or bad it is, other than the intuitive sensations it gives.

Well, let´s see how the model continues developing.
Cheers,
Aleatorylamp