Jump to content


Photo
- - - - -

RoF test flight data


  • Please log in to reply
69 replies to this topic

#41 MiG-77

MiG-77
  • Posts: 2651

Posted 17 October 2010 - 12:43

Anyway it closest that we currently have ;) Also is it clear than most of these thin wing aircrafts had CLmax around 1-1,1 and thick wings around 1,3-1,4. -> most of the aicrafts are wrong or test/data too inaccurate.
  • 0

#42 =IRFC=AirBiscuit

=IRFC=AirBiscuit
  • Posts: 2455
  • LocationNaples, FL USA

Posted 17 October 2010 - 12:47

Piecost, is there any way you could regenerate your chart images to focus on the range in question a little more? For example, the stall chart could start the X axis at 40kmph instead of 0? Thank you for posting this!
  • 0

=IRFC=Air Biscuit

http://quetoo.org


#43 ImPeRaToR

ImPeRaToR
  • Posts: 7902

Posted 17 October 2010 - 13:57

[…]i really think that develompent team should think about change some beta testers which could have influence in FM changes. Im no wonder why some ROF planes fly like now in ROF.

:)

They should hire you, since you are such a great guy and spend so much time flying rise of flight!
  • 0

#44 Chill31

Chill31
  • Posts: 1891

Posted 17 October 2010 - 14:08

MiG, on this attached picture, I do not get 1.03 for Cl at any point on the graph, please explain how you have reached 1.03 Cl max.

Also, Cl max isn't necessarily max lift EFFICIENCY. It is the max lift produced before stall of the airfoil. Max lift efficency would be L/D (lift over drag) where the most lift is produced for the least negative effect (drag).

Also, MiG, could you please either post all of these airfoil documents that you have or post where I can DL them myself? Thanks!

Attached Files


  • 0

#45 Kwiatek

Kwiatek
  • Posts: 680

Posted 17 October 2010 - 14:16

[…]i really think that develompent team should think about change some beta testers which could have influence in FM changes. Im no wonder why some ROF planes fly like now in ROF.

:)

They should hire you, since you are such a great guy and spend so much time flying rise of flight!

I have spend a lot of time flying ROF some time ago since i still got hope that FM/performacne issues ROF of planes would be corrected according to aerodynamical laws and historical accuracy. But after one year i lost my patience. I dont want to fly in or against any (German or Entente) arcadish or wrong modelled plane.
  • 0

#46 Kwiatek

Kwiatek
  • Posts: 680

Posted 17 October 2010 - 14:22

Chill31 look here. There are 2 the same Cl/Cx polares for Spad:

The second is coming from the same source ( data) which posted Mig for RAF14.
These diagrams unfortunale had wrong discribtion values for Clmax.

Look for both charts.

First you have ~1.03 Clmax at 13.5-14 deg AoA

Second you have ~0.103 Clamx at 13.5-14 deg AoA.

So you see that second chart have just error in numering Cl values.

Attached Files


  • 0

#47 Chill31

Chill31
  • Posts: 1891

Posted 17 October 2010 - 14:27

Thanks Kwiatek, where do you get these pictures?
  • 0

#48 Kwiatek

Kwiatek
  • Posts: 680

Posted 17 October 2010 - 14:31

I like to be in the theme :)

Some of them i got from one of Neqob member. Im dunno why betatesters dont recive such data from development team? I think such data are important in FM development.

Im still interesting what critical angle of attack values have planes ( expecially some German ones know from ufo manouvers)
  • 0

#49 MiG-77

MiG-77
  • Posts: 2651

Posted 17 October 2010 - 14:42

MiG, on this attached picture, I do not get 1.03 for Cl at any point on the graph, please explain how you have reached 1.03 Cl max.

1,03 is for Eiffel 53 airfoil. Picture is from RAF14 (Camel) airfoil.

Also, Cl max isn't necessarily max lift EFFICIENCY. It is the max lift produced before stall of the airfoil. Max lift efficency would be L/D (lift over drag) where the most lift is produced for the least negative effect (drag).

We didnt talk about lift efficiency anywhere? Or did we?

