IMFD 5.1g Autoburn Problems

[Bullethead]

Howdy-

I recently got IMFD 5.1g and have encountered a problem with the autoburn feature. Here's what happens...

1. Set up course change (orbit eject, mid-course, plane change, orbit insert, etc.) as before.
2. Go to the BV page of the Course program
3. Select Autoburn
4. Ship orients in what I assume is the correct direction then fires main engines. So far, so good.
5. When burn time gets down to about 0.4 seconds or so, main engines stop, and the ship goes back into attitude correction, which is usually considerably off what it was burning on. Main engines fire for a split second, then more major attitude changes, etc. This process may continue through 4 or 5 iterations, with the burn time never reaching zero. I usually stop the autoburn program after the main burn is done.
6. No matter whether I stop the autoburn or let it continue as long as it wants to, I never end up in the right place. For instance, when going from Europa to Callisto, I usually end up missing Callisto by about 1 radius of that moon. With prior versions of IMFD, I always hit the target dead on.

 

[jarmonik]

That sounds familiar. You can shutdown the Autoburn when it starts to re-orient the vessel and burn last 0.4 seconds manually with linear-RCS.

BTW, are you using default (stock) DeltaGlider ?

 

[Bullethead]

Thanks for the quick reply.

That sounds familiar.

It didn't happen in 5.1e, which is why I mentioned it.

You can shutdown the Autoburn when it starts to re-orient the vessel and burn last 0.4 seconds manually with linear-RCS.

I've tried that, too, and it's the same problem doing it manually. The problem seems to be that when the remaining burn time gets down to about .4 seconds, the ship can no longer decide which way it's supposed to be oriented. I center the + in the BV screen, and just use taps on KP+ to finish off the burn. But as soon as I touch KP+, the BV screen's + instantly goes all the way to the edge of the screen. So I stop and reorient with the remaining burn time still like 0.35 seconds, and try again. Same thing happens. IOW, I go through the same process that the AB program does, a bunch of split-second main burns divided by many reorientations, all trying to get rid of that last 0.4 seconds of burn time.

BTW, are you using default (stock) DeltaGlider ?

Yes.

 

[jarmonik]

It didn't happen in 5.1e, which is why I mentioned it.

In that case it must have something to do with the fuel-efficiency factor that is implemented in 5.1g it could have an effects you mentioned.

Thanks about reporting this.

 

[jarmonik]

I just tested some burns from Europa to Callisto tutorial but couldn't repeat the problem. Could you post a scenario just a little before the burn.

How much there is dV still remaining when the problem will occur ?

 

[Bullethead]

OK, here's a log from a run I just did. Included in the zip is the scenario I'm using. It was made by Rdenny and came with the video tutorial he did for the Europa-Callisto trip. I tried to do exactly what he did in that video.

Strange to relate, IMFD worked better this time than it has before. I was trying to capture the map issues I mentioned in another thread, but this time the map worked perfectly, which is an all-time first for me under Vista. But this does show the problems I had with the autoburn.

I should have mentioned before, but the autoburn seems to work fine for the Orbit Eject program. I set the MFDs up like in the video, with Course (Target Intercept, Planet Approach, and Orbit Insert) on the right and Orbit Eject and Map on the left. The Orbit Eject autoburn in the left MFD has always works fine. The ship burns all the way through it without stopping to reorient. But when doing burns in the right MFD with various forms of the Course program running, the ship always reorients before the end.

So, things don't go haywire until the 1st mid-course correction burn, which is done in the Course MFD. This one and the plane change burn afterwards both reoriented and I ended up finishing them manually. Same with Planet Approach burns. In both cases, after just a little manual burning, I would have had to have reoriented again before finishing the burn, but I just stopped figuring I was in the ballpark.

The Orbit Insert autoburn did something I've never seen it do before. Every single other time I've run this scenario, this autoburn has not done anything at all, so I've had to do the burn completely manually. The only times I've had Orbit Insert do any autoburn is when heading to Mars, not Callisto.

This time, however, the Orbit Insert autoburn did something, but it wasn't correct. It turned prograde instead of retrograde, and fired the main engines at about 25% power. This had no effect at all on the burn time remaining figure, which remained at 0.104 the entire time. I let this run for 10-15 seconds. I finally I stopped the autoburn and did the orbit insertion manually.

Last, I used the Orbital / Circularize program in the left MFD to tidy the orbit up. This autoburn also resulted in the ship wanting to reorient to finish the burn.

 

[Bullethead]

I've noticed something that might bear on this issue of the ship reorienting towards the end of an autoburn. I think this is a change from earlier versions of IMFD.

