PDA

View Full Version : Fw 190 A8 dive performance comparison: C++ code simulated and PF V4.04m



Holtzauge
02-18-2007, 08:19 AM
Below is a comparison of dive performance for a Fw190 A8 done with a C++ simulation code I have hacked together as compared with flown values in PF V4.04. The PF curve may be off somewhat since I only flew one track and only eyeballed the dive angle. However, I believe that the margin of error that this introduces does not affect the conclusions as outlined below.

The C++ program models subsonic drag rise due to compressibility, exhaust thrust, propeller efficiency at both low high disc loading, losses at high prop tip speeds, ram effects on engine performance etc. I have chosen the Fw190A8 because there is a lot of data on max speeds, climb rates etc around that can be used to calibrate the modeling in the code. I have used 1.42 ata boost and a weight of 4300 Kg for the A8 in the simulations. This gave me a top speed of about 545 Km/h at sea level and 643 Km/h at 6.2 Km so I believe I am pretty close in the modeling (below the drag rise at about M=0.7) but I am still tweaking the modeling. Above M=0.7 I have guessed the drag rise (right now it reassembles the P-51) so I would appreciate if someone could help me out here with a Cdo/Mach or TAS chart for Fw190A and/or the D series.

As can be seen, there is a surprising degree of agreement between the PF and C++ simulated IAS from 250 Km/h up to about 500 Km/h. At 500 Km/h at 8-8.5 Km altitude the Mach reaches 0.68 and there should be a substantial increase in drag due to compressibility effects as the simulation shows. However, PF does not seem to model this. I did a test where I removed the drag rise in the C++ code but left in the decrease in propeller efficiency due to high prop tip speeds. The tip speeds reach around M=1.1 in the simulation. This is shown by the "C++ no mach" curve. If this is compared with the PF IAS curve it seems to indicate that the loss of propeller efficiency due to supersonic prop tip speeds are not modeled in PF either.

I have also included the max allowable IAS from the Fw190 Bedienungsvorschrift LDvT2190 A-1 bis A-8 for the altitudes 9,7,5 and 4 Km altitudes. As can be seen, the PF IAS values are way above the figures allowed by the POH. The C++ figures are quite a bit below which I would also expect since a 45 degree dive does not seem excessive and it seems reasonable to expect that the a/c could handle a power dive under these conditions and not shred the wings like happened to me in PF. (That's why the light blue curve ends at 4 Km) http://forums.ubi.com/images/smilies/16x16_smiley-wink.gif

BTW: Does anyone have the Fw190D9 propeller efficiency chart Schpam posted a while back? The link is broken and I would appreciate if someone who has it posted it again or can give atip where to find it.

http://img.photobucket.com/albums/v210/Holtzauge/Fw190A8divePA1.jpg

JG4_Helofly
02-18-2007, 08:41 AM
WoW! Thanks a lot!

I have been waiting for something like that for years to finaly know how close il2 is to RL.

Is it possible to simulate other flight conditions like acceleration or zoom climb?

MEGILE
02-18-2007, 09:00 AM
This agrees with previous tests someone did on a P-47. I can't remember who... but IIRC there is a P-47 dive performance chart floating around somewhere.

JG4_Helofly
02-18-2007, 09:06 AM
Yes, This test showed that the dive acceleration of the p 47 is way to high in the game ( probably all other planes have the same problem ).

Maybe Oleg will model these missing effects in Bob which slow down the aircraft in a dive.

VW-IceFire
02-18-2007, 09:06 AM
Nicely done http://forums.ubi.com/groupee_common/emoticons/icon_smile.gif

I think that fits my general view that there is far more right in the sim than wrong but that not everything is modeled yet. As time goes along...more computer power will allow for more calculations.

MEGILE
02-18-2007, 09:43 AM
http://mitglied.lycos.de/jaytdee/testgraph/p47dive.JPG

It was by JTD.. but I couldn't find the original P-47D30 chart.

Holtzauge
02-18-2007, 09:49 AM
Originally posted by JG4_Helofly:
WoW! Thanks a lot!

I have been waiting for something like that for years to finaly know how close il2 is to RL.

Is it possible to simulate other flight conditions like acceleration or zoom climb?

Yep, I can simulate, acceleration, sustained and momentaneous turn rate, climb rate, zoom peformance and of course speed. When I find the time and get the necessary data I plan to do comparisons of different a/c.

