Looking for a real Mosquito compass deviation card, anyone?

[Lau]

On 12/4/2023 at 5:11 PM, Lau said:

@JBDCS

Most compass deviations have a mix of negative and positive values accross the sinusoid, interesting to see that the Manchester compass deviation only uses positive values. 

 

 

 

It seems that I have indeed read the Manchester compass swing card too fast and my above statement is wrong. There are indeed negative values in column VI

[JBDCS]

Okay, first off sorry I've not posted back for a while, I have continued testing things out though.

The testing is very complicated (or maybe I'm overcomplicating it) because:

• it's very difficult to read the compass(es) accurately in the cockpit
• the readings are affected by pitch and bank as I've already mentioned
• the readings are also affected by inertia effects making the needle move around

I've now worked out how to get the compass readings from the devices in game and log them to file along with the other heading information, so that fixes the first problem, but the other two still make things difficult.  Anyway, by doing that and graphing out the differences between a setup with zero deviations and one with deviations set I think I'm reasonably happy that it is working how I expect (I'll put more details of that in another post sometime).

So now that I've got that far, I've started work on a program for playing with the setups, and generated some deviation settings to simulate the readings for the Manchester in the 1944 RAF manual (p190).  It's still very much work in progress, but the screenshots below should show what I'm aiming for, and at the end I've listed the deviation settings you could use in your scripts for these deviations (though I've not actually tested those settings yet).

The first screenshot shows the deviations from the standard Mosquito compass_device.lua.  These are the generic deviations which seem to be used across all modules in game.  It shows the deviations against each course in steps of 10 degrees, along with a graph in the style of the RAF compass log graph on p191.

The second screenshot shows the deviations for the Manchester, entered manually against the compass points, and adjusted for coefficient A (ie. the deviation figures match those in column vi in the manual), alongside the graph for those deviations.  The only difference is that the table in the manual has north west first but my table has north first so that the rows line up roughly with the graph.

The third screenshot shows the same data as the second, this time showing the deviations script which can be used in the device .lua files to set these deviations, and on the right a compass deviation card in the style of the card described also on p191.

[Note: I realised after taking this screenshot that the deviation figures in the lua script are wrong - they should all be divided by the K_deviation value, since they're multiplied by it after that).  I've fixed it now in the script at the end of this post, but this screenshot still shows the incorrect values.  Also found that lua does not accept numbers beginning with +, so have removed those in the script too but again on this screenshot shows the old version.]

(I haven't worked out yet how to suitably position the ADD and SUBTRACT labels, so for now the card just shows A or S.)

The idea is that images of these graphs and/or cards can be added to the kneeboard or somewhere else.  I'll also add the USAAF style deviation card from that manual (which is just a simple list) and maybe other formats later.

 

Anyway, if you want to try playing with these deviations, here's the code you need to copy and paste into your compass_device.lua and/or remote_compass_device.lua files, replacing the similar existing code (but see the warning below this code):

deviation = {
	  0.0, -1.34,
	 10.0, -0.56,
	 20.0, 0.47,
	 30.0, 1.52,
	 40.0, 2.35,
	 50.0, 2.75,
	 60.0, 2.77,
	 70.0, 2.53,
	 80.0, 2.09,
	 90.0, 1.53,
	100.0, 0.77,
	110.0, -0.34,
	120.0, -1.52,
	130.0, -2.43,
	140.0, -2.79,
	150.0, -2.73,
	160.0, -2.39,
	170.0, -1.87,
	180.0, -1.28,
	190.0, -0.44,
	200.0, 0.62,
	210.0, 1.67,
	220.0, 2.46,
	230.0, 2.79,
	240.0, 2.77,
	250.0, 2.51,
	260.0, 2.07,
	270.0, 1.52,
	280.0, 0.8,
	290.0, -0.26,
	300.0, -1.39,
	310.0, -2.31,
	320.0, -2.75,
	330.0, -2.74,
	340.0, -2.42,
	350.0, -1.92,
	360.0, -1.34,
}
K_deviation = 0.74627
i = 0
for k, v in pairs(deviation) do
  i = i+1
  if math.fmod(i + 1.0, 2.0) > 0.0 then
    deviation[i] = deviation[i] * K_deviation
  end
  -- to radians
  deviation[i] = math.rad(deviation[i])
end

WARNING - If you edit the compass_device.lua files, they will fail the integrity check (you get messages in dcs.log about IC fail for those filename), so be sure to make a backup copy before editing them so that you can copy that back if you want to run on multiplayer servers which enforce the IC.  I think this also means that your modified files will be moved to a backup folder when the game is updated and the files would revert to the standard ones, so you'd need to copy your modified version back into the game directory after updates.

Edited December 17, 2023 by JBDCS
Fixed errors in the deviations script

[Bozon]

Normally I am happy if I can hold course within +-5 degrees…

 

[JBDCS]

Yes I agree, nice to get things right and great that the game/sim takes these things into consideration, but can any of us really fly that precise a heading?

[stuart666]

Im getting better. But no.

As a slight aside, it would be helpful if we could have a top down compass view, just so we can set things up properly. Ive found Ive been finding magnetic heading off the map view, even in vr using the compass is near impossible.

I might say, thanks for looking into this JBDCS. Im not sure Ill ever use it as a mod, but if someone is identifying it as potential perhaps the devs might consider integrating it.We live in hopes.

 

[Bozon]

6 hours ago, stuart666 said:

Im getting better. But no.

