Youen

[old thread] Trajectories : atmospheric predictions

Recommended Posts

Kerbalstuff is still showing 1.4.3 and not 1.4.5.  The link on the first page still shows it as a proper link to get the 1.0.5 compatible version.  People will need to use Github only for now.

Share this post


Link to post
Share on other sites

I've made a file for CKAN stock aero users to upgrade to the new version that should work fine with future upgrades. File-> Install from .ckan should add the data. If you've got the previous version installed thanks to an earlier KSP, you'll need to upgrade as a second step.

Share this post


Link to post
Share on other sites

I was hoping to quickly fix the remaining issues with FAR, but unfortunately couldn't find the time to do that in the past few days. This is why I didn't release the new version on kerbalstuff, and didn't update the version info for AVC either. I don't want people who use FAR to install the new version and get their ships broken as soon as they look at the map view. I'll let you know as soon as I have news on that issue.

@MAV3RICK I'm not directly involved with CKAN, they do their stuff on their own, and apparently do a good job at that because all the mods seem to be always correctly referenced at the correct version :-) The best they could do at this time would be to mark Trajectories as incompatible with FAR, but allow installation for stock aerodynamics ; but that's probably not worth the trouble as the situation should be completely fixed in the coming days (I hope).

On 12/12/2015, 03:25:23, SlabGizor117 said:

Awesome!  Is there an ETA on the non-map trajectory prediction?

No, and as far as I know, no one is even thinking about planning to work on that feature at this time. But anyone with programming skills is welcome to.

Edited by Youen

Share this post


Link to post
Share on other sites

@Youen, thanks for the reply - I thought it was the mod developers who had to update the metadata for CKAN to show it is compatible? I figured the FAR incompatibility might be why though. Thanks again!

Share this post


Link to post
Share on other sites

Any idea when a FAR compatible version will release?

Share this post


Link to post
Share on other sites
On ‎14‎.‎12‎.‎2015‎ ‎12‎:‎47‎:‎49, Youen said:

No, and as far as I know, no one is even thinking about planning to work on that feature at this time. But anyone with programming skills is welcome to.

has been done & tested before - but never released...

 

edit: Found: it was @BahamutoD:
https://www.youtube.com/watch?v=jrUpbGvXrnQ
- well its even on your wishlist already :D
no chance to get what bahamuto has done yet and complete it?

Edited by Speadge

Share this post


Link to post
Share on other sites

Just a short message to let you know I'm still investigating the issues with FAR. I've found out it's commit c284239493b837ab3e649660252416eadd6aab9d in the FAR codebase that breaks vessels when used in the prediction. The issue is easy to reproduce with the Aeris 4 stock vessel, just take off, go to map view, and the vessel explodes instantly if Trajectories is installed and enabled.

More specifically, these two lines added in FARWingAerodynamicModel.cs seem to be causing the issue:

part.dynamicPressurekPa = 0.0005 * density * velocity.sqrMagnitude;
part.submergedDynamicPressurekPa = part.dynamicPressurekPa;

I'll try to understand exactly what is going on, and prepare a pull request for Ferram to fix the next FAR release.

Edited by Youen

Share this post


Link to post
Share on other sites

Trajectory prediction is not showing up (at least not on Kerbin). Also log is spammed with "[Exception]: MissingMethodException: Method not found: 'DragCubeList.ForceUpdate'." (taking into account that error counter in Trajectories windows keeps going up, I think it's the cause). Any ideas where to look? I'm using stock aero, 1.0.5.

Looks pretty similar to this post, though without exception detector stuff.

Edited by Mystique

Share this post


Link to post
Share on other sites
On 12/18/2015 at 8:02 AM, Youen said:

Just a short message to let you know I'm still investigating the issues with FAR. I've found out it's commit c284239493b837ab3e649660252416eadd6aab9d in the FAR codebase that breaks vessels when used in the prediction. The issue is easy to reproduce with the Aeris 4 stock vessel, just take off, go to map view, and the vessel explodes instantly if Trajectories is installed and enabled.

More specifically, these two lines added in FARWingAerodynamicModel.cs seem to be causing the issue:


part.dynamicPressurekPa = 0.0005 * density * velocity.sqrMagnitude;
part.submergedDynamicPressurekPa = part.dynamicPressurekPa;

I'll try to understand exactly what is going on, and prepare a pull request for Ferram to fix the next FAR release.