Maybe I should model and compare Spifire 14 and 109K4 turn rate http://forums.ubi.com/images/smilies/winky.gif

Seriously, right now I lack data on prop efficiency and exhaust thrust for both the Me109, and Spit though. I also lack a Cdo/mach drag diagram showing the drag rise for the Me109 at high subsonic speeds.

Perhaps someone can help out with this?

Holtzauge
02-18-2007, 10:23 AM
Originally posted by Megile:
http://mitglied.lycos.de/jaytdee/testgraph/p47dive.JPG

It was by JTD.. but I couldn't find the original P-47D30 chart.

Is there any info on altitude as well? It's a bit tricky to interpret the chart without that.

BTW: There is a nice article in Flight magazine March 1945 that shows the P-47 doing around 510 Mph instrumented TAS at 14000ft diving from 34000ft.

MEGILE
02-18-2007, 11:16 AM
Originally posted by Holtzauge:


Is there any info on altitude as well? It's a bit tricky to interpret the chart without that.

BTW: There is a nice article in Flight magazine March 1945 that shows the P-47 doing around 510 Mph instrumented TAS at 14000ft diving from 34000ft.

I presume JTD performed the test as close to the real parameters as possible.
Maybe you should send him a PM.

JG4_Helofly
02-18-2007, 02:37 PM
Now dive and zoom comparison between the 190 and the spit would be very interessting. After so many years of showing same reports and energy charts from completly different planes ( Josf http://forums.ubi.com/groupee_common/emoticons/icon_wink.gif ) it would be cool to finaly know how it should be.

I am looking forward to your tests

MEGILE
02-18-2007, 02:40 PM
Originally posted by JG4_Helofly:
Now dive and zoom comparison between the 190 and the spit would be very interessting. After so many years of showing same reports and energy charts from completly different planes ( Josf http://forums.ubi.com/groupee_common/emoticons/icon_wink.gif ) it would be cool to finaly know how it should be.

I am looking forward to your tests

Back in the day JTD did say he was going to test other planes, but alas I don't think it ever happened.

Kurfurst__
02-18-2007, 02:51 PM
Good stuff, Holtzauge ! Grat's! http://forums.ubi.com/images/smilies/25.gif

stalkervision
02-18-2007, 04:11 PM
Excellent work there Holtzauge! http://forums.ubi.com/groupee_common/emoticons/icon_cool.gif Your making us all hungry for more of your most excellent research.. http://forums.ubi.com/images/smilies/clap.gif

Holtzauge
02-19-2007, 02:19 PM
Thanks for the positive feedback guys, it's appreciated http://forums.ubi.com/groupee_common/emoticons/icon_smile.gif

How about some input data for the Spitfire and Me109? If I'm gonna model them I need data on prop efficiency and exhaust thrust. The prop efficiency data should have efficiency on the one axis and either speed, advance ratio or propeller speed coefficient Cs on the other axis. The exhaust thrust is usually presented with thrust on the one axis and altitude on the other. The drag rise would be nice to have too ( e.g a chart with Cdo on y-axis and mach up to about 0.8-0.9 on the x-axis).

On the PF FM issue I agree with Icefire that PF has some merits as is but I do not think we need to wait for more powerful computers to the model subsonic drag rise. The FM in PF has to include some sort of acceleration algorithm and all that is needed is to include a "if M > 0.7 then " line and reduce the acceleration abruptly when these speeds are reached. The same goes for propeller efficiency at high speeds: even a simple linear reduction as prop tip speeds go up would get us much closer to reality.

It's not a doomsday machine we are talking about. Just a few lines of code. As Dr Strangelove would have put it: it only requires the WILL to do it.

Sadly, my prognosis is that BoB will stagger us with a beautiful graphic package hiding the same limited FM but I hope to be proven wrong.....

JG4_Helofly
02-19-2007, 03:02 PM
For fw 190 datas ask Crumpp, he is the expert and has probalbly some drag charts for you.

Speaking about bob, Oleg said that the FM will be impoved, so maybe these missing FM components will be add in the new sim.
And if it's realy so easy to programm these things I wonder why Oleg did not implant them in il2.
Let's wait one year, than we will see if planes can still dive to surrealistic speeds.

