Jump to content

Modified cfg and air for DG808S


bruin

Recommended Posts

Well Guys (and dolls),

Being fed up with the default aerodynamics of the DG808S I had a close look to what's wrong with it and compared the data in the cfg and the airfile with data of the real life thing.

I ended up with modified files with close to the real life data and (after tweaking the drag parameters a bit) resulting in:

- L/D 50 at 132km/h (with waterballast, flaps neutral)

- L/D 39 at 153km/h (with waterballast, flaps neutral)

- L/D 23 at 188km/h (with waterballast, flaps negative)

Do you think I can publish the files here? Or is there a chance MS will try to lock me up in jail for it?

Bert

Link to comment
Share on other sites

This is a *great* discussion topic. I'm sure we can circulate a tweaked .air and .cfg file without Microsoft getting their knickers in a twist, as afaik all .air files are based ultimately on a tweaked MS original anyway.

I could write a vario that would help tune the parameters by displaying L/D in realtime - would that help?

Did you tweak the .air file or alter the text parameters in the .cfg ?

Did you look at the effect of the airbrakes also? I'm interested to know if it's possible to both reduce their effect (that I'm sure) and define some parameter to have the effect of dropping the nose a bit when the airbrakes are open - the real airbrakes both increase drag and reduce lift and the lift reduction is slightly forward of the C of G so the nose going down compensates for the speed reduction of the increased drag. In the FSX DG you just have the huge drag increase and the speed dies away, which would be fatal in a real glider.

A couple of comments on which parameters to tweak to adjust gliding performance would be helpful.

good stuff - Ian

Link to comment
Share on other sites

@ Peter:

Thanks for the article. Reading it I think it's a yes to "may I publish?", provided I include the MS-asset statement in the redme.

@ Ian:

I could write a vario that would help tune the parameters by displaying L/D in realtime - would that help?
Thanks for the offer. Yes it may help, but don't invest much of your valuable time in it. I can live with the "IGC to SeeYou-method".

Did you tweak the .air file or alter the text parameters in the .cfg ?

Yes, both files needed to be changed. The fuselage-length and the position of CoG in the .air file were the essential ones in the .air, the drag-parameters were the main things in the .dfg.

A couple of comments on which parameters to tweak to adjust gliding performance would be helpful
Too much to write down in this post. But basically I started with gathering specs of the real thing, changed the .air and .cfg according to that data as much as possible. The parameters I had no RL-info of, I left as they were. And finally I tuned the L/D-performance with the parasite- and induced- drag parameters in the .cfg (For the Astir I also had to tweak the cruise_lift_scalar to get a good result).

Did you look at the effect of the airbrakes also?

Nope, not yet. I'll post the file and than we can discuss what further improvements we would like and see if we can discover how we achieve them.

Bert

Link to comment
Share on other sites

Here is version 1.0 of DG808S modified flightdynamic files.

Any comments are welcome

Bert

Edit:

The attached zip is removed, due to updates in the following posts in this topic

Link to comment
Share on other sites

Bert - can you comment on what you changed? I see in the aircraft.cfg we have

[flight_tuning]
cruise_lift_scalar     = 1.0
parasite_drag_scalar   = 0.75  //was 1.0
induced_drag_scalar    = 0.75  //was 1.0
elevator_effectiveness = 1.0
aileron_effectiveness  = 1.0
rudder_effectiveness   = 1.0
pitch_stability        = 1.0
roll_stability         = 1.0
yaw_stability          = 1.0
elevator_trim_effectiveness = 1.0
aileron_trim_effectiveness  = 1.0
rudder_trim_effectiveness   = 1.0

We want to try and have everyone on the same version of the flight config so my suggestion is we include 'beta' in the name of these ones (or alpha :)

Ian

Link to comment
Share on other sites

Ian,

When you compare the original aircraft.cfg with the one I have cooked, you can see there are more changes than the one you posted.

Firstly I made changes:

- in the section [airplane geometry];

- in the section [waterballast_tanks].

- In the new aircraft.cfg there is also a new section at the bottom called [airwrench]. That section shows the results I made in the .air file. With the progam Airwrench.exe I changed the fuselage_length and the CoG-position, which were not properly defined in the original .air file.

- I may have made more changes. I cannot tell for sure, because in the proces I may have lost the original FSX-files.

