Aircraft performance curves

[Christer Bergström]

To the ongoing discussion on performance curves in the book section, I’d like to add my 5 cents:

Most of the published performance curves are unreliable, due to different reasons. Some of the most important are these:

Above all, nearly all performance curves are the result of clinical trials.
Secondly, many performance curves are highly unreliable because they are the result of test flights of captured, worn and often damaged (and repaired) aircraft where - due to the unused “enemy’s” lacking experience - the best simply wasn’t brought out of the aircraft. (See examples below.)
Thirdly, some performance curves originate from test flights with captured aircraft where we are not informed of the equipment (load) with which these tests were made. (See examples below.)
Fourth: One and the same aircraft had different performance due to different stages of engine wear, and depending on how much fuel and ammunition the aircraft carried. (Performance could also differ depending on polishing or worn paint, etc.)
Five: These so adored performance curves show the theoretically maximum speed attainable after a certain time in level flight, i.e. straight forward. Mostly the curves don’t say how long time it took the aircraft to reach that speed at level flight - which would be a very important factor to know. Since those performance curves are used to compare fighter planes, they are of fairly little use, because hardly any fighter combat is flown only at level flight.
Six: Curves for acceleration would be more useful, but then we would need lots and lots of tables showing acceleration from different speeds at different angles (dive, climb) and at various degrees of turning or recovering from turns, at various propeller pitch and trim applied, etc.
Seven: Two aircraft of the same type rarely had exactly the same performance, due to a multitude of reasons. (See examples below.)
Eight: How an aircraft is flown in combat differs from pilot to pilot - depending on how skilful he is to handle the aircraft in order to utilize its performance to a maximum - and that creates a huge difference between the best pilot in the best plane at the lowest weight from the worst pilot in the most worn plane at maximum load.

Now please don’t get me wrong. I understand that such realistic performance curves probably are not available, and I would not make a fool out of myself and demand that something should be published which I don’t know if it is available.

But I am questioning the strong demands for simple performance curves, as though these would be of any particularly great value.

The clinical performance curves which people here talk about might be amusing, and may give a weak hint, but for someone deeply interested in the technical details - someone who really is interested in the actual performances in real combat - they are of no greater use than any cursory school book in history is to a historian.

I will give you some examples which will show just how unreliable these performance curves are:

A Bf 109 F-4, test-flown by Messerschmitt at Augsburg, 29th November 1941 showed a performance which differed the following, compared with a test flight at Rechlin with the same aircraft type:

About 15 kph slower at sea level.

About 5 kph faster at 2000 m altitude.

About 5 kph slower at 4000 m altitude.

About 10 kph slower at full pressure altitude of 6200 m.

About 15 kph slower at 8000 m altitude.

The flight weight is with 2890 kg identical with the flights in Rechlin. Without knowing more details of the test planes, the differences cannot be explained.
http://www.beim-zeugmeister.de/zeugmeister/index.php?id=26&L=1

Then we have the calculated performances of the Bf 109 F-4, by Messerschmitt in Augusburg, 1 July 1942:

And again in comparison with the Rechlin speed measurements:


About 18 kph slower at sea level.

About 9 kph slower at 2000 m altitude.

About 8 kph slower at 4000 m altitude.

About 35 kph slower at full pressure altitude of 6200 m.

About 44 kph slower at 8000 m altitude.

http://www.beim-zeugmeister.de/zeugmeister/index.php?id=27&L=1

So if the span could be as wide as 44 kph (27 mph) even under ideal conditions - then of what use are these performance curves really? Other clinical tests have showed that the maximum speed at level flight for one and the same aircraft during the same flight (i.e. with the same degree of engine wear etc) could differ 30 kph between max load and nearly empty. Add that to the 44 kph above, and you get a span of 74 kph (46 mph). Now add the other factors, such as engine wear and different pilot skills, and you will start approaching reality.

An article at virtualpilots.fi should give the “performance curve believers” more to consider:

” When talking about the Messerchmitt 109 performance, we must take into account that many western sources are simply wrong. They are based on original wartime allied test flights flown with damaged planes, or with such equipment that the planes do not represent a normal fighter variant. Also western sources often fail to quote the used power setting. Was the engine runnign on continuous, 30 minute or 5 minute power? Western performance numbers (US/FAF/RAF) are always quoted with maximum power settings. Luftwaffe standard was to test all climb and level speed performance with the 30 minute setting, which really gives a more "real life" performance. (. . .) A good example this are the 109 F-4 tests.