Also, MiG, could you please either post all of these airfoil documents that you have or post where I can DL them myself? Thanks!

NACA report 93 and NACA report 244. You can google them ;)
  • 0

#50 Chill31

Chill31
  • Posts: 1891

Posted 17 October 2010 - 14:46

Kwiatek, could you please zip the ones you have and post them here so I can DL them?

Unfortunately being a beta tester doesnt mean I get to change anything. Very rarely will I get to "change" a flight model. The RoF community has more power (by providing data in the form of numbers, quotes, charts like these here) to change the flight models. And once Neoqb is finished with the next major revision (RoF will be almost completely new game again) I think they will focus on more of the details such as flight moodels. This is why it is VERY important to gather data NOW so that we are ready to help them when they call for it.

Kwiatek, if you have 'given up' on RoF, I recommend that you re engage yourself since your knowledge of aerodynamics is above average. It has been a busy couple of years for NEOQB and they dont have all the time or man power that they would like, so some features had to be left alone in favor of creating new features. The time will come though when refinement of old parts of the game (FMs for example) will become the priority. I hope you (and anyone else who has a strong knowledge of aeronautics) will be there to help

MiG, thank you! If we did not, then my mistake and hopefully anyone who reads it may find it useful anyway.
  • 0

#51 piecost

piecost
  • Posts: 1318

Posted 17 October 2010 - 20:17

Chill31 asked: Piecost, what equation are you using for Cl max? Also, may I ask your background? Why have you concluded that the Pfalz D3 and Fokker D7 cannot stall?

I used CLmax = ( Mass x 9.81 ) / ( 0.5 * air_density x Velocity^2 x Area)

where: Mass (kg)is taken from the RoF store, for the light weight I used Empty weight + 80kg for the pilot and for heavy I used the Takeoff weight

Air density = 1.22506 kg/m^3 - Sea-level International Standard Atmosphere conditions

Velocity = indicated air speed - converted to meters per second

Area = wing area from RoF Store

The Pflaz DIIIa (at low weight) and Fokker DVII (at low and high weight) lack sufficient elevator power to generate a high enough angle of attack to stall when slowly decelerating slowly in straight and level flight. They can be made to stall by pulling the stick aft at higher speeds. This is why in "real life" stall testing it is vital to decelerate at 1 knot per second on the approach to the stall.

My background is a degree in Aeronautical Engineering and working as an aerodynamicist for 14 years, many of those performing aerodynamic modeling. I have also performed some flight test analysis and handling qualities and a fair bit of wind tunnel work.

I am interested to see if any of you guys try to repeat my stall speed tests, there may be some differences in technique or equipment which influence the results.

Finally, with regard to CLmax, looking at aerofoil data is an excellent starting point but it only part of the story. You also need to consider the 3-dimensional effects of planform and configuration. The shape of the wing planform and the interference of the other wing (on a biplane) will influence the variation of CL over the wing span. The first part of the wing to achieve the sectional CLmax will cause the wing to stall. But, in reality, some other thing may cause a stall before the theoretical maximum is achieved. For instance, if the junction between the wing and the fuselage is poorly designed the wing may stall at a lower CL/incidence.

In conclusion, there is no theoretical or computer based method (that I have heard of) to predict CLmax for an aeroplane. The best method is to gather information for similar aeroplanes and to make an educated guess.

Large aerospace companies will make an estimation based on their previous aircraft, later perform wind tunnel tests and finally perform flight tests. It is unlikley that these will give the same results.
  • 0

#52 MiG-77

MiG-77
  • Posts: 2651

Posted 18 October 2010 - 04:55

Made calculations with your stall speed test, but with weight of 904kg for D.VIIF(store page weight for Fokker D.VIIF) and wingarea of 21,4m2 (store page wingarea + 1m2 from "third wing" between wheels).

CLmax = 1.45

Made same calculation with Fokker D.VII with 880kg (from french test report) take of weight (909kg in store page is very odd figure).

