Flight model updates for 4.12
Work example for P-36 and P-40
General
Just like the previous patches, 4.12 will see some updated and some new flight models, mostly
planes that previously had been AI controlled only, or new planes. In the past we've often
been asked what we do when we create or change flight models, and up to now we haven't
given a detailed answer. This is simply because giving a sufficiently detailed answer is quite a
bit of work on its own, and we didn't find the time - until now.
4.12 will have the P-36 flyable. Having been an AI controlled plane up until now, the flight
model was very crude and needed an update. Since it shares the air frame with the P-40, an
update for the P-40 was thought to be mandatory as well, for consistency. With this example,
we'd like to share what the flight model making process typically involves.
Research
General
Extensive research is the first step with every flight model done or redone, and to give you an
idea, we've compiled an incomplete list of sources which we checked for the update of the P-36
and P-40 flight models. Most valuable are of course the primary sources, but also secondary
sources often contain good information, be it about operational use, field modifications or just
the bibliography. A couple of websites proved very useful, some of which have uploaded copies
of first hand sources. Many sources still remain unlisted, for instance hard to reference primary
snippets, or applicable general information, with very few specifics.
Typically we go through thousands of pages of varying quality, more for some planes that are
very well documented, less for others were sources are very hard to find. As you can imagine,
research can be very time consuming, so we are grateful for the help we sometimes receive
from the community, making hard to get sources available to us.
1st hand sources - test reports, manuals, technical specifications
technical order No. 01-25C-1 - handbook of operation and flight instructions for the P-36A pursuit airplane,
Nov 25th, 1938
technical order No. 01-25CB-1 - handbook of operation and flight instructions for the P-36A and P-36C
pursuit airplane, revised May 25th, 1940
technical order No. 01-25CB-1 - pilot's flight operation instructions P-36A and P-36C airplanes, revised
Feb 15th, 1943
technical order No. 01-25CE-1 - pilot's handbook of flight operating instructions for model P-40B and P-
40C, Jul 25th, 1942
technical order No. 01-25CK-1 - handbook of operation and flight instructions for the models P-40K and P-
40K-1 fighter airplanes, Aug 20th, 1942
AN 01-25CN-1 - pilot's flight operation instructions for Army model P-40N series British model Kittyhawk
IV airplanes, Sep 25th, 1944
AN 01-25CN-2 - erection and maintenance instructions for Army model P-40N series British model
Kittyhawk IV airplanes, Aug 30th, 1944
air publication 2012A pilot's notes - Curtiss H-75A-4 (P36) aeroplane Cyclone GR1820 -G205A Engine,
Oct 1940
air publication 2014A volume 1 - Kittyhawk I aeroplane, Aug 1941
air publication 2014A pilot's notes - the Kittyhawk I aeroplane, Jul 1941
pilot training manual for the P-40, 1944
operation and service manual - Wright Cyclone 9 aircraft engines, 1944
technical order No. 02-35GC-1 - handbook of operation instructions for the model R-1820-65 engine and
associated models, Oct 10th, 1942
technical order No. 02-5AD-1 - handbook of operation instructions for the model V-1710-35 engine and
associated models, Oct 25th, 1941
maintenance manual - R-1830 S1C3G Twin Wasp, Feb 1960
operators handbook - Twin Wasp C3 engines, Jul 1943
instruction manual - Pratt & Whitney engines R-1830-43 & 65, Nov 1943
handbook - operation and maintenance for Allison V-1710 "F" type engines, Apr 1st, 1943
sales brochure - Curtiss Hawk 75 pursuit airplane, unknown date
preliminary handbook - operation and service instructions for the model P-36A pursuit airplane, 1938
sales brochure - Curtiss Hawk 75-A pursuit airplane, unknown date
notice sommaire generale - l'avion de chasse Curtiss Hawk 75A-1, unknown date