Thank you for all your time and effort in fixing the issue with FAR. The dedication is greatly appreciated as well as the status updates! :D 

Edited by Svm420

Share this post


Link to post
Share on other sites
10 hours ago, Mystique said:

Trajectory prediction is not showing up (at least not on Kerbin). Also log is spammed with "[Exception]: MissingMethodException: Method not found: 'DragCubeList.ForceUpdate'." (taking into account that error counter in Trajectories windows keeps going up, I think it's the cause). Any ideas where to look? I'm using stock aero, 1.0.5.

Looks pretty similar to this post, though without exception detector stuff.

Do you use the latest version of Trajectories from github? (the one marked as pre-release, 1.4.5, that you can download here). The version from Kerbal Stuff is not compatible with KSP 1.0.5 yet.

Share this post


Link to post
Share on other sites

Also, I forgot to tell, Trajectories 1.4.5 should work fine with KSP 1.0.5 and FAR v0.15.5.2_Helmbold.

The vessel explosion bug has been introduced in FAR v0.15.5.3_von_Helmholtz, but obviously new FAR versions also fix bugs and improve things, so the choice is yours: stock aerodynamics, use an old FAR version, or don't use Trajectories (or wait until the explosion bug is fixed).

Edited by Youen

Share this post


Link to post
Share on other sites
On 12/20/2015 at 10:48 AM, Youen said:

Do you use the latest version of Trajectories from github? (the one marked as pre-release, 1.4.5, that you can download here). The version from Kerbal Stuff is not compatible with KSP 1.0.5 yet.

Yep, seems like that did the trick, thanks :)

Edited by Mystique

Share this post


Link to post
Share on other sites

so I was just quickly messing with things and it seems like the pre-release 1.4.5 works with FAR 0.15.5.4 "Hoerner". Ive only shot a rocket into space and back but it seems to land in the right spot with no kraken attacks... is there something specific that makes it happen or is everything working well now?

Share this post


Link to post
Share on other sites
On 12/18/2015 at 8:02 AM, Youen said:

Just a short message to let you know I'm still investigating the issues with FAR. I've found out it's commit c284239493b837ab3e649660252416eadd6aab9d in the FAR codebase that breaks vessels when used in the prediction. The issue is easy to reproduce with the Aeris 4 stock vessel, just take off, go to map view, and the vessel explodes instantly if Trajectories is installed and enabled.

Hah, so that's what was causing the weird bug! I never would have guessed that Trajectories/FAR would be the culprits. Thanks for the heads-up, Youen!

Share this post


Link to post
Share on other sites
On 12/19/2015 at 5:33 PM, Svm420 said:

Thank you for all your time and effort in fixing the issue with FAR. The dedication is greatly appreciated as well as the status updates! :D 

I'll second this.  I'm really looking forward to having FAR and Trajectories playing nicely together again!

Share this post


Link to post
Share on other sites
On 20/12/2015 at 5:33 PM, RedDestiny said:

so I was just quickly messing with things and it seems like the pre-release 1.4.5 works with FAR 0.15.5.4 "Hoerner". Ive only shot a rocket into space and back but it seems to land in the right spot with no kraken attacks... is there something specific that makes it happen or is everything working well now?

I think the vessel explosion will only occur if it has wings.

EDIT: actually, you're right, the problem has been fixed in Hoerner. I should've tested each version before digging in commits. So it seems the only problematic version is von Helmholtz...

Edited by Youen

Share this post


Link to post
Share on other sites

I've released v1.4.5 on Kerbal Stuff, and updated version information in KSP-AVC as well. Just remember to update FAR if you currently have version von Helmholtz.

Share this post


Link to post
Share on other sites
On 12/21/2015 at 11:09 PM, Youen said:

I've released v1.4.5 on Kerbal Stuff, and updated version information in KSP-AVC as well. Just remember to update FAR if you currently have version von Helmholtz.

@Youen, I just wanted to thank you for your wonderful mod. Together with FAR, it works extremely accurately. I could set up my Kerbin deorbit just after exiting Mun's SOI with Body mode (these "crazy" lines make a lot of sense when you figure them out), and deorbit with a capsule (without any wings or anything) on top of VAB building. Talk about accuracy. Amazing job, Youen. Thank you again for providing this exceptional tool.