CLmax = 1.48

As we can see, data used makes big difference in these. Example that "third wing" is not usually listed in wingarea and weights may vary a lot depending on source.
  • 0

#53 ImPeRaToR

ImPeRaToR
  • Posts: 7902

Posted 18 October 2010 - 07:14

Wasn't the BMW engine lighter?
  • 0

#54 MiG-77

MiG-77
  • Posts: 2651

Posted 18 October 2010 - 08:09

Wasn't the BMW engine lighter?

When it was dry, yes. With oil/water no. Difference was ~7,5kg when dry.
  • 0

#55 JoeCrow

JoeCrow
  • Posts: 4145

Posted 18 October 2010 - 11:01

I have spend a lot of time flying ROF some time ago since i still got hope that FM/performacne issues ROF of planes would be corrected according to aerodynamical laws and historical accuracy. But after one year i lost my patience. I dont want to fly in or against any (German or Entente) arcadish or wrong modelled plane.

So you keep telling us.
  • 0

#56 Chill31

Chill31
  • Posts: 1891

Posted 18 October 2010 - 14:15

Piecost, thank you for your help!
  • 0

#57 NickM

NickM
  • Posts: 1625

Posted 18 October 2010 - 16:28

Yes, thanks Piecost, very useful.

With the D.VIIF, if you give it full power then gently raise the nose it never actually stalls or becomes unstable. It settles into a nose-high attitude and descends towards the ground like a parachute. I can't get other aircraft to do this because they become unstable and tend to roll off to one side or the other. Provisionally, I put this down to the "changing effect of rudder near the stall", where even the slightest perturbation in yaw causes an immediate roll out of the stable attitude (but this is just a guess). The D.VII flight model seems to retrain high stability in yaw even at extreme angles of attack and very low airspeed, which allows it to parachute to the ground.

Cheers,

Nick
  • 0

#58 =Fifi=

=Fifi=
  • Posts: 10329

Posted 18 October 2010 - 20:53

I didn't read all 6 pages, and my english isn't good enough…but have you noticed on Mig 77 turn test that N28 is a worst turner than EIII :o :shock:

That's hilarious…and pretty sad (for me)
  • 0

#59 piecost

piecost
  • Posts: 1318

Posted 18 October 2010 - 23:41

Corner Speeds and Load Factors

A test was carried out to determine the Corner Speed of each aircraft. This, when combined with my earlier estimate of CLmax was used to determine the ultimate and limit load factor for each aeroplane.

Corner speed (maneuvering speed) is the maximum airspeed at which all control surfaces can be deployed to their maximum deflection. It is usually the elevator that is critical and the corner speed is defined by the ultimate load factor (“n” is the number of “g”s the aeroplane can pull) and CLmax (the higher the CLmax the lower the corner speed). Corner speed does not vary with weight but the ultimate load factor is defined at maximum take-off weight.

It is a useful speed to remember as it sets the highest speed at which the controls can be used without risk of pulling the wings off (assuming no combat damage) The corner speed may in practice be limited by the maximum “g” which the pilot can withstand before graying/blacking out, this is modeled in RoF.

To find corner speed I performed a series of dives and pull-outs, starting at a low test speed and increasing the speed of the test by 5kph until the wings failed. Once I had found the approximate speed for loosing the wings the test was repeated a couple of times to improve the accuracy.

The tests used a Logitec Wingman Extreme joystick and Simped F16 rudder pedals with the default joystick & rudder responses.

The large gauges were used with a Post-it stuck to the monitor with 1kph graduations marked. All speeds are IAS taken from the large gauge.

The following mission was used:
Quick mission: Skirmish: Lille
____________Altitude: 3000m
____________Time: 12:00
____________Clouds: 5000m
____________Wind: 0m/s
____________Turbulance: 0m/s
____________Weather: Clear

The corner speed tests were carried out after stall testing and so the altitude varied, but the IAS accounted for the change in air density with height. The load tested; was

100% Fuel + 100% Ammo (no bombs)
5% Fuel + 0% ammo (no bombs)

On the start of the mission the controls were moved to full extent, as the sim seems to perform some kind of calibration, this maximized the available control deflections.

From straight and level flight with the throttle set to zero and radiator shutters closed (where fitted), the aeroplane was put in a shallow dive, keeping the wings level and zero rudder deflection, with no attempt to zero the slip bubble (where fitted). The time was set to half speed. The dive angle was varied to achieve the test speed quickly enough to avoid too much height loss without over speeding.

