He 111's can't climb together

[NETSCAPE]

KG 55 Testing.rar Use autopilot and 8x time compression.

 

A group of 3 He 111's takes off on an airfield - course is directly east. They are given a distance of 40km to form up, then the climb to 5000m starts.

 

The first obvious issue is the planes all veer off to the right. Why? There is no wind. Watching the wingman form on lead at this point is comical, but I gave them 40km to do it...and they barely have time to form up over that distance due to the derpy nature of the AI zig-zagging and jumping in altitude for no apparent reason.

 

The second issue is the big one. The AI wingman cannot climb with the flight lead. Anyways, sit back and enjoy the show. The wingman will continually fall behind their flight lead. Before I allowed that aforementioned 40km of non-climbing distance to form up together - the wingman were 5km, yes 5km behind flight lead by the time the climb was completed at 5000m altitude. My last test resulted in them falling behind 3km or so, then circling around all the sudden, before the waypoint was reached... then flight lead turned around to fly towards them and I went ahead and alt+f4'd right outta there. 

 

Let me know if there is something I'm missing here...

[JimTM]

I cannot open the .rar file using 7-zip. Can you post it as a .zip file?

[NETSCAPE]

1 minute ago, JimTM said:

I cannot open the .rar file using 7-zip. Can you post it as a .zip file?

 

This has happened a few times with people downloading my stuff. It just means you need to update your program.

 

The mission is a sample. Take off, V formation, 2 waypoints, one to give time to form up, the other to climb.

[JimTM]

Try this mission:

JimTMTestNETSCAPEHe111ClimbV2.zip

 

The planes still veer a little to the right, but after 10km or so they come back on course. After 20km, the wingman were within 1km of me. After 30 km, they were within 0.3 km of me. They were well behaved the whole time, including at the transition to the climb to 5k.

 

Here are the changes I made:

- Added 10s timer before takeoff command (any small value would also do).

- Changed the two event messages to a single one, going to a 0-second timer, which then triggers the first waypoint and a 2-second timer. The 2-second timer then triggers the formation command. I think this setup is key to solving the problem.

- Changed the waypoint diameters to 200m and the speed of waypoint 1 to 210, which is what I climb out at. Not sure if this makes any difference to the behavior though.

- Added some icon translators so I could track the course closely.

 

You might try going with just one waypoint at 5k and see if the wingman can close up properly in a reasonable distance.

 

Hope that helps you out.

 

Cheers!

 

Jim

 

Edited June 8, 2018 by JimTM

[NETSCAPE]

6 minutes ago, JimTM said:

Try this mission:

JimTMTestNETSCAPEHe111ClimbV2.zip

 

The planes still veer a little to the right, but after 10km or so they come back on course. After 20km, the wingman were within 1km of me. After 30 km, they were within 0.3 km of me. They were well behaved the whole time, including at the transition to the climb to 5k.

 

Here are the changes I made:

- Added 10s timer before takeoff command (any small value would also do).

- Changed the two event messages to a single one, going to a 0-second timer, which then triggers the first waypoint and a 2-second timer. The 2-second timer then triggers the formation command. I think this setup is key to solving the problem.

- Changed the waypoint diameters to 200m and the speed of waypoint 1 to 220, which is what I climb out at. Not sure if this makes any difference to the behavior though.

- Added some icon translators so I could track the course closely.

 

You might try going with just one waypoint at 5k and see if the wingman can close up properly in a reasonable distance.

 

Hope that helps you out.

 

Cheers!

 

Jim

 

 

Jim, the first thing I tried was one way point, it was the worst result by far - they fell behind even faster. Even if you alleviate the odd turn to the right, the climb issue is still present is it not? Basically I wanted a couple guys to look it over before making another bug report. Keep in mind this isn't me building a mission, just testing. I've been in the ME enough to know I need a spreadsheet of AI flight speeds/payload/alt/climb rates ect to save time before embarking on a large project.

Edited June 8, 2018 by NETSCAPE

[JimTM]

2 minutes ago, NETSCAPE said:

 

Jim, the first thing I tried was one way point, it was the worst result by far - they fell behind even faster. Even if you alleviate the odd turn to the right, the climb issue is still present is it not? Basically I wanted a couple guys to look it over before making another bug report.