notice de manoeuvre - l'avion de chasse Curtiss Hawk 75A-1, 1939
report A.&A.E.E. 749a, part 1, Mohawk I A.R.645, Cyclone GR1820 - G.205A, oil cooling and cylinder
temperature tests, Dec 15th, 1940
report A.&A.E.E. 749a, part 2, Mohawk I A.R-645 and A.R-678, Cyclone GR1820 - G.205A, performance
trials, Apr 5th, 1941
report A.&A.E.E. 749a, part 4, Mohawk I A.R.640, Cyclone GR1820 - G.205A, tests with air cleaners, Mar
13th, 1941
report A.&A.E.E. 749a, part 5, Mohawk I A.R.645 and A.R.678, Cyclone GR1820 - G.205A, fuel
consumption measurements and handling tests, Apr 5th, 1941
report A.&A.E.E. 749b, part 1, Mohawk II A.R.631, Twin Wasp R-1830-SC3-G, oil cooling and cylinder
temperature tests, Dec 18th, 1940
report A.&A.E.E. 749b, part 2, Mohawk II A.R.631, Twin Wasp R-1830-SC3-G, brief performance trials,
Mar 6th, 1941
air corps technical report 4338 - performance test of Curtiss Y1P-36 airplanes, Aug 28th, 1937
engineering section memorandum report - Curtiss P-36B airplane, Sep 25th, 1939
engineering section memorandum report - P-36A , Oct 31st, 1940, Jan 7th, 1941, May 16th, 1941, Jun
19th, 1941
official summary of characteristics for P-40, P-40B, P-40E
official performance summary for P-40, P-40B, P-40E
reports A&A.E.E. 783, parts 1 to 19, in particular
• report A&A.E.E. 783, part 5, Kittyhawk A.K.572 - weights, loading data and leading particulars, May
23rd, 1942
• report A&A.E.E. 783, part 6, Kittyhawk A.K.572 - fuel consumption trials, May 13th, 1942
• report A&A.E.E. 783, part 8, Kittyhawk A.K.572 - rate of climb and position error measurements, May
27th, 1942
• report A&A.E.E. 783, part 11, Kittyhawk I A.L.229 - take-off and landing trials, with and without overload
fuel tank, Sep 2nd, 1942
• report A&A.E.E. 783, part 12, Kittyhawk I's A.K.572 and A.L.229 - engine cooling trials, Sep 3rd, 1942
• report A&A.E.E. 783, part 15, Kittyhawk I A.L.229 - weight and loading data, Sep 30th, 1942
• report A&A.E.E. 783, part 17, Kittyhawk I A.L.229 and A.K.572 - handling trials, Dec 3rd, 1942
• report A&A.E.E. 783, part 19, Kittyhawk Ia L.T.573 - level speed performance with normal cowling fitted,
Jun 20th, 1943
C.S.I.R. report F.8, Kittyhawk aircraft, 1942
D.T.S. technical bulletin no. 31, P-40N performance, 1944
William H. Phillips - comparison of aileron characteristics as determined in flight tests of P-36, P-40,
Spitfire and Hurricane pursuit airplanes, Nov 16th, 1942
2nd hand sources - books on the subject
Francis H. Dean - America's hundred thousand, 1997
Lionel Pursyn - Curtiss Hawk H-75 in French service, 2010
Lionel Pursyn, Kari Stenman, Andrew Thomas - P-36 Hawk Aces of World War 2, 2009
Michal M. Mietelski - Samolet mysliwski Curtiss Hawk 75, 1987
Marek Rys - Curtiss P-36 Hawk, part 1, 2000
Marek Rys, Seweryn Fleischer - Curtiss P-36 Hawk, part 2, 2000
Jiri Chodil, Seweryn Fleischer - Curtiss P-36 Hawk, part 3, 2000
Kalevi Keskinen, Kari Stenman, Klaus Niska - Curtiss Hawk 75A, 1975
Peter M. Bowers - The Curtiss Hawk 75, 1966
Ernets R. McDowell - Curtiss P-40 in action, 1976
Marek Rys - Curtiss P-40 Tomahawk, Kittyhawk, Warhawk, part 1, 2000
Zbigniew Kolacha, Marek Rys - Curtiss P-40, XP-46, XP60, part 2, 2000
Krysztof Janowicz - Curtiss P-40, part 3, 2000
Terill Clements - American Volunteer Group Colours and Markings, 2001
TsAGI - samoletostroenye v SSSR 1917- 1945, part 2, 1994
particularly useful websites
http://www.gc2-4.com/histavions.htm
http://www.wwiiaircraftperformance.org/P-36/P-36.html
http://www.wwiiaircraftperformance.org/P-40/P-40.html
http://www.raafwarbirds.org.au/targe...e/p40_data.htm
http://www.enginehistory.org/reference.shtml
Sorting information
General
The above sources provide an abundance of data. However, all this data does not necessarily
support each other. Part of the package always is contradicting and even wrong information,
so the information has to be sorted, verified against each other, checked for plausibility, put in
the proper context. Eventually target qualities have to be derived.