As the speed approached the target, the time was set to 1/8th speed and as the target speed was reached full aft stick was applied.

A first attempt at a technique was to pause the sim once the target speed was reached, apply full aft stick and then un-pause. It was hoped that this would generate a “stepped input” of full aft stick (an instantaneous application of full-up elevator) which would remove a source of human error. However, it was found that upon un-pausing the model of the stick in the cockpit did not reach full aft deflection.

The corner speed for each aeroplane is given in the graph on the left.

———————
Notes from the test

The start of the red band on the large gauges ASI approximates corner speed, but only matches corner speed for the Nieuport 28.

I could not find the corner speed of the Nieuport 11 as the maximum dive speed was lower than the maximum dive speed for structural failure.

I can’t remember why I have no result for the Pfalz DXII for 5% fuel and 0% ammo but noted that the aeroplane lacked sufficient nose-down elevator authority to dive at high speed.

It is apparent that for most aircraft that the corner speed is constant with light and heavy weights, indeed it should not vary with aircraft mass. However, the DVII, DVIIf, SPAD and Nieuport 28 all had a 5 to 10 kph higher corner speed at low weight. Perhaps these aeroplanes have the a forward Centre of Gravity at low weight? This would be due to an aft fuel tank and ammo position. An aft centre of gravity position makes an aircraft more sensitive in pitch, hence it may be possible to pull the wings off at a lower speed.

I performed a more detailed examination on the Camel (since its my favourite plane). In hindsight, I should have tried one of the DVII's

I found the corner speed for the Camel at a number of weights, with the assumed centre of gravity shown below.

Camel, engine at idle
________________________________________________Test 1____Test 2____Test 3____Test 4
Heaviest________100%Fuel+100%ammo+4 bombs________210_______205_____212_______212
Heavy___________100%Fuel+100%ammo+0 bombs_______215_______215_____215
Lightest________5%Fuel+0%ammo+0bombs_____________215_______215_____215
Forward CoG_____5%Fuel+100%ammo+0bombs__________215_______215_____212_______215
Aft CoG_________100%Fuel+0%ammo+0bombs__________215_______212_____213_______212

The scatter in the results are too great to conclude that there is any difference in measured corner speed with forward and aft CoG.

Calculation of Ultimate Load Factor

The ultimate load factor “n” was estimated by the formula: n = ( Corner speed / Stall Speed )^2

The stall speed was taken from my earlier post (rounded to the nearest 0.1CL) for:

100% fuel + 100% ammo + 0 bombs
5% fuel + 0% ammo + 0 bombs

The results are plotted in the middle graph.

———-

The results vary between about n=5.5 and n=15 I was surprised with how high these were !

The largest values of ultimate factor were for the planes with large difference in corner speed at low and high weights. I do not believe the light weight results for the Fokker DVII, DVIIF, Dr1, Camel and SPAD. The Pflaz DIIIa has the next highest value of about 13 which is a fairly consistent for both weights. The majority are around 8 to 10.

Comparing the results to real-life

Quoting from:

Aeroplane Structures
A J S Pippard & J L Prichard
published by Longmans Green & Co, July 1919

" Experience during the war showed that for main planes, reasonable load factors ranged from seven and five on the front and rear trusses respectively in the case of fast scouts and fighters to four all over for heavy bombers."

This, however, refers to limit and not ultimate load factor.

Limit & Ultimate Load

Aeroplanes are designed for the limit load. The varies depending on the use of the aircraft. It is the load which might be experienced once by a single aeroplane in a fleet of many over the entire life of the fleet. The aeroplane may be damaged (i.e. bent and may be written-off) but it should not catastrophically fail (i.e. the wings fall off). My test has not found limit load since i found the breaking speed of each plane.

The ultimate load is calculated:

Ultimate Load = Limit Load x Factor of Safety

Quoting from Pippard & Prichard again:

"Any increase over the minimum is an added guarantee of safety, and provides a reserve against the use of inferior materials and workmanship which may at some time or another creep into any particular aeroplane, however careful the inspection."

