Board index FlightGear Development Spaceflight

Space Shuttle - Development

Discussion about development and usage of spacecraft

Re: Space Shuttle - Development

Postby eatdirt » Wed Jun 30, 2021 8:39 pm

I'd suggest to check why you're exploding, aka do 'limits' to 'soft'


God, of course... thanks!

Now I know, it is simply a "orbital structure exceeded" and the reason is the trajectory. For some reason, at the time of the expected MECO, the Shuttle pitches up. It actually sorts of follow the display. The problem is that if I am a bit late at MECO, it pitches up like crazy and starts looping as hell very fast. No surprise I am exploding almost instantaneously. This is happening while flying late a manual MECO. But it also occurs in full AUTO with only one engine running. This situation happens sometime if a failure scenario leads to 2 engines lock-up, I am switching throttle to manual to cut them off before too much gs build up, then move back to AUTO to get the last running engine doing his job for MECO, and bang.

PS: I am running next now, but I can do some test on 2020.* as well.

Edit: Just realized that it may be the preset ascent stage 2 in the mission file, it has indeed a kick in the last point. I am toying with it, adding a point at same altitude to see if that improves the final attitude.
eatdirt
 
Posts: 1012
Joined: Wed Aug 15, 2018 3:06 pm

Re: Space Shuttle - Development

Postby eatdirt » Wed Jun 30, 2021 11:10 pm

All right, interesting. So the pitching up at MECO is indeed because the last point in ascent traj 2 is really at normal MECO. If I add another one, number 15 here, at greater speed, the attitude slightly past normal MECO is fine.

Code: Select all
<point n="14">
      <x type="double">25800.0</x>
      <y type="double">345000.0</y>
   </point>
   <point n="15">
      <x type="double">30000.0</x>
      <y type="double">345000.0</y>
   </point>


But the explosion is not due to this. Two snapshots, voluntarily passed normal MECO in throttle manual, the RCS are firing and the Shuttle is put in CSS while the engines are still on. I suspect, it transits to another DAP? Maybe that's the joystick bypassing every logic...again :-/

Two snapshots separated by a fraction of second (recorded at 5 fps, the next one is fully red and I explode).

Image

Image
eatdirt
 
Posts: 1012
Joined: Wed Aug 15, 2018 3:06 pm

Re: Space Shuttle - Development

Postby Thorsten » Thu Jul 01, 2021 7:23 am

The problem is that if I am a bit late at MECO, it pitches up like crazy and starts looping as hell very fast.


That kinda sounds like the CoG/vectoring issue when the engines can't gimbal enough to avoid a torque - except in your screenshots you have 3% propellant left AND the center engine is good, so that's not when this is expected to be triggered.

So let me try to clarify - the issue is when you're on manual throttle but automatic pitch/roll? That's what the screenshots seem to indicate...

I'd guess what happens is in this case that pitch/roll guidance is undefined because the code terminates when reaching MECO, so the last signal sent to the engines is 'stuck' there, i.e. rather than controlling rates, the engines fire in whatever position they were when the code ends.

I don't think I've ever tried manual throttle and the rest on auto, usually I go fully manual and that ought to work out fine because then the signals come from the stick all the time. Could you try switching to fully manual a few seconds before MECO to confirm whether this is the issue?

@GinGin - I'm actually not sure what the expected behavior is when you force engines to run manually past the MECO point while under AUTO guidance - do we simply hold the last attitude? Do we drop AUTO in all channels and force MAN?
Thorsten
 
Posts: 12490
Joined: Mon Nov 02, 2009 9:33 am

Re: Space Shuttle - Development

Postby GinGin » Thu Jul 01, 2021 5:01 pm

So the pitching up at MECO is indeed because the last point in ascent traj 2 is really at normal MECO


I hardly see how it might be linked.
The thing you edited is for the displayed trajectory only. It is not taken account by the guidance in autolaunch.nas.

The mission file parameters that change the ascent guidance logic are the following ones.

*Trajectory loft will change the Meco altitude ( normal one is around 345000 feet)
*Gamma meco will change the targeted flight path at Meco hence the vertical speed ( tan(gamma_meco) x Meco inertial speed).
The lower is gamma Meco parameter , the lower the pitch up moment to have the correct vs before Meco will be.