 

No, By the time I reached waypoint 1, the wingmen were in tight formation. They stayed that way through a few minutes of climbing to 5k. I spawned out after a few minutes.

Edited June 8, 2018 by JimTM

[NETSCAPE]

You have to fly over 100km before reaching the proper altitude. Sounds like you didn't see them climb enough to actually observe the issue. Even if they get formed up tight enough, which even I had them doing - the climb issue will still be there unless your speed change has an affect. I'll try a few speeds right now. Not sure what your rig is but my cpu can handle this map and 3 he 111s at 8x time compression easy. 

 

EDIT: I just watched mine then yours back to back. As I suspected, your changes had no affect on the odd AI swerve to the right. The speed change of the climb waypoint is irrelevant since that's hardcoded stuff. Climb is climb, at least in bomber and attack planes - we don't get to decide the speed. 

 

Thanks for trying it out though... I think it's another one for the AI/Autopilot Bug report...

 

Reducing fuel to 50% equated to the wingmen "only" falling behind 1.4km by the time 5k alt was reached. It alleviated the swerve to the right quite a bit since they all take off faster, by faster I mean when the game registers they are airborne. 

Edited June 8, 2018 by NETSCAPE

[JimTM]

No problem.

[Yogiflight]

From my experiences in the old campaign system, where you had to fly as leader, I fully agree with NETSCAPE. This same issue apeared to me when flying Ju87 and He111. Even after bringing my flight into formation after takeoff, they usually lost contact somewhen during the climb, no matter how fast I climbed. I noticed a few times, when I looked back at the right time, that they simply turned to a side for about 90° and then turned back. It was worse with more bombload and steeper climb angle, so I never took the heaviest bombloads and climbed with about 3m/s. Sometimes it was even so bad, that, after I reached my flight altitude, I could not see my wingmen anymore, as they were flying so far behind me. With the flightleader being AI, the climbspeed and -angle of the aircrafts are, as NETSCAPE mentioned, hardcoded, no matter, what distance between the waypoints, the AI wil climb the same way, and if the waypoints are very far spread, fly horizontally to the next one. This is my experience from flying PWCG, were the FW190 starts to climb from a speed of 300km/h in a steep angle to climb with 270km/h, and then flys to the next waypoint with 300km/h horizontally. But not only during climb the aircrafts turn away, this happens during cruising as well.

[JimTM]

I tried running my version of the mission all the way to waypoint 2. Here's what I found:

 

- As before, after the initial 2 or 3 km diversion to the right, the wingman got back on course, closed up, and kept tight formation to waypoint 1.

 

- After waypoint 1, the wingmen gradually fell back to about 3km back, near waypoint 2. The left hand wingman (looking forward) fell back faster than the right hand one. They flew smoothly, with no sudden direction or altitude changes.

 

- At about 5km from waypoint 2, the wingmen suddenly collided.

Edited June 8, 2018 by JimTM

[NETSCAPE]

3 hours ago, JimTM said:

- At about 5km from waypoint 2, the wingmen suddenly collided.

 

dear god...

[NETSCAPE]

further reducing the fuel load to 1000L helps tremendously. When the altitude got reached the wingman were around 1-1.25km behind. Due to the decreased weight they reformed within 15km distance (flight lead was cruising at 280kph). I haven't even got to the point of trying my new formation logic (series of WPs with different speeds dictated by proximity values of flight leaders) but I can assume by these tests today and yesterday that I will need to have fuel loads of less than 2000L, probably even 1000L.

[Yogiflight]

Today I was flying two missions in my He111 career. In the first mission two of my wingmen crashed, I guess when they cut my wide turn about a minute after takeoff, only one of the three remaining wingmen was flying with me, the other two fell back and were not rendered anymore. In the second mission all four wingmen fell back out of render distance, however I climbed both times with less than max duration power ( down to 2000 RPM and 1.0 ATA). It might be, that I was climbing too fast for the obviously hardcoded climbspeed, as I was climbing in a shallow angle with about 230- 250km/h. In the first mission I used 2000l, in the second 1500l of fuel. It doesn't make sense to fly career missions with this AI. I hope you can get it working better, NETSCAPE.