”Bf 109 F-4 . . . Practically all its performance reports stem from a single British test flown with a damaged airplane with derated engine. All other test "reports" are copied from this one test. (. . .) The American test, “Combat Evaluation Report Nr. 110” for the Bf 109 F, 7th February 1943”, are only a compilation of the British test reports sent to the USA and no American flight tests were flown with F-4s. And to top it, the transferred report is riddled with errors in converting the numbers and drawing the performance curves. For example the reported climb rate is the British climb time for 16,500 feet converted to 15,000 feet. Also in the American summary are existing further serious transfer errors.”

http://www.virtualpilots.fi/en/feature/articles/109myths/

Very interesting! Now compare with this - you can check the source for yourself:

“In the book "War Prizes" by Phil Butler is told, that the Bf 109 F-4 with the built number 7232 made a wheels up landing in Beachy Head on 20.5.1942. After replacement of the damaged engine the plane was flown on 24.10.1943 at the first time by British pilots. Measurement results of test flights with this plane are not documented, another Bf 109 F-4 was never captured.
Where did the British performance data from August 1942 then came from? Most probably the real flown measurement results of a Bf 109 F-2 were taken as base and extrapolated due to the engine card of the DB 601 E. It has to be mentioned, that the measurements of the Bf 109 F-2 performance were faulty, since the engine was troublesome during all tests and had to be repaired several times. The British authors of the Bf 109 F-2 test report had pointed out clearly, that because of this the measured performances should only be seen as minimum performances.”
http://www.beim-zeugmeister.de/zeugmeister/index.php?id=32&L=1

Back to virtualpilots.fi:


“Comparitive Trials between Me109E and British Fighter Aircraft, RAE (?), 14 August 1941
Here we have two interesting reports. They're actually a 1941 report from tests conducted in September of 1940 from an aircraft that was captured by the French in 1939. (. . .) The 109 tested is claimed to be "Me 109E-3 Werk-Nr 1304" which is documented to have been captured. However, there is some discrepency as to WerkNr 1304 actually being an Me109E-1. So what have they tested? E-1? E-3? E-4? Did they test one of the crash landed, damaged planes? (. . .) Another problem is with the test itself, when compared to a Spitfire. Overall the accuracy of the test suffers from the fact that it was flown with a crash landed plane wirh a worn, several years old engine producing less power than usual. It was then flown against a brand new Spitfire with a 1940 engine. As shown by the test data, the turns were made in the 120mph range which is too slow for the 109 slats to be deployed, which doesn't compare the maximum turning abilities of each aircraft. (. . .) The French flew their test with the same Me 109 E-3 (E-1?). The test results are not available in English, but to author's knowledge their recorded performance numbers are higher than in the British tests.

(. . .)

“The Bf 109 G-6 model's performance numbers are usually quoted from a flight test flown by mr. Brown. The actual plane was a 109 G-6/U2, which is a three cannon night fighter variant with night fighting equipment. Authors now take these numbers, drop away the information that it was a 3-cannon night fighter and voila, we got weak performance numbers for the G-6. The 3-cannon night fighter G-6 made 621 km/h in 30 minute power setting. A clean G-6 does 635-640 km/h with 30 min setting and 650+ km/h with 5 min WEP setting. So you can see that the wing cannons not only decreased speed, but they decreased the climb rate, roll rate and overall agility of the plane.”

http://www.virtualpilots.fi/en/feature/articles/109myths/

[

[Juha]

Christer,
Only a few short comments. It is well known that there were speed differences between individual a/c of same type. And for example fully open radiator flap(s) dropped the the max speed of Bf 109G circa 50km/h and that of LaGG-3 c. 15km/h, this info is from Finnish tests. So on warmer day the max speed tended to be lower. So one got usually better figures in Nov. than in July.

It of course clear that the performace curves flown by captured a/c are somewhat doubtful. Much depends how much info the test organisation had.

Acceleration in different situations are possible to calculate, if some basic data is available.

On graphs; there are in Dean's America's Hundred-Thousand some which have both manufacture's and USAAF's/USN's graphs on same type and surprise, surprise, manufactures´ graphs tended to show better performance.

And as I wrote before here in Finland we have a book, Raunio's Lentäjän Näkökulma II, which shows the area where the max. speeds of the FAF's MS-406s were after major overhauls plus the manufacture's speedgraph. The book also gives the variation of max speeds of overhauled FAF's Fiat G.50s, info that Brewster 239s usually were very near to manufacture's figures, Curtiss H75As always clearly slower etc.. So I and surely also Jukka are well aware of the problems and merits of performance graphs and I still think they are very usefull but must be used with care.
BTW all the info in Your message was familiar to me before so I still think the graphs useful.