- Finally I had to make changes in the section flight tuning. After having made the other changes I had to reduce the drag parameters, otherwise the L/D would have been be too optimistic.

As for adding beta of alpha in the name: the name of the aircraft.cfg can not be changed. The .air can, but it requires also a change in the .cfg file to let the .cfg refer to the right .air file. But I don't believe there is a way to prevent cheaters to change those two files further to get better contest-results. And to be honest, I don't want to spend my time on trying to prevent that.

Bert

Link to comment
Share on other sites

A hash code of the active aircraft.cfg and the .air-file could mean something, perhaps provided by Ian's logger in simprobe.

Still, it wouldn't prevent others from doing the same what CumulusX!/Simprobe are doing with lift.

Cheers,

Peter

Edit: Related to publishing the modifications, I think that there shouldn't be any issue with the CFG-file at all, because it's open text anyway. I'm not 100% sure about the .air file, since MS earlier reasoned, that its format is not exhibited because of IPR constraints with the manufacturers of the aircraft, of which corporate know-how is included. Because there is also a statement about reverse engineering, someone could be critical how far reverse enegineering is a pre-requisite for modifications of .air-files.

Airwrench is most probably on the basis of reverse engineering, but you only used this tool rather than developing this tool yourself. At the end of the day, there are lots of modified air-files out in the wild and by now I've not yet heard about legal actions against it. Yet, I'm not a lawyer ...

Link to comment
Share on other sites

okeydoke here's the 'diff' of the aircraft.cfg:

[flight_tuning]
cruise_lift_scalar     = 1.0
parasite_drag_scalar   = 0.75  //was 1.0
induced_drag_scalar    = 0.75  //was 1.0
elevator_effectiveness = 1.0
aileron_effectiveness  = 1.0
rudder_effectiveness   = 1.0
pitch_stability        = 1.0
roll_stability         = 1.0
yaw_stability          = 1.0
elevator_trim_effectiveness = 1.0
aileron_trim_effectiveness  = 1.0
rudder_trim_effectiveness   = 1.0

// ...snip

[Water Ballast System]
//Tank.n =  Gallons,  Long,    Lat,   Vert,  iValve

//Tank.0  =  7.79, -2.75,   0.0, 0.0, 1          //Front Fuselage
//Tank.1  =  3.57, -3.28,   0.0, 0.0, 2          //Rear Fuselage

//Tweaks: waterballast tank.0 and tank.1 modified according to specifications real life
Tank.0  =  7.93, -2.75,   0.0, 0.0, 1          //Front Fuselage
Tank.1  =  3.43, -3.28,   0.0, 0.0, 2          //Rear Fuselage

Tank.2  =  9.25, -0.60, -10.5, 0.0, 2          //Left Outboard
Tank.3  = 16.38, -0.66,  -4.5, 0.0, 1          //Left Inboard
Tank.4  = 16.38, -0.66,   4.5, 0.0, 1          //Right Inboard
Tank.5  =  9.25, -0.60,  10.5, 0.0, 2          //Right Outboard

NumberOfReleaseValves = 2
DumpRate = 0.18494                            //Gallons per second

[AirWrench]                      //Added by airwrench after updating fuselage length and CoG
Aircraft_Length= 22.572
nose_position_lon= 6.529

So these changes are relatively few - but I'm sure they move the actual performance in the right direction. I've taken a look at the 'spoilers' parameters in the air file and it seems you can adjust both the lift/drag effectiveness and the rotation 'moment' (i.e. pitch down).

As a thought, it might be worth using AirWrench to *display* the air file parameters, but consider editting them with AirEd. The rationale is simply that AirEd is a simple editor that only changes a single value in the air file at a time (well, there's a one-to-one mapping between the field you edit and a corresponding value in the file) while AirWrench (I think) makes multiple simultaneous changes within the file so you are slightly disconnected from the changes you're making. Probably the optimum is using a bit of both but it's early days for me looking at this stuff...