The factor of safety is often taken to be 1.5 since this this is based on the approximate difference between the yield and ultimate stress of aluminum (i.e. the difference between bending and breaking a metal aeroplane). The Ultimate load is the one at which the structure will catastrophically fail (ie. the wings fall off). This is what I measured.

Now, my 1919 British aeroplane structures book does not specifically mention the factor of safety used. But a value of 1.5 is reasonable.

I therefore calculated the limit load factor by dividing the ultimate load factor by 1.5, see the graph on the right.

Ignoring the dubious Fokkers at low weight, the numbers match the five to seven from Pippard and Prichard.

Since information is lacking; I guess that the developers grouped each aeroplane into catagories based on their year of design (1916 designs were stronger than 1914 designs) and reputation for strength. It is most likely that the assumed limit load factors are all integers.

Based on my testing the best guess of the RoF limit loads are as as follows:

________________limit load factor
Nieuport 11.C1______not calculated
Nieuport 17.C1______4
Fokker EIII__________4

DH2_______________5
SE5a______________5
Sopwith Dolphin_____5
Albatros DII________5
Albatros DIII_______5

Sopwith Pup_______6
Sopwith Triplane____6
Sopwith Camel______6
Nieuport 28.C1_____6
SPAD XIII.C1_______6
Albatros DVa_______6
Fokker DVIII________6

Fokker Dr1_________7
Fokker DVII________7
Fokker DVIIF_______7
Pfalz DXII__________7

Pfalz DIIIa_________8

I am surprised that the SE5a and Dolphin have a lower limit load than the Pup and Triplane, I would expect them to be the other way round. I am similarly surprised that the Fokker Dr1 has a factor of 7, it had a reputation of being of delicate construction. It is interesting that the RoF SPAD has a reputation for strength, the results show that this is due to the high corner speed rather than the strong construction. Just as the Pfalz had a high CLmax (low stall speed) it also has a high corner speed suggesting that it has a limit load factor of 8!

A final comment is to be suspicious of my results they are sensitive to the test technique, the joystick quality (not good, its rather old) and stick calibration. There also may be one or two calculation errors in there as well.

Attached Files


  • 0

#60 piecost

piecost
  • Posts: 1318

Posted 28 June 2011 - 17:59

E-M Plot for RoF Dr1 & Camel - lacking sustained turn curves

See attached an E-M plot for the RoF Fokker Dr1 & Sopwith Camel. The plot lacks the important sustained turn curves, but it is still of interest.


Sustained Turn

MiG-77 performed sustained RH turn test some time ago and measured identical turn rates of 40º/second at 78mph for the Camel and 68mph for the Dr1. I have performed a LH turn test on the Dr1 at low weight and can perform sustained turns of 160ft radius with ease.

Much more comprehensive tests will be required to generate a sustained turn curve, testing should be carried out in both directions for the rotaries.


Data

The attached table shows the values I assumed in compiling the plot, based on the RoF store and testing I have performed

Maximum take-off weight and wing area were taken from the RoF store.

Two weights are given for each aircraft; 100% fuel & 100% ammo (no bombs) and 5% fuel & 0% ammo. These give extreme ranges of weight, not necessarily representative of combat conditions but confirm the important of load-out to E-M performance.

The mass of the ammunition and pilot (80kg) were estimated to give the 5%fuel & 0% ammo case. This was chosen as the minimum mass case that could be tested.

The CLmax values were based on tests (see RoF test flight data post#16 ) in straight and level flight with engine at idle.


Lift limited Maximum Instantaneous Turn Rate

The 1g CLmax (engine idle) has been used to determine this curve. However, it may not be appropriate for two reasons:

1. Underestimation of full throttle CLmax - The increased lift and elevator authority due to slipstream at full throttle may give significantly higher values.

2. Limitation by “departure from controlled flight” rather than stalling.


Corner Speed

The lowest speed at which structural failure occurs was found from full elevator pull-ups with the sim at half speed and engine idle. Idle was used since corner speed is coincident/higher than engine limiting speed for some aircraft.

The Dr1 has the lower corner speed of 127mph compared to than 134mph for the Camel.