In 5.1g, when you do an autoburn, the ship does not hold its pitch and yaw attitude during the burn. I'm talking about the long, uninterrupted part of the burn prior to the last .4 seconds. If you look out the window during this phase of the burn, you'll see your ship's nose is moving around in a circle via pitch and yaw movements relative to the background stars. The roll attitude, however, remains constant. The circle around which the nose moves appears to have a radius about equal to one of the horizontal lines on the nose indicator of the docking HUD, and the speed is about what you'd get if you held an RCS key down for about 1 second.

Despite this constant movement of the nose, however, the + on the BV screen remains white and in the center of the target. It's only when the main engines stop firing at the last .4 seconds that the + on the BV screen moves at all. Then it instantly turns green and jumps to the edge of the screen, after which the ship goes into to its period of many small burns punctuated by major attitude changes. So I'm thinking that during the main part of the burn, trajectory inaccuracies build up due to thrusting as the ship's orientation is changing. Then, when the autoburn program has to calculate how to manage the last .4 seconds, it notices it's gotten off course, and is unsure how to correct the problem.

As noted in a previous post, Orbit Eject autoburns work as intended, going all the way through without interruption and no reorientations toward the end. Thus, I'm thinking the problem might lie with the orientation system for burns in arbitrary directions. I'm thinking the Orbit Eject burn is prograde, and somehow it's easier to hold that. But the other burns are at funky angles derived mathematically, and autoburn thing seems to have difficulty pointing the ship in those directions and keeping it there.

BTW, I'm using the AbsoluteKillRotMFD. But I've always used that, and I never noticed this orientation instability with earlier versions of IMFD.

 

[jarmonik]

I've noticed something that might bear on this issue of the ship reorienting towards the end of an autoburn. I think this is a change from earlier versions of IMFD.

Here is IMFD 5.1e
http://koti.mbnet.fi/jarmonik/51e.zip

If you look out the window during this phase of the burn, you'll see your ship's nose is moving around in a circle via pitch and yaw movements relative to the background stars. The roll attitude, however, remains constant.

This is a failure of an attitude controller. It should be more like a spiral leading into a fixed attitude. This problem should also appear in IMFD51e. It's most likely that this is caused by a higher frame rate of the new computer. There has been a lot of problems with attitude controller and it still doesn't work.

Then it instantly turns green and jumps to the edge of the screen, after which the ship goes into to its period of many small burns punctuated by major attitude changes.

This more like a critical problem but I haven't been able to reproduce it. It sounds like a computational presicion problem. This often happens when you have less than one milli-second of burn remaining. Are you sure not confusing seconds and milli-seconds. How much delta velocity is still remaining ? This might also indicate a hardware problem in the FPU if it can be repeated only with the new computer.

 

[Bullethead]

This more like a critical problem but I haven't been able to reproduce it. It sounds like a computational presicion problem. This often happens when you have less than one milli-second of burn remaining. Are you sure not confusing seconds and milli-seconds. How much delta velocity is still remaining ? This might also indicate a hardware problem in the FPU if it can be repeated only with the new computer.

I'll get back to you on this. It's too late at night now to get into it.

Thanks for answering my posts.

 

[wovik]

I experienced the same problem as Bullethead using IMFD 5.1g regarding the many itterations near the end of a burn. I was using the DGIV modified with mark V engines and the largest fuel and O2 loadouts. I believe it was the Dv figure which was below .4. I just stopped the burn at that point. It was so small it didn't make much difference on my long trip to Pluto.
I also find the orbit insert does not work and had to manually do this. It appears to increase the eccentricity instead of reducing it.
I did not have the nonspherical gravity checked on the options screen. I forgot IMFD wants this but it got me there.
I find IMFD is a great tool even with it's problems.

Jack

 

[jarmonik]

Does this problem appear with IMFD5.1e ?

http://koti.mbnet.fi/jarmonik/51e.zip

 

[wovik]

I don't recall it happening useing 5.1e. I started with 5.1d but 5.1e came out a day later when I first started with Orbiter. It is the one I learned on.

Jack

 

[NukeET]

Does this problem appear with IMFD5.1e ?

http://koti.mbnet.fi/jarmonik/51e.zip

Haven't seen this problem with ver 5.1e.

I just finished development of another AMSO scenario pack using it.

 

[polaris149Tiberius]

May or may not help

May or may not help

I dont know if this helps or not but I noticed that burns seem to get kinda mixed uo (when doing transfers at least) and I am watching the burn from the external view. This was in IMFD v 1.5d I think. Ill check the version when I test it again but it seems to NOT be fixed in version 1.5g.

Does this have ANYTHING to do with this issue or am I reaching for the stars? -if youll pardon the phrase.