Code: Select all
<ballistic-climb-ang-bias-deg type="double">0.0</ballistic-climb-ang-bias-deg>
<trajectory-loft-ft type="double">0.0</trajectory-loft-ft>
<gamma-meco type="double">0.5</gamma-meco>



@Thorsten: I am not sure either about the real logic if we don’t manually press the pushbuttons at Meco .I will have a look in the ascent handbook and flight rules about manual throttling.
I would say that we can force the rcs to null the rate to avoid weird things as a first measure.
GinGin
 
Posts: 1580
Joined: Wed Jul 05, 2017 11:41 am
Location: Paris
Callsign: Gingin

Re: Space Shuttle - Development

Postby Thorsten » Thu Jul 01, 2021 6:16 pm

I would say that we can force the rcs to null the rate to avoid weird things as a first measure.


I suspect the issue is that the SSME gimbal angles get stuck - the RCS won't do anything against that kind of force, we need to continue SSME gimbal past nominal MECO (if my analysis is correct...).
Thorsten
 
Posts: 12490
Joined: Mon Nov 02, 2009 9:33 am

Re: Space Shuttle - Development

Postby eatdirt » Thu Jul 01, 2021 9:13 pm

The thing you edited is for the displayed trajectory only. It is not taken account by the guidance in autolaunch.nas.


Good to know, it was pure coincidence then...

I've tried full manual, throttle and attitude, same conclusion, bang. The pitching up is very strong if the center engine is off, and mild if it is on (hence my confusion of what the ascent points were doing).

if MECO guidance is signing off while engine are running, what happens? Do they get folded up? That would explain. That thing exactly happens at Mach 26. I'll read a bit the auto_launch.nas to understand a bit more.

Edit: auto_launch.nas, line 502:
Code: Select all
if (getprop("/fdm/jsbsim/systems/ap/automatic-sb-control") == 1)

The alternate case is not dealt with, so as soon as apoapsis target is reached, MECO signs off, and the rest occurs while engine are still running? That's why I am seeing the RCS firing, that's the next instruction.
Maybe a waiting loop waiting for manual_meco or fuel exhaustion would do here? I really do have a thing for flying the exception :)
eatdirt
 
Posts: 1012
Joined: Wed Aug 15, 2018 3:06 pm

Re: Space Shuttle - Development

Postby Thorsten » Fri Jul 02, 2021 8:16 am

I've tried full manual, throttle and attitude, same conclusion, bang. The pitching up is very strong if the center engine is off, and mild if it is on (hence my confusion of what the ascent points were doing).


That looks like the CoG issue, except we really shouldn't have that with 3% propellant left and a working center engine (in fact, we shouldn't have that at all, I remember going down to near zero propellant in abort tests...)
Thorsten
 
Posts: 12490
Joined: Mon Nov 02, 2009 9:33 am

Re: Space Shuttle - Development

Postby eatdirt » Fri Jul 02, 2021 9:05 pm

That looks like the CoG issue, except we really shouldn't have that with 3% propellant left and a working center engine


Edit: I've made more tests and, possibly some fixes preventing MECO to sign off prematurely: with 3 engines running, it is all good, the shuttle maintains attitude. Only when the *central* engine is off, it pitches-up and gets destroyed. And, that could be expected, especially when the ET is close to empty?
NB: with MECO signing off too early, current dev snapshot, pitching up and exploding was occurring even with 3 engines running. My best guess is really the engines were getting folded up, or at least not gimballing anymore, while still running.

I've made a merge request for the MECO signing-off, testing on thrust, please reject it if you think it is bogus. But that works good for 3 engines, tested till ET at 0%, Mach 28, 3.2g; engines ran out of fuel [and turbo pumps damaged :)]

https://sourceforge.net/p/fgspaceshuttl ... quests/32/

PS: auto TAL might have the same premature signinf off in manual throttle, but I did not test.
eatdirt
 
Posts: 1012
Joined: Wed Aug 15, 2018 3:06 pm

Re: Space Shuttle - Development

Postby GinGin » Sat Jul 03, 2021 12:35 pm

@EatDirt : Do you observe the same behaviour with the stable release ?
GinGin
 
Posts: 1580
Joined: Wed Jul 05, 2017 11:41 am
Location: Paris
Callsign: Gingin

Re: Space Shuttle - Development

Postby eatdirt » Sat Jul 03, 2021 1:43 pm

@EatDirt : Do you observe the same behaviour with the stable release ?