As a slight aside, it would be helpful if we could have a top down compass view, just so we can set things up properly. Ive found Ive been finding magnetic heading off the map view, even in vr using the compass is near impossible.

I might say, thanks for looking into this JBDCS. Im not sure Ill ever use it as a mod, but if someone is identifying it as potential perhaps the devs might consider integrating it.We live in hopes.

 

Don’t use the P8 compass, it is intended as a backup. The RI is the main magnetic compass. It too is a bit tricky because the needle is hight above the plate so perspective changes the readings - either move your head to look straight at it (still easier than the P8) or use the rotating arrow, and bring the needle parallel to it (correct from any viewing perspective).

[stuart666]

Yeah you are undoubtedly right. I keep forgetting about it for some reason...

[JBDCS]

Yes the R.I. Compass Repeater is certainly the easier instrument to read but one important note is that that one does read as if the compass it repeats from is fixed in the airframe and not gimballed at all, so with that one it's even more important to make sure you're flying dead level when reading the compass because without a gimbal the reading will be affected a lot by pitch and bank.

For example in my test results detailed earlier in this thread for the P8M at various headings on the ground, with the aircraft pointed east or west there is a couple of degrees error due to the pitch up being more than what the compass can handle.  For the repeater compass, this same test gives an error of more than 20 degrees in heading at 11 degrees pitch, which is what would be expected if it's fixed in the airframe.

I've tried adding a max_pitch_bank line to that compass setup file, but it seems to have no effect, so looks like it's not that simple to set up.

This is a bit strange when the DCS flight manual says during engine warmup to check that the P8M and repeater compasses agree and to set the gyrocompass from them, since on the ground as it is neither of them will usually read correctly and there will usually be a difference between their readings, often quite large.

There's no information in the manual about the compass device which the repeater reads from, so no idea whether this is realistic or not, but someone will perhaps know more.

[Lau]

@JBDCS

On 12/17/2023 at 10:01 AM, JBDCS said:

Okay, first off sorry I've not posted back for a while, I have continued testing things out though.

The testing is very complicated (or maybe I'm overcomplicating it) because:

No needs for apologies, I am amazed by the work you have put out to use the little room there is to MOD DCS, so we can enjoy a working compass deviation. What I also appreciate is that your work will allow players to customize compass deviation to their personal preferences. Far reach out JBDC, you are really used to go above and beyond in everything that tickles your curiosity and I  can confirm that you have indeed a natural talent in overcomplicating things, which is something very rewarding to witness. So thank you indeed for your passion and support Sir.

The Manchester compass deviation sinusoid seems indeed more natural than the default Mosquito one, I am not surprised that ED did not get to that level of realism and at the same time, greatfull that the engine allow us to push that little realism boundary a notch further.

I can read the reactions to what we are trying to accomplish here with the question: Are we good enough virtual pilots to make the most of that level of precision? My answer is yes, that is within the limitations of DCS. It all depends on what definitions of FUN we have, for me improving my virtual flying skills is very rewarding, I design my little missions with that consideration in mind. If I get lost or do not arrive on time to support the ground troops then I lose the game, which makes me strive to get it right the next time. So to succeed in the mission, I indeed battle not only with guns and bombs, but also with precision flying: time, speed, altitude, fuel management, IMC conditions, compass deviation, mag variation, wind drift, temperature variables, dead reckoning navigation, etc…. It is the combination of factors that makes the experience  ultimately interesting and therefore FUN to fly from my perspective.

Time permitting, I will test the Manchester compass deviation sinusoid and report back. Ultimately I will however move to the figures I have submitted, not only to compensate for the little size of DCS maps like Normandy 2.0 but, simply to make the exercise more interesting and less forgiving. 

I can't stop laughing when I get too happy because all my numbers add properly during a mission and that suddenly, I get jumped over by enemy fighters, a real threat these challenges and luckily from the comfort of home.

I usually use the repeater for quick reference during BFM/CAS, the DG for precision navigation and easy of read, making sure it is aligned on a regular basis with the P8.

Remove all the helpers and get serious about flying warbirds in DCS ; )

Best,

Lau

 

 

 

 

[stuart666]

2 hours ago, JBDCS said:

Yes the R.I. Compass Repeater is certainly the easier instrument to read but one important note is that that one does read as if the compass it repeats from is fixed in the airframe and not gimballed at all, so with that one it's even more important to make sure you're flying dead level when reading the compass because without a gimbal the reading will be affected a lot by pitch and bank.

For example in my test results detailed earlier in this thread for the P8M at various headings on the ground, with the aircraft pointed east or west there is a couple of degrees error due to the pitch up being more than what the compass can handle.  For the repeater compass, this same test gives an error of more than 20 degrees in heading at 11 degrees pitch, which is what would be expected if it's fixed in the airframe.

I've tried adding a max_pitch_bank line to that compass setup file, but it seems to have no effect, so looks like it's not that simple to set up.

This is a bit strange when the DCS flight manual says during engine warmup to check that the P8M and repeater compasses agree and to set the gyrocompass from them, since on the ground as it is neither of them will usually read correctly and there will usually be a difference between their readings, often quite large.

There's no information in the manual about the compass device which the repeater reads from, so no idea whether this is realistic or not, but someone will perhaps know more.

Was it like that in reality? Im surprised that, having the extra room, they didnt take that opportunity. But then I suppose if they had, it would make