Thanks in advance

 

[jarmonik]

I have made a small correction in the IMFD. Does it fix the problem ?

http://koti.mbnet.fi/jarmonik/IMFD51h.zip

 

[jarmonik]

I am still waiting a test report for 5.1h

 

[Zeehond]

I am still waiting a test report for 5.1h

I didn't have the attitude problems described above in the g version and the h version seems to work fine too. Only problem is orbit insert, that doesn't work for me as reported by other people above.

 

[Bullethead]

I am still waiting a test report for 5.1h

Sorry it's taken me so long to get back to you, but I've been busy lately. And I greatly appreciate you taking the time to listen to me .

Anyway, I got 5.1h and tried once again at rdenny's Europa-Callisto scenario using the standard DG as provided in the scenario I attached above. I noticed several problems. When a problem appeared, I paused the game, hit ALT-TAB, and saved the log. Then I returned to the game and kept on with the scenario. Thus, each log attached is cumulative of all that's happened since the start.

I noticed the following problems:

1. Orbit-Eject Burn Reorients
Before 5.1h, autoburns using the Orbit-Eject program were the only ones that didn't reorient at the last 0.4 seconds remaining. In 5.1h, the Orbit-Eject burns now reorient serveral times at this point, just like all the other burns. All autoburns now reorient 2-3 times. The autoburn program finally stops when it recenters the + in the BV bullseye after the 2nd or 3rd reorientation. At this point, there's between 0.03 and 0.3 seconds of burn time remaining.

The final amount of burn time remaining doesn't seem to matter. The ship always reorients 2-3 times and the autoburn program always stops when the + is back in the center.

2. Map Display
On this Vista machine (I still haven't had a chance to check an XP box), in 5.1e my ship's path and the intercept point did not appear in the Map display. In 5.1g, my path worked by not the intercept point. Now in 5.1h, it's back to 5.1e: neither my path nor the intercept point.

There's also a new problem in 5.1h. Now the p-(target) function of the "center map view" doesn't work. The p-(target) function now only goes to the current postion of the target, as if I hadn't used the p- parameter.

3. Reorientations During Correction Burns
All this is the same as in 5.1g. The ship never stabilizes its orientation during the main part of the burn. Its nose still moves around in a circle during this part of the burn. And then, as noted above, when the burn time gets down to 0.4 seconds, it reorients 2-3 times and finally quits when the + recenters.

4. Orbit Insert Doesn't Work
When the ship is in position to do the Orbit Insert autoburn, either nothing happens or the burn lasts forever without the displayed burn time decreasing at all. That last is what's shown in the logs. While this non-stop autoburn was happening, my orbit's inclination got further and further from the value entered in the Planet Approach program, and my orbit just became a tighter hyperbolic without ever becoming even elliptical. The Pe was always below the surface and got more so as the the burn progressed. When it became obvious that this was an incurable problem, I pulled the plug on the scenario.

 

[jarmonik]

1. Orbit-Eject Burn Reorients
Before 5.1h, autoburns using the Orbit-Eject program were the only ones that didn't reorient at the last 0.4 seconds remaining.

There is absolutely nothing changed in 5.1h that shoud effect in orbit-eject.

2. Map Display
Now in 5.1h, it's back to 5.1e: neither my path nor the intercept point.
There's also a new problem in 5.1h. Now the p-(target) function of the "center map view" doesn't work.

There is absolutely no changes made in 5.1h that could effect in map program

3. Reorientations During Correction Burns
All this is the same as in 5.1g. The ship never stabilizes its orientation during the main part of the burn. Its nose still moves around in a circle during this part of the burn.

Circular motion of the nose is caused by a problem in AttitudeControl.DLL that could appear at higher frame rates. Also time acceleration could effect into the problem. I don't remember seeing this causing vessel re-orient problems before. But if the problem lies in the Attitude Control yuo can execute burns manually without Autoburn. If there is still re-orientation problems when doing it manually. Then, it is unlikely a fault of the Attitude Control or the circular motion of the nose.

You could try to increase AB RCS Treshold from 1.0 to 3.0
It is located in Configuration page and IMFD5.cfg

4. Orbit Insert Doesn't Work

Unfortunately it doesn't. This failure should cure it self when I get the other things in order.



Right now I don't know is this:

• a Vista specific problem
• a Computer hardware specific problem. (broken memory, FPU...)
• a Problem in the IMFD

ATTENTION:
Readers of this thread:

Could you reply if the IMFD works or if it dosn't work in Vista !

 

[EtherDragon]

I tried IMFD5.1g just now in Vista64 ultimate - and it worked without any problems. I have not thouroughly tested it, so I couldn't say if every function works. But first impression is yes, IMFD5.1x works in Vista.