Juha

[Christer Bergström]

Juha, those particular performance curves which show the variation of max speeds of overhauled FAF aircraft seem to be quite interesting.

However, what I've seen here is people just asking for those plain performance curves which we have seen so many examples of in cheap few dozen-page aircraft monographies. I have heard complaints that there are no performance curves at all, but I have never heard the people who are so interested in those performance curves ever discuss the veracity of ordinary performance curves.

What I'd like to know is: For what are those ordinary performance curves so useful?

[Juha]

Hello Christer
The use of graphs made it easy to understand for example by a glance why Spit LF Vs were built, even if its max speed was lower than that of Spit F. V. From graph one sees that in lower altitude the LF V used in the test was clearly faster than the F. V used. Because both a/c used were probably in good condition and were flown by test pilots it is possible to draw a conclusion that also an average LF V is faster at low level than an average F. V and also up to circa what altitude a LF V is faster if both pilots handled their planes more or less same skills. Exact altitude of the transfer is of course only valid to those 2 a/c tested at those conditions were test was made but it also gives a rough estimate for “average” a/c. So those speed graphs give a basis from where one can become add more variables (condition of engine, of airscrew, of skin, plus pilots ability, air temperature etc).

Because this is LW and Axis forum I give the G.50 figures. The max speeds of Fiat G.50s test flown by FAF at 5000m varied between 403 and 430km/h, most important reason to variations was the airscrew used. These tests were flown with 790 mmHg boost pressure not with the max permissible temporary overboost pressure of 890 mmHg but Finns calculated that 430km/h with 790mmHg means circa 445 km/h with 890mmHg. Fiat’s figure for max speed was 484km/h at 5000m. So FAF Fiats were 40 – 60km/h slower that the manufacture claimed. And from a speed graph flown by G.50 FA-31 one sees that the difference to Fiat’s figures was biggest between 2500 – 7000m and smallest around 2000m.
Juha

[Jukka Juutinen]

Originally Posted by Christer Bergström

Juha, those particular performance curves which show the variation of max speeds of overhauled FAF aircraft seem to be quite interesting.

However, what I've seen here is people just asking for those plain performance curves which we have seen so many examples of in cheap few dozen-page aircraft monographies. I have heard complaints that there are no performance curves at all, but I have never heard the people who are so interested in those performance curves ever discuss the veracity of ordinary performance curves.

What I'd like to know is: For what are those ordinary performance curves so useful?

In short: they are a very compact way to present plenty of information in easily understood manner. Tell me which is the best way to describe aircraft speed, rate of climb and time to climb at various power settings from the sea level up to the service ceiling? You can have all this data in one compact graphic presentation requiring less space than covered by a cigarette box. That is the beauty of performance curves!

[Christer Bergström]

Since you admit the relative weakness of the simple performance curves (another one, not mentioned here, being that most of them are published without any source reference), I still fail to understand for what they are so useful. Knowing about the unreliability of the performance curves, such information can be explained with less inaccuracy in a few lines of written text - like you just did yourself, Juha:

At low altitude the LF V was clearly faster than the F. V.

Jukka used a quantitative method, explaining that a performance graph table requires less space than covered by a cigarette box. Well, the eleven worda above require even less space than a cigarette box, and contains less inaccuracy. I would also prefer a praiseworthy description as Juha's below rather than one of those ordinary and unreliable performance curve graphs:

The max speeds of Fiat G.50s test flown by FAF at 5000m varied between 403 and 430km/h, most important reason to variations was the airscrew used. These tests were flown with 790 mmHg boost pressure not with the max permissible temporary overboost pressure of 890 mmHg but Finns calculated that 430km/h with 790mmHg means circa 445 km/h with 890mmHg. Fiat’s figure for max speed was 484km/h at 5000m. So FAF Fiats were 40 – 60km/h slower that the manufacture claimed. And from a speed graph flown by G.50 FA-31 one sees that the difference to Fiat’s figures was biggest between 2500 – 7000m and smallest around 2000m.

We have heard people here talking about their technical knowledge. I don't say that aircraft performance curves is totally superfluous, but it requires technical knowledge to be able to use those curves properly, and I find it surprising that no one of those who have such high thoughts about their own technical knowledge have discussed the weaknesses of these ordinary performance curves. I would say that to someone without the high level of technical knowledge which has been implied by members on this forum, a table with performance spans like in Juha's example above is of far greater use.