AirWrench is here: http://www.mudpond.org/AirWrench_main.htm

AirEd (v1.52) is here: http://pagesperso-orange.fr/hsors/docs/aired152.zip

You need an updated AirEd.ini file, supplied with AirUpdate here: http://www.mudpond.org/AirDat.ZIP

Regarding the spoilers, in the .air file there are parameters for:

Spoiler Drag (Cd_ds=205)

Spoiler Lift (CL_ds=-0.01)

Spoiler Pitch (Cm_ds=-0.015)

in AirWrench these are normalised to

Spoiler Drag = 100.1

Spoiler Lift = -1.0

Spoiler Pitch = 15 (Nose Up)

as far as I can tell the relationship between the AirWrench values and the 'real' .air file values is 1:1 and linear.

So for spoilers I'm guessing we should

* decrease the drag value,

* make the lift value more negative,

* and make the pitch 'Nose Down' to some amount...

I'll do a couple of experiments.

I'd suggest another area of the flight envelope worth looking at is at low speed - the stock DG flies very nose-up and the stall speed seems set very low.

Re the 'beta' tag, I just meant in the name of the zip file to minimise future confusion - I know it's impossible to stop 'cheating'.

Ian

Link to comment
Share on other sites

small note - looking at the air file for the Wolfgang Piper LS8-18, the spoilers have a 'Nose Down' parameter of 20 rather than the stock FSX DG with 'Nose UP' = 15 - they can't both be right and my faith is in Wolfgang...

Ian

Link to comment
Share on other sites

[AirWrench]//Added by airwrench after updating fuselage length and CoG

Aircraft_Length= 22.572

nose_position_lon= 6.529

I think the above changes in the .air file made the major performance-difference (from L/D 24 to L/D above 50).

As far as I have seen these changes can only be made with airwrench (see tab balances. In the evaluation-version of airwrench those are the only parameters you can really change). In the .cfg and in aired.exe I was not able to find the parameters for it.

Yesterday I found that in the uploaded version there was something wrong with the shape of the wing (too much arrow-shape). I am not on my home computer right now, so I cannot give more details right now on how to change that.

Bert

Link to comment
Share on other sites

Bert this is *great* stuff - I think we can really make a difference. My suggestion is we aim for a single target 'virtualsoaring.org' upgrade package for the DG808S - e.g. new .air file, new aircraft.cfg and I'd like to suggest we'd include a new panel.cfg at the same time so we can incorporate at the very least a total energy vario.

I've built a TE vario that also displays L/D in real-time, and included it in this zip with your aircraft.cfg and .air files.

This includes your aircraft.cfg and dg808s.air, as well as a new panel.cfg and the gauge. Simplest is to just drag-and-drop the folders into FSXMAIN but as before you can browse the zip and move the individual files.

The zip file actually includes the *original* dg808s.air and aircraft.cfg (renamed) to help you out if you've lost yours. As these are useless without FSX I'm assuming Microsoft won't come after me too aggressively, plus the readme includes the 'usage' words.

Note - I've done a binary comparison of Bert's dg808s.air and the original, and they are *identical*, so all the changes are so far via the aircraft.cfg.

I have *not* incorporated my 'spoiler (=airbrake)' changes into the .air file yet as they need proper testing, but the early indications are it will be easy to make a basic improvement along the lines I suggested earlier and as already implemented in Wolfgang Piper's aircraft.

To illustrate the current state of the 'beta' here's the L/D comparison AT ZERO FLAP, FULL BALLAST:

knots    L/D original    L/D beta 00 (Bert's aircraft.cfg so far)
50        49                 73
60        42                 61
70        35                 46
80        28                 37
90        22                 29
100      18                 25

These measurements are within maybe +/- one or two - i.e. this is just an indication at this stage that we're making progress.

We need to work out how to bend down the polar at the low-speed end of the polar.

Ian

virtualsoaring_dg808s_beta_00.zip

Link to comment
Share on other sites

My suggestion is we aim for a single target 'virtualsoaring.org' upgrade package for the DG808S - e.g. new .air file, new aircraft.cfg and I'd like to suggest we'd include a new panel.cfg at the same time so we can incorporate at the very least a total energy vario.
Maybe it is better to include a complete new panel folder (like panel.TE) with your modified panel.cfg and TE-vario in it.and leave the panel folder as it is. That way people can choose between the original (or Peter's compensated) vario and your TE, without having to make changes in the panel.cfg, but in the [FLIGHTSIM]-section.

Note - I've done a binary comparison of Bert's dg808s.air and the original, and they are *identical*, so all the changes are so far via the aircraft.cfg.