Ultimate Load Factor

The heavy Dr1 and Camel have factors of about 9g, it is likely that the difference is due to limitations in my testing and assumptions. The lightweight ultimate load for each aircraft are coincidently at just above 11g. I expect that both aircraft share a factor of 9g at high weight and the Camel should have a higher factor at light weight due to it’s greater disposable load. I believe that the developers group the aeroplanes into sets sharing common ultimate load factors (integer values), with a trend of increasing strength with date of design with, perhaps, some allowance made for reputation for strength/weakness. I have assumed that the Camel Ultimate factor corresponds to a weight without bombs.


Maximum Dive Speed

The maximum dive speed was determined by vertical diving, with throttle at idle at ½ time speed. The Camel has a significantly higher maximum speed 186mph than the Dr1. The Dr1 has a maximum dive speed of 137mph, only 10 mph above corner speed. However, maximum dive speed is not a very useful measure on these aircraft where the engine will be damaged before it is reached.


Dive Speed limited by maximum engine revolutions

For many RoF aircraft the maximum speed for full throttle is vital in the E-M diagram and replaces the maximum dive speed as a limit. The quoted speeds are for a warm engine (50% temp on the 2D gauge). Care is needed to avoid overcooling and the associated reduction in critical rpm.

The Fokker Dr1 has a full throttle maximum dive speed of 112mph, 15mph below corner speed and 25mph below maximum dive speed. The Camel has a maximum throttle speed coincident with corner speed at 134 mph and 52mph below maximum dive speed.

A further increase in dive speed (without engine damage) may be attained by switching the engine off. This could be argued to be irrelevant to the E-M diagram since the engine is not providing any power, but it represents a potentially useful tactical limitation.

The Dr1 with the engine off can attain the full diving potential of the airframe (for what it is worth – the Dr1 is a poor diver).

The Camel may attain an extra 9mph with the engine off (compared to full throttle) at 143mph and can out-dive (engine off) Dr1 by 6mph. This is not a great advantage, though an engine destroying dive to 186mph may be a viable last resort.

Attached Files


  • 0

#61 Chill31

Chill31
  • Posts: 1891

Posted 29 June 2011 - 01:50

It is an interesting convo you and Josf are having. However, the application to WWI aircraft is limited at best. Even if you could get 14Gs out of a DR1 without pulling the wings off, achieving corner speed is virtually impossible. The only lesson you get from E-M in WWI fighter is "more speed is better" since it is 1) impossible to achieve a speed higher than corner in a fight scenario and 2) maintaining corner costs so much E that you are at an immediate disadvantage if you try to hold it.

I also dont agree that the DR1 can out turn the Camel in either direction.

My conclusion is that corner speed is impractical to talk about for WWI fighters, losely applicable in WWII fighters, and a key factor in jet power fighters based on the ability to achieve/maintain corner speed.
  • 0

#62 piecost

piecost
  • Posts: 1318

Posted 29 June 2011 - 08:21

You have a point. I quickly tried to determine corner speed on the Camel, Dr1 & Pflaz DIIIa in a descending spiral as josf suggested and could only achieve onset of black-out at about 120 mph. I simply loose control of the plane trying to turn tighter and faster - I am flicking into/out of the turn due to slip/skid. It is due to due to limitations in my piloting technique. I conclude that the lift limited turn curve I plotted should be disregarded.

A very interesting topic, I am learning allot. Anyway, I'd better go to work.
  • 0

#63 Chill31

Chill31
  • Posts: 1891

Posted 29 June 2011 - 23:28

I do want to say that it is interesting! and I'm not intending to inhibit learning by any of us. I'm just looking at how to apply this discussion to WWI combat and ROF, and it is bleak from my point of view.
  • 0

#64 NewGuy_

NewGuy_
  • Posts: 4114

Posted 29 June 2011 - 23:36

I have made series of tests with game planes during last half year or so. You should remember that these are just my test results and may not be "the thruth", but they should close enought to be used as general quide of plane perfomances.