WWMaxGunz
02-19-2007, 03:15 PM
Originally posted by Holtzauge:
Thanks for the positive feedback guys, it's appreciated http://forums.ubi.com/groupee_common/emoticons/icon_smile.gif

How about some input data for the Spitfire and Me109? If I'm gonna model them I need data on prop efficiency and exhaust thrust. The prop efficiency data should have efficiency on the one axis and either speed, advance ratio or propeller speed coefficient Cs on the other axis. The exhaust thrust is usually presented with thrust on the one axis and altitude on the other. The drag rise would be nice to have too ( e.g a chart with Cdo on y-axis and mach up to about 0.8-0.9 on the x-axis).

On the PF FM issue I agree with Icefire that PF has some merits as is but I do not think we need to wait for more powerful computers to the model subsonic drag rise. The FM in PF has to include some sort of acceleration algorithm and all that is needed is to include a "if M > 0.7 then " line and reduce the acceleration abruptly when these speeds are reached. The same goes for propeller efficiency at high speeds: even a simple linear reduction as prop tip speeds go up would get us much closer to reality.

It's not a doomsday machine we are talking about. Just a few lines of code. As Dr Strangelove would have put it: it only requires the WILL to do it.

Sadly, my prognosis is that BoB will stagger us with a beautiful graphic package hiding the same limited FM but I hope to be proven wrong.....

I see you're doing this in as scientific a method as possible.

"if M > 0.7 then"

looks more like basic to me

Really you're going to end up with your approximation vs what can run at playable framerates.
And you probably won't be comparing same-same models but charts volunteered vs 46 planes.

What kind of interval are your FM calcs going to be iterating? 1 sec? 1/some-power-of-10 sec?

At least the curve will be a different shape, that much is certain!

With any luck some of the guys in or graduated from AE school might have a look at the source?

Holtzauge
02-20-2007, 12:35 PM
Originally posted by WWMaxGunz:

"if M > 0.7 then"

looks more like basic to me

Really you're going to end up with your approximation vs what can run at playable framerates.
And you probably won't be comparing same-same models but charts volunteered vs 46 planes.

What kind of interval are your FM calcs going to be iterating? 1 sec? 1/some-power-of-10 sec?

At least the curve will be a different shape, that much is certain!

With any luck some of the guys in or graduated from AE school might have a look at the source?

In the C++ simulation I did I use 1/100 s iteration interval since I do not have to have real-time data. I can afford to wait a couple of seconds http://forums.ubi.com/groupee_common/emoticons/icon_smile.gif

I'm not suggesting that this (shorter iteration interval) has to be done to improve the PF FM. A simple equation (no iteration needed) modeling an abrupt drag rise would not be perfect but would improve the simulation immensely. The same could be done to reduce prop efficieny at high speeds. Right now it looks like props in PF are exempt from mach effects.

I think having 950 Km/h IAS on the dial at 4 Km resulting from a measly 45 degree dive sucks!

To put it mildly, the modeling in PF definitely has some room for improvement in this area.

Really, I think you and Icefire have a bit of a negative mindset about what is possible to simulate and not and need to perk up a bit http://forums.ubi.com/images/smilies/16x16_smiley-wink.gif

M_Gunz
06-27-2007, 11:54 AM
Holtzuage, you have good elements of simulation there. And we know that no sim, IL2 included,
can be perfect. The big deal is how much error and where and then importance of that to the
sim and scope of the sim -- at least that's what I can come up with!

I have been told I am wrong on this with no actual grounds, that it is very possible that as
in reality the IAS guages read high at high speeds. I say that because of the detail level
used in for example the slip ball guages we have. Oleg did write to me (gasp, oh no, I can't
PROVE this!) about the slip guages that the model takes into account what the ball is made of
and what particular fluid was used BY COUNTRY. The ball moves as real, it is not just an
indicator of actual slip (slip in devicelink is position of the indicator, not always actual
slip as devicelink is made for supporting external instruments or painting guages on screen
rather than raw data output, BTW) that is only updated infrequently. Knowing THAT I have no
trouble thinking that MAYBE our IAS guage and speedbar read as the real one would and need
conversion to get CAS before any TAS is possible to find. So is the plane really going 950?

What the WW readout means... is that raw data or not? If to posit a bug in compression that
shows supersonic planes as has been shown is it not just as likely a bug in the WW TAS which
is not entirely consistent anyway?

I've said it before how really GOOD it would be to have some answers from Oleg (even if
through someone else to translate for purposes of saving Oleg time!).