It may sound like I've lost my mind by saying this, but if you compare the two air.files, the .air file appears to interact on the active aircraft.cfg and vice versa. In other words: if you don't re-activate the original aircraft.cfg also, the original .air file will show the same values as the new one. That may be the reason you didn't see differences between the original and the new .air file.

This strange behaviour I found out when I implanted the new DG808S.air into my Astir folder in which I have an aircraft.cfg with the Astir-data and simply renamed the DG808S.air to Astir CS.air. Reading the .air with aired.exe after the implant/rename I saw that most of the values in the .air had already been changed to the Astir values in the aircraft.cfg.

I definitely made changes in the .air file, like I mentioned earlier (see also the added [airwrench]-section in the aircraft.cfg)

Bert

Link to comment
Share on other sites

include a complete new panel folder (like panel.TE) with your modified panel.cfg and TE-vario in it.

Great idea - that way it's just a folder copy and the aircraft.cfg is being changed anyway...

Re: dg808s.air

eeek - does FSX overwrite the .air file ? All I did was rename my dg808s.air to dg808s_original.air, install your version as dg808s.air, and then compare the two with the dos prompt 'comp' command to do a binary compare and they came out the same. Somehow I've got to find a *definite* original copy of dg808s.air and do another compare if there's a chance mine was overwritten.

Ian

Link to comment
Share on other sites

I think there is some bug in the AirUpdate.exe which is used by AirEd, too. Sometimes it required several attempts before the data got actually written in the file.

You may also consider working on the lift coefficient tables either at 404 or 1546 (the latter is perhaps for turbopropos only). This allows better to realise the correct pitch over speed than the flight tuning entries (zero-lift pitch and curve steepness), and also transition to stall (maximum lift coefficient). Btw, we used to have the DG100 in our club which always flew pretty much "nose-in-the-sky" during approach which got more than one cadet a bawl, because of flying "too slow".

regards,

Peter

post-176-1223315218.png

Link to comment
Share on other sites

eeek - does FSX overwrite the .air file ? All I did was rename my dg808s.air to dg808s_original.air, install your version as dg808s.air, and then compare the two with the dos prompt 'comp' command to do a binary compare and they came out the same. Somehow I've got to find a *definite* original copy of dg808s.air and do another compare if there's a chance mine was overwritten.
Reïnstall the original aircraft.cfg and your .air will go back to original, however the new one will be changed to the original.air as well. You can start from the beginning by getting back to the original .cfg and .air and reinstall the new files from the .zip file I published.

I think there is some bug in the AirUpdate.exe which is used by AirEd, too. Sometimes it required several attempts before the data got actually written in the file.

Yes, there sometimes is some spooky stuff going on when working on the .air file. On some occasions I changed something in aired.exe, saved the file, reopened it in aired.exe and saw the changes I made. But when I opened the changed .air in Airwrench.exe the old values were still there. Beats me...

You may also consider working on the lift coefficient tables either at 404 or 1546

Regarding to those parts of the .air I had to surrender. They go way beyond my knowledge so I decided to leave them to real experts.

BTW Ian, the L/D-reading in your vario works fine for testing-purpose.

Bert

Link to comment
Share on other sites

I will host a session on MP in the advanture lobby (Session ame SOAR, host is BravoRomeoTango). If you join in we can chat, that will make communication go quicker.

Bert

Link to comment
Share on other sites

I can't make the hosted session tonight...

I'm glad the L/D gauge is working ok - it certainly makes it easy to check the flight performance at any point.

I think we should try and tweak the 404 table a bit - Peter's graphic explains it pretty well but I'm sure there'll be hard-to-comprehend collateral impact... at the very least we can compare the dg808s one with those from other gliders to see how much they vary. Assuming your testing confirms the high L/D at <50 knots we have to reduce the CL at that speed (i.e. at high angles of attack). The updated polar at higher speeds seems pretty good to me.

It was easy to improve the airbrakes/spoilers - reducing the drag (new Cd_ds=80) and increasing the 'negative lift' (new CL_ds=-0.04). So far I've fixed the sudden drop-off in speed when they open and the airbrakes are no longer speed-limiting and I've made sure when fully open the spoilers give a glide ratio of about 7:1 at normal approach speed. But like the polar, I have not enough sink at very low speeds (high angles of attack). I'm wondering if the solution will be to fix the 404 table rather than just for the spoilers.