Below we've used speed data of Allison V-1710-39 powered P-40's for an illustration of typical
variance of researched data. We've simply put various speed data into the chart, the data
already filtered to showing supposedly clean airframes (no external stores) on maximum
admissible power.
Similar compilations, though not graphically as interesting and often not as extensive, will be
done for many more aspects, like climb performance, weights, stalling characteristics, take off
and landing behaviour, and lots more.
Example: Speed data for P-40 with Allison V-1710-39 engine
We've found no less than 22 data sets, mostly top speed at altitude figures, but also some
detailed information, mostly British and Australian tests. Eventually we decided to go for the
speed performance as tested by the Australians, which seems to show the best performance at
low altitude, and planes are supposed to be modelled to best performance. Performance
around the full throttle altitude achieved with that plane appears a bit low in comparison with
the other data. The Australian test is coloured red in the above chart.
Actual flight model making
General
This is the easy part - the derived parameters can be entered into the engine and plane flight
model files. If the research gave good results, and a few particularities of the game engine
have been taken into account, historically correct input data will result in an fairly accurate
flight model right away. Still, next phase is testing, and then there will usually be a couple of
iterations until the desired results are reached. But even with all the iterations, the actual flight
model making is about 1% of the total work.
Example: Speed performance of P-40E
Weare trying to stay inside of 3% of determined target speed performance. This figure is
based on typical variances and tolerances found in the actual aircraft. The above chart shows
that for the most part we can get it much more accurate, however, at around full throttle
altitude, there's a noticeable difference. That's not a problem, because for one we already
considered the full throttle altitude to be lacking a bit and second, the difference is only about
10 km/h - well within our 3% margin. Also added is current performance with version 4.11.1.
Testing
General
Testing might be the biggest task when it comes to flight modelling, right along with research.
The changes have to be verified in game, for all planes effected, and all quantities involved.
This is a time consuming and complex task, because a change in the flight model aimed to fix
a detected problem, might cause a new problem where so far everything was OK. So even a
small change might require a whole lot of new testing.
Usually, we conduct climb, speed, take off, landing, ground handling, stall, spin, overheat,
dive, roll, turn, trim and other tests, and of course take the plane into virtual combat to see if
it behaves as expected when it matters. All in all, a lot of effort, which is luckily helped with by
our beta testers.
Below we've attached a couple of charts showing in game climb performances, compared to
old in game performances and the historical reference. Even though it is a simple test, it is
easy to spend a day on confirming projected performance for a plane family as large as the
P-36 / P-40 family - a climb to high altitude may easily take 15 minutes. Do it at several
speeds and several fuel loads, and you can imagine where the time goes.
Hawk 75A-3 climb
Reference is the median between a British and a US test, conducted at different weights but at
the same power settings. Using the average weight of both tests, and same boost, rpm, speed
profile and radiator setting (roughly 92% pitch, 100% power), we are now within 0.5 m/s for
most of the altitude. It is also apparent that the new flight model is much closer to the real
thing that the one it is replacing (il-2 compare data 100% ).
Hawk 75A-4 climb
Reference is a British test. Recreating this test in game under the same conditions, we again
are within 0.5 m/s for the most part. It compares favorably with the old flight model, which is
nowhere near the recorded performance.
P-40E climb
Target for touching flyable models is to at least not make it worse. The P-40 climb chart might
serve as an example. Both in game climbs have been done at the same settings, mirroring the
reference climb, as closely as possible. As can be seen, there's little to chose between the two,
the biggest advantage of the new flight model being an accurate boost gauge.
Hawk 75A-4 cylinder temperatures in a climb
Some people have been wondering about the overheating, in particular with the new model
introduced with the version 4.11. If data is available, we try to match it as good as possible, as
the above chart shows. For comparison we've added the 4.10 version, with both flight and
overheat model from back then. Neither the new model nor the new engine parameters are
perfect, but they are a lot more close than they used to be.
Сообщение от Sita
(c)mamali
Пока там беты, релиз кандидаты, ...лето наступит
Ваши права
Ответить с цитированием