General test methods used: No wind/turbulence so those dont affect test, warmed up engine on. Everything else in realistic settings. Speeds, etc looked from simple gauges. In climb test, climb started from plane max speed at deck (flying just above trees). In turn tests, turn speed I just tested with 10km/h intervals so actual best might be ie 134,55467km/h ;) Also those turns are sustained turns and I made couple of circles before timing started.

Great stuff Mig. ;) Thank you for making this. Gavagai and Chill made performance figures for the various machines too. Maybe, the three of you can compare notes and experiences. Thank you again! :S!: MJ
  • 0

Something something SPAD. Something something then dive away. 


#65 piecost

piecost
  • Posts: 1318

Posted 30 June 2011 - 00:42

Chill, if aspects of E-M theory are not relevant to WW1 then that in itself it is an interesting and useful conclusion.
  • 0

#66 piecost

piecost
  • Posts: 1318

Posted 06 July 2012 - 16:05

Measuring the Drag of the RoF SE5a

When A.P. revised the RoF SE5 flight model he took the time and effort to be open with the community and discuss the revisions:

SE5a FM - review & fixes!

A.P. showed that he matched the RoF SE5a drag polar to data taken from R&M 603. This gave the opportunity to see whether a player (me) could perform glide tests and accurately derive the drag of any RoF plane.

I tested the RoF SE5a on the Lake Summer map with zero wind and turbulance and measured the time taken to descend from 2000m to 100m at different speeds. The change in height divided by the time taken gave the vertical speed and combining this with the indicated airspeed the drag polar could be constructed (see attached).

The graph shows drag coefficient on the vertical axis against lift coefficient on the horizontal. Comparing the test derived drag to that of the drag polar used in the flight model shows a massive difference at low KL (high speed).

This difference is explained by the drag of the stationary propeller in the R&M 603 data. They used a broomstick jammed into the propeller to keep it still. Whereas in RoF at speeds above 100KPH the propeller windmills, presumably creating less drag the faster it turns.

R&M 603 gave an estimate for the stopped propeller drag and A.P. used this value in the flight model. So I took my tested drag polar and corrected it to represent an SE5a with a stopped propeller. Where the speed was below 100KPH the RoF and real SE5a both had a stopped propeller and no drag was added. At 300 KPH I added the full propeller drag, and between 300 KPH and 100 KPH I varied the propeller drag linearly with speed. The dotted red line shows the RoF tested data corrected for stopped propeller drag. It matches remarkably well with the R&M 603 drag data.

So, the conclusion is that we can glide test RoF aeroplanes to determine the drag, but the effect of the propeller drag is massive and must be accounted for. We only know what the propeller drag is for the SE5a. (It might be reasonable to assume that the same value can be used for all planes released subsequent to the SE5a F-M revue). The propeller drag may be larger than the drag differences between different planes!

Attached Files


  • 0

#67 piecost

piecost
  • Posts: 1318

Posted 06 July 2012 - 16:17

The Increase in Drag due to the Lewis being Tilted Back

I was intrigued to find out the effect of pulling the Lewis gun down on the foster mount. So I repeated the glide tests at 4 speeds for the gun horizontal and tilted back. The number of speeds is not high enough for accuracy, but the results were consistent. See the attached graph. I forgot to pit on the test points, this is a bad mistake since we cannot tell how much scatter was in the data relative to the size of the thing we are trying to measure.

The dotted line shows an increase of drag of about 8% at zero lift with the Lewis tilted. The drag coefficient increase is constant over the whole KL range and thus at all speeds. This means that there is no effect on the induced drag.

Attached Files


  • 0

#68 Chill31

Chill31
  • Posts: 1891

Posted 23 January 2013 - 03:00

Chill, if aspects of E-M theory are not relevant to WW1 then that in itself it is an interesting and useful conclusion.

Long time responding to this thread, sorry.

It does of course depend on the max G available. For example, a Dr1 with a level stall speed of 65 kmh will stall at 145kmh at 5Gs and 195kmh at 9Gs, and this is indicated airspeed. So if the Dr1 structure is capable of 9Gs, then corner speed is almost entirely impractical. However, if 5Gs are the safe limit, then corner speed is losely applicable because you can at least achieve it.