Ian

Link to comment
Share on other sites

I would say, start with the 404 to make the "normal" flight characteristics as real as possible, than add on the spoilers with drag and lift. I have no personal experience with the 808, but the earlier DG's were notorious for their weak spoilers. They 1000 is again pretty good.

Basically, 100's achieved more or less the required 7:1, but the experience of sink was only sufficient at exactly the recommended approach speed, which was around 92-95 km/h for the 100. When you was faster, then sink didn't at least increase that much that one could expect from most other gliders.

In addition, ground effect was really significant. If you aproached beyond 100 km/h you could expect a looooooong flare.

Btw, the more negative lift, the more you have to pull up nose for steady flight.

regards,

Peter

Link to comment
Share on other sites

Ian,

I will leave changing the 404 up to you then, for as I said I have no clue what to do.

On my side I did change the wing_sweep parameter in the [airplane geometry]-section of the .cfg to:

wing_sweep = 1.0 // was 5.3 (Degrees, wing leading edge). those 5.3 degrees was just wrong (I don't know how it got there in the first place), in airwrench the wing looked like one of an F16.

To get somewhat equal performance as before I also upgraded the drag parameters to:

parasite_drag_scalar = 0.8

induced_drag_scalar = 0.8

If you change the above three parameters in your .cfg we will be equal again. Changing the .air with these changes is not necessary.

BTW: Flying with your panel.cfg I can't open the GPS.

Bert

Link to comment
Share on other sites

Flying with your panel.cfg I can't open the GPS

oops - that's where the SDI C4 goes in my development for the Aerosoft Discus... I've fixed panel.cfg and I'm in the process of creating a new beta_01 zip file which has the air changes discussed so far and the gauges implemented as a complete separate panel as you suggested...

Ian

Link to comment
Share on other sites

Bert - updated beta_01 produced and is in your email....

it combines the changes you've mentioned plus tweaks from me for the airbrakes and the lift-vs-angle-of-attack table (low-speed area). Also I've shifted my gauge stuff into a separate panel and made the air file changes applicable to a 'variant' of the DG808S only so the original airfile doesn't get impacted. When the beta goes to prod we can discuss whether to use a variant or simply alter the main dg808s.air... for beta_01 the air file is in virtualsoaring.air and the panel is in panel.virtualsoaring.

My limited flying so far suggests the 'virtualsoaring' DG already feels better than the FSX default and we've done the right thing in only tweaking a few things.

I'm planning on improving the set of instruments - mainly with the Cambridge vario going to Netto and the Winter to TE - it will be a very convenient package for anyone to install including the air/config changes. I think we have *one* shot at producing a good package - if we end up with multiple variants of the 'virtualsoaring' DG we'll be worse off rather than better.

We should collect any other comments on the air model of the stock FSX DG - we've pretty much fixed the polar and the airbrakes - I'm dubious about the stall behaviour but I'm not sure this is fixable with our knowledge.

Ian

Link to comment
Share on other sites

status so far:

* polar - looking pretty good - a big improvement on the default FSX DG

* airbrakes/spoilers - also now adjusted to a more realistic performance. Warning.. they're not speed-limiting in a vertical dive any more!

* flaps - reasonable progress here, an improvement over the default anyway.

* instruments - TE vario done so far.

To do:

* finish the instruments - the idea is to include Netto (on the Cambridge) as well as TE (on the Winter).

* a decent amount of testing

In a nutshell, we've already got a glider we could 'ship' but we'll try and round off the package this weekend. It's actually been fairly tricky to make the polar more accurate at *both* ends of the speed range, and also have the flaps perform the way they should. But the default FSX DG is miles off so it was not hard to do better than that...

Ian

Link to comment
Share on other sites

Ian,

Suggestion for the to do list:

- rewrite all my glider missions , because with the modified DG, the old ones will be a peace of cake :winks:

Bert

Link to comment
Share on other sites

Guest
This topic is now closed to further replies.
×
×
  • Create New...

Important Information

We have placed cookies on your device to help make this website better. You can adjust your cookie settings, otherwise we'll assume you're okay to continue. Privacy Policy & Terms of Use