Yes! I also confirm my best guess, it goes fast, but in the external views, the engines are getting folded up while running. That's just it, premature end of auto_launch guidance in manual throttle mode!
eatdirt
 
Posts: 1012
Joined: Wed Aug 15, 2018 3:06 pm

Re: Space Shuttle - Development

Postby GinGin » Tue Jul 13, 2021 4:24 pm

@EatDirt: I merged your commit request.

I fixed the typo in autonomous mode.
Could you try it again on a center MFD ?

I updated the runway coordinates for Gander and Kef for a more precise TAEM.

Image
GinGin
 
Posts: 1580
Joined: Wed Jul 05, 2017 11:41 am
Location: Paris
Callsign: Gingin

Re: Space Shuttle - Development

Postby GinGin » Fri Jul 23, 2021 7:51 pm

I pushed a commit with a slightly updated logic for final flare that avoid some pitchdown near the ground due to low qbar filter.
Also some more code about Entry guidance original simulation not activated for now.


If you like the challenges, you can either try Paro (number 91) or Innsbruck (number 92) for TAEM and final approach scenarios.
A bit of headwind and short speedbrake logic is almost mandatory.
I used those fields to fine tunned some TAEM guidance parameters for final approach flare.

Manual final flare at Paro runway 33 to land really close to the stall speed (20° of AOA) make it possible to not overshoot the runway
Image


Innsbruck case is well handled for both QFU (26/08) with Short Speedbrake selected and a bit of headwind.
Image
GinGin
 
Posts: 1580
Joined: Wed Jul 05, 2017 11:41 am
Location: Paris
Callsign: Gingin

Re: Space Shuttle - Development

Postby Vinny002 » Wed Aug 25, 2021 11:36 pm

Hi guys! How is the FG space shuttle coming along? Thanks!

Cheers,
Vincent
Vinny002
 
Posts: 89
Joined: Mon Jul 01, 2019 9:55 pm

Re: Space Shuttle - Development

Postby GinGin » Fri Sep 03, 2021 7:04 pm

I started some work about the Shuttle original ascent guidance, named Power Explicit Guidance
Basic concept there: https://www.orbiterwiki.org/wiki/Powered_Explicit_Guidance

One major function required for it was the following one: a Conic State Extrapolation Routine.
A routine to conically extrapolate any spacecraft inertial state vector either backwards or forwards as a function of time or as a function of transfer angle. [...]
It serves as a preliminary navigation tool and as a method of obtaining quick solutions for targeting and guidance functions.


Original document there, quite entertaining: https://archive.org/details/NASA_NTRS_A ... 4/mode/2up

I have tested it quite a bit, it seems to work well enough for a fast inertial SV extrapolation with error below a percent.
In red the State Vector at time t0.
In Green the forecasted State Vector 9mn forward.
In Pink, the true State Vector position 9 mn later.

Pink and Green are coherent.
It might be even useful to calculate in advance the Inertial Attitude Position for a burn (?)

Image
GinGin
 
Posts: 1580
Joined: Wed Jul 05, 2017 11:41 am
Location: Paris
Callsign: Gingin

Re: Space Shuttle - Development

Postby eatdirt » Sat Sep 04, 2021 7:58 pm

Hi guys,
since a while we have two HST directories containing effects and texture files, one is supposed to be abandoned:Models/HST. The one currently in use is in Models/PayloadBay/HST. Unfortunately, I've spotted that a few effects in Models/PayloadBay/HST are still pointing to other effects and files which resides in Models/HST. Therefore, I have made some clean-up in that merge request:

https://sourceforge.net/p/fgspaceshuttl ... quests/33/

It renames Models/HST into Models/HST_obsolete and fixes all paths to be self-contained into Models/PayloadBay/HST. Nothing fancy, but that should ease future code and effects development for things in Models/PayloadBay (in prevision for 2020.4.0, encompassing canvas seems to be coming!!).

Cheers,
Chris.
eatdirt
 
Posts: 1012
Joined: Wed Aug 15, 2018 3:06 pm

PreviousNext

Return to Spaceflight

Who is online

Users browsing this forum: No registered users and 1 guest