The only thing I would like to suggest, is landing prediction in Body-fixed mode after SOI changes. Right now, predictions end up when trajectory hits atmosphere if that body is outside current SOI. If predictions would work even when planning that far ahead, it would make it possible to make accurately Mun/Minmus return to Kerbin burn, timing direct reentry to KSC. Right now, you can only know if the KSC will be on the right side only when returning to Kerbin's SOI, which is basically a little too late to change periapsis timing unless you pack a lot of extra dV.

Also, having these inter-SOI landing predictions would allow timing direct reentry at interplanetary trajectories, for example, timing arrival at Duna for a direct reentry to a specific spot (like a base) months before actual arrival, instead of traditional step-by-step aerocapture and reentry. This would allow for more time and fuel efficient, as well as more challenging transfers.

 

Just a suggestion. Please let me know if you'd consider such an idea.

Share this post


Link to post
Share on other sites

I think it's possible, actually I disabled it because it was buggy but it used to almost work if I remember correctly. The only problem is the lack of time to make it work correctly (as you may have guessed by the time it took me to figure out how to fix a bug that was actually already solved before I even started to look at it).

Share this post


Link to post
Share on other sites
On Wednesday, December 02, 2015 at 10:41 PM, Vorg said:

The only problem I'm having with it is it's based on the current ship, not how the ship will be when you do a de-orbit burn and do another staging. It needs a way to run based on the craft after it has staged one more time.

I just thought about a workaround for that staging issue: you could use docking ports to decouple your capsule, place a manoeuver node with the correct reentry trajectory, dock again, make your burn, and then undock again... A bit complicated of course.

Share this post


Link to post
Share on other sites
On 2014/12/23 at 4:25 AM, Youen said:

Someone reported it is not compatible with real solar system, but a very quick test showed it worked for me ; let me know your findings

In my case it doesn't work when I set my shuttle's Periapsis to 60km in order to have a long and flat trajectory toward the "Kerbal" Space Center at Cap Carnival. It says "Predict stopped. Too many literations←I can't remember that word." and ends the predicted trajectory midway in the atmosphere.

It was with the last version of it, and I am testing with this version.

Share this post


Link to post
Share on other sites
5 hours ago, Youen said:

I think it's possible, actually I disabled it because it was buggy but it used to almost work if I remember correctly. The only problem is the lack of time to make it work correctly (as you may have guessed by the time it took me to figure out how to fix a bug that was actually already solved before I even started to look at it).

That's great. I realize this probably isn't a priority, but if you ever have time to make this work, it would be cool. Thank you again, and good luck!

Share this post


Link to post
Share on other sites

This has likely been asked before but, is it possible to use this with the maneuver planner, or does it still only work with the current trajectory?

Share this post


Link to post
Share on other sites
On ‎23‎/‎12‎/‎2014 at 6:25 PM, Youen said:
  • It is not possible at this time to predict a trajectory for a future stage. The prediction is made for the current vessel only. This is a most wanted feature but unfortunately also quite complicated to implement

This saddens me, but your tool is already amazing.

Share this post


Link to post
Share on other sites
On 28/12/2015 at 3:47 PM, Wolfwood said:

This has likely been asked before but, is it possible to use this with the maneuver planner, or does it still only work with the current trajectory?

It will use your maneuver nodes to draw the predicted trajectory, but it won't alter the stock trajectories. So for example, after an aerobrake, the stock trajectory diverges from the predicted one, and you can only place a node on the stock trajectory. But if you do so, you'll see the result on the predicted trajectory. Maybe that's not clear, but then the best way is you try and see by yourself :)

On 23/12/2015 at 10:29 AM, 01010101lzy said:

In my case it doesn't work when I set my shuttle's Periapsis to 60km in order to have a long and flat trajectory toward the "Kerbal" Space Center at Cap Carnival. It says "Predict stopped. Too many literations←I can't remember that word." and ends the predicted trajectory midway in the atmosphere.

It was with the last version of it, and I am testing with this version.

Indeed, when a trajectory is very long inside atmosphere, the prediction stops at some point because these calculations are expensive. It has been suggested to add a setting to change the maximum predicted time, but haven't had time to do that yet. Also, the longer your in-atmosphere trajectory is, the more inaccuracies you'll accumulate.

Share this post


Link to post
Share on other sites
Guest
This topic is now closed to further replies.