Also, the value of performance curves may differ from aircraft to aircraft. For the Me 262, which based its success on high speed, a comparison between average performance curves of the Me 262 and the P-51D could be quite useful.

However, for the Me 109 and its most common tactic of "boom & zoom" attacks, the initial climb rate from various accumulated speeds would be more useful in order to compare its effectivity against enemy aircraft like, for instance, the Airacobra I.

In a prolonged fighter combat which took the shape of a turning combat - which in WW II was carried out at speeds of about 350 - 450 kph - a fighter aircraft's top speed at a certain altitude is of fairly little value; in such cases we rather need to know the competing aircrafts' roll rates, initial turning rates, sustained turning rates, and initial rates of climb.

I am afraid that the staring at performance curves and simple speed statistics (like 391 mph or 387 mph) has led the focus away from those more crucial facts for the understanding of an aircraft's relative value.

Besides, I particularly wonder for what there is such a great need of performance curves for a bomber or ground-attack aircraft. Here we have talked about various top speeds of fighter planes at low or high altitudes. But for what do we need to know a Hs 129 ground attack aircraft's different average level flight top speeds at clinic tests at 17,000 ft, 22,000 ft, 24,000 ft, and 28,000 ft respectively?

[Jukka Juutinen]

You need rate of roll data? Then what is a better way to present that data than a curve showing rate of roll on the vertical axis and the IAS on the horizontal axis? You can easily have several curves at various stick forces on the same graph making comparison easy. Or what is your objection to curves this time?

[Juha]

Christer
the beauty of graphs is that one sees at glance roughly how much faster than Spit F. V LF V was at certain level, up to what level it was faster and how must faster F. V was above that at any given altitude. To explain that in writing one need a table which is more difficult to gasp.

Usually there are at least a speed graph, time to altitude graph and rate of climb graph for a certain type/version of a/c in books.

Beauty of a speed graph is that at level flight max speed is attained when max. thrust possible to develop with the engine-propeller combination (plus exhaust stack jet thrust) is equal to the drag. From that one can deduce much, if the thrust is smaller a certain amount because for ex. engine wear, propeller pitch problem, propeller wear because of stone hits etc, the max speed is lower by certain amount which is possible to calculate. Same way if the a/c skin is carefully waxed the drag of the plane is a little bit smaller and the plane is a certain amount faster. All is pure physics, not simple physics but anyway calculable.

There are books which explain the basics on all this, certainly also in Swedish.

I agree that the performance graphs might be more important for understanding fighters than ground attack planes.

HTH
Juha

PS as Jukka wrote, one can presents many phenomes by graphs, they are like compressed info.

[Graham Boak]

It is perfectly true that some of the available information is not fully defined, or of doubtful origins, but this does not lead to the conclusion that all should be discarded. A comparison of several such datasets - the more the better, surely? - will soon lead to discovering which are realistic and which are not. However, even the best can be viewed as a measure of the capabilities of the aircraft, if idealised compared with service conditions.

The development of an aircraft cannot be understood and appreciated without understanding the technical changes that did or did not happen, that were or were not possible at the time. Some part of this is most easily expressed in terms of these fairly simple curves, annotated as necessary.

From these charts can be obtained basic drag data, which when coupled with engine power, weight and lift can be used to produce all the intermediate climb, dive and acceleration points you suggest. Perhaps you are not aware that there is a simple conversion between climb rate and acceleration?

Every aircraft ever built would take a little while to accelerate the last few kph/mph to meet a true Thrust=Drag level speed. (Early high subsonic jets possibly excepted, due to the rapidity of the drag rise.) I don't see the value of this comment.

Comparisons are also best made using these charts. The relative performances of the Spitfire Mk.V, LF Mk.V, Fw 190A and Spitfire Mk.9 can be shown very easily on such a figure, especially the differing variations with altitude. To express all this in words would take many sentences, and would lack the clarity of the picture. Comparative areas of combat advantage are clear, with their key points of changeover.

ideally, it would be best to see the full flight envelope in terms of Specific Excess Power contours, but this technique was not available at the time. Some studies have been done to generate this but apart from one on the P-40 vs Zero I can't bring any to mind. Modern sources rather lack the knowledge available in the original aerodynamic offices. Without this we can only do the best we can with the data available, knowing that the air has not changed, and neither have the laws of physics. The more data, the better the job.

[Kutscha]

A picture, err graph, is worth a thousand words of text.