Heading tape and HSI not indicating correct course in Caucasus

[TobiasA]

On the caucasus map, the heading tape and HSI are off by about 14-16°.
On the marianas, you fly straight east or west (90 or 270°) and the latitude does not change. In the caucasus, you fly the same course, but the latitude does change and TacView shows that you fly or heading of about 14 degrees more. Radials on the HSI are also affected.
Wind is not so much an issue in the free flight missions in which I tested the behavior.
We noticed when someone in our squad invited us for a navigation training and l got a bit curious as to why I flew about 15 degrees off the desired radial.

You have to fly heading ~75 in the caucasus to go east.

It also does not happen in Syria.

It is however related to the theater, I think, since the error is about the same in the Viggen... Feel free to move it.

Tracks and TacViews included, the TacView in the Viggen was the usual "spawning with heading 0, then turn to 090" and you can see the initial heading is already off. It is just easier to see in the F-16 when you look at the INS page.

heading_tape_caucasus.trk heading_tape_caucasus.acmi heading_tape_marianas.acmi heading_tape_marianas.trk caucasus_test_Viggen.zip.acmi

Edited November 21, 2021 by TobiasA

[TobiasA]

Edit: Well, it even seems to be checkable by using the ruler in the F10 map vs. the cockpit indications. Then, the magnetic compass will show the heading on the map (the one of the ruler). I think (just a wild guess) that the sign of the magnetic deviation is the other way around.

Because when I enter a magnetic deviation of W06.8 instead of E06.8 I'll head straight east when flying 090.

[Machalot]

Last year I had an issue similar to this which was the result of a corrupted MagVar table in my Save Games folder.  I eventually fixed it by deleting the contents of ...\Saved Games\DCS.openbeta\Data\MagVar\TabularData

What clued me in was that all the files in there are the same size, but one of mine was an odd size.

Next time I started DCS it rebuilt those tables and everything was fine.  Maybe worth a try.

 

Edited November 21, 2021 by Machalot
Added link to referenced post

[Frederf]

Lines of constant latitude, longitude aren't the same thing as ruler directions NESW. It's a necessary compromise of the flat geometry of the simulated environment.

[TobiasA]

vor 5 Stunden schrieb Frederf:

Lines of constant latitude, longitude aren't the same thing as ruler directions NESW. It's a necessary compromise of the flat geometry of the simulated environment.

Well yes but not 14 degrees. 

[Frederf]

At the edges of the map difference between ruler and longitude gets about 9 degrees in magnitude. Combined with 6 degrees variation that can be >15.

Edited November 22, 2021 by Frederf

[TobiasA]

vor 39 Minuten schrieb Frederf:

At the edges of the map difference between ruler and longitude gets about 9 degrees in magnitude. Combined with 6 degrees variation that can be >15.

 

It is exactly the same in the middle of the map. The ruler is about 4-5° off in the main area, that is probably due to the conversion flat - round. Also, on the edge of the map, you will always have a slight deviation due to that. The north "opens up" so to say, so lets just fly north in the middle of the map.

The start heading is 000, and it will also show TRK 360 on the "TacView globe" in the middle of the map. As soon as you correct your course to go north, you go about 7° off.
In the middle of the map, I go north if I enter a magnetic deviation of exactly 0.

[TobiasA]

I do agree that the Tacview globe and the flat DCS map do differ, but it is a little too small to explain a 14 degree error. 

[Frederf]

At Vaziani the ruler measures lines of longitude about 351°. At Anapa the ruler measures lines of longitude about 357°.

Realize that when you go ruler 360 you aren't traveling along a constant longitude which you would be in the real world 360 true. DCS calls 360 ruler = 360 true but that's not the case when compared to the longitude lines. Ruler 360 are lines of constant Z coordinate which DCS calls "true north" for the purposes of heading. All of DCS is based on the idea that going along the XZ grid is true course even though the L/L grid isn't aligned with the XZ axes.

It's possible to drag a flight path along the 46°E longitude line thinking that this must be true north by definition. Then the cockpit shows DCS 344 magnetic. 360 - 344 = 16. There's your 15 degree error.

[TobiasA]

vor 52 Minuten schrieb Frederf:

At Vaziani the ruler measures lines of longitude about 351°. At Anapa the ruler measures lines of longitude about 357°.

Realize that when you go ruler 360 you aren't traveling along a constant longitude which you would be in the real world 360 true. DCS calls 360 ruler = 360 true but that's not the case when compared to the longitude lines. Ruler 360 are lines of constant Z coordinate which DCS calls "true north" for the purposes of heading. All of DCS is based on the idea that going along the XZ grid is true course even though the L/L grid isn't aligned with the XZ axes.

It's possible to drag a flight path along the 46°E longitude line thinking that this must be true north by definition. Then the cockpit shows DCS 344 magnetic. 360 - 344 = 16. There's your 15 degree error.

I am talking about the heading tape after the magnetic correction being 15 degrees off. 

[TobiasA]

As an addition: I know the issue of converting spherical map data to a flat map, I was part of a terrain development team on another sim before it had three letters. That is not the issue here, the issue is that I am more accurate without magnetic correction than with it in man cases. 

And even with spherical errors, the north does match in the middle of the map. 

Sorry if I used a wrong approach to it. 

Edited November 22, 2021 by TobiasA

[Frederf]

DCS north and longitude north do coincide but only at LongitudeCenter = 32.96 as defined in \Mods\terrains\Caucasus\Caucasus.MapCoords.lua

Notice the original post is comparing DCS to TacView tracks. TacView looks like its external heading label references DCS ruler direction (aka DCS true) and the pseudocockpit display is an approximation of expected DCS magnetic heading.

"You have to fly heading ~75 in the Caucasus to go east."

If you define going east as not changing latitude then that's ~081 DCS ruler over near Vaziani which would be about a 075 magnetic on the cockpit heading tape. Note that TACAN radials in Caucasus are defined according to magnetic heading so the VAZ TACAN 090 radial is aligned on a ~075 F-16 HUD magnetic tape course/heading.

 

[TobiasA]

vor 7 Stunden schrieb Frederf:

DCS north and longitude north do coincide but only at LongitudeCenter = 32.96 as defined in \Mods\terrains\Caucasus\Caucasus.MapCoords.lua

Notice the original post is comparing DCS to TacView tracks. TacView looks like its external heading label references DCS ruler direction (aka DCS true) and the pseudocockpit display is an approximation of expected DCS magnetic heading.

"You have to fly heading ~75 in the Caucasus to go east."

If you define going east as not changing latitude then that's ~081 DCS ruler over near Vaziani which would be about a 075 magnetic on the cockpit heading tape. Note that TACAN radials in Caucasus are defined according to magnetic heading so the VAZ TACAN 090 radial is aligned on a ~075 F-16 HUD magnetic tape course/heading.

 

Yes, but the heading tape should only be off by the amount of the "map stretch error" and not by the magnetic deviation. 

Which means it should display 082 or something like that. 

Because the heading tape (and the heading indication of other planes and the HSI as well) is corrected for magnetic deviation via the MAGV page (or other input devices in other planes). 

[TobiasA]

I agree that my assumptions about a sign error just canceled out the map stretch error, it looks like the compass is showing true north instead of the magnetic north at least on parts of the map. 

If I set it to 0 near Vaziano I am going true NSEW, with a rotation of the map stretch error.