The reason I argue that it is mostly impractical is due to the very short turn radii of WWI aircraft and the ability of pilots to see the enemy at long ranges. WWI pilots can turn around so quickly that there is really no "competition" in the turn circle. You should always meet your enemy head on. At this point, you'd have to look at E-M charts to see if it was worth trying to max perform the turn rate. My feeling is these planes expend energy too quickly trying to sustain Max G which leaves the guy who didnt choose to go 100% angles in a much better energy state to either reposition for attack or get away.
  • 0

#69 Chill31

Chill31
  • Posts: 1891

Posted 23 January 2013 - 03:26

This difference is explained by the drag of the stationary propeller in the R&M 603 data. They used a broomstick jammed into the propeller to keep it still. Whereas in RoF at speeds above 100KPH the propeller windmills, presumably creating less drag the faster it turns.

Depending on propeller pitch, a stopped propeller may produce less drag than a spinning propeller (not saying that is the case for the prop you tested). I'd have to go run the numbers again to find the actual blade angle for WWI props. But here is a decent little graph in this link: http://www.casa.gov....eName=32-33.pdf" onclick="window.open(this.href);return false;">http://www.casa.gov....cripts/nc.dll?W … =32-33.pdf
  • 0

#70 piecost

piecost
  • Posts: 1318

Posted 25 January 2013 - 18:49

Chill, thanks for the response - there is no rush! I am slightly embarrassed to respond so quickly, but this is easier than doing more useful things.

Regarding E-M theory.

I certainly found that transitioning from IL2 to RoF (offline) I kept ending up head-on. I never really "got" the point of sustained turns, other than in demonstrations of E-M theory, rather that the diagrams gave useful indications for other "energy" manoeuvres. I got the impression that turn fighting was crucial, but, as you say, sustained turns were not practical. Were rapidly descending Lufbery turns the norm? I find that in RoF when gaining the upper hand on the AI that they perform a "corner speed" descending turn. Sometimes I cannot tell whether their engine has stopped or not. It makes so little difference!

I believe that the high G turns are over-cooked in the Sim since most of the planes are modeled too strong (able to withstand 8-9G rather than 5-7G Ref: Current ROF Airplanes Flight Model Discussion Topic. post #311).

To quote from R&M 670:

".. a number of accidents occurred due to wings of Camels breaking during rapid turns, and the possibility of this is shown very clearly in the Table. The calculated load factor of 5.4 was that existing on these machines at the time these accidents occurred; this value was subsequently appreciably increased"

Current ROF Airplanes Flight Model Data Topic.
post #37

Note that this is before the concept of limit and ultimate load. These aircraft would not be considered strong enough to be aerobatic nowadays. (Having said that; I can't quite remember if Achim Engels was certifying his Fokker DVIII as aerobatic). I do not advocate weakening the RoF flight models; since I lack a force-feedback joystick and would probably keep ripping my wings off! A British mock dogfight indicated much lower G levels (see attachments). They were generally below 3G - was this the hardest the pilot was prepared to pull with sufficient percieved safety margin? Or, was there no advantage in energy draining hard turns? I had not carefully examined the data but cannot see any obvious max-G turns. I do not know what tactics/maneouvres were used but trust that they were representative.


Propeller Drag

I tried to open the link and got a nc.dll file instead of a pdf.

quoting from Vonrd_JS109

"And a windmilling prop does create more drag according to my CFIs (never had the nerve to actually test it out!). I was taught to raise the nose and slow until the prop stopped (given sufficient altitude of course) and then resume best glide speed"

ROF Planes glide too well?
post #46

It makes sense to me that a windmilling prop can create more drag since it acts like a windmill; extracting energy from the air to turn the engine. I conclude that this is not modeled - at least on the SE5a. It might be a good thing to add, but could be difficult to estimate. The friction of a windmilling engine may be much higher than a running engine due to lack of oil circulation, cold temperatures etc. I am not sure of the operating conditions of the propeller in this condition - it could be extreme - stalled propeller blades, hence difficult to reliably predict the drag and torque. On aircraft lacking accurate drag data windmilling drag is a very minor detail.

Attached Files


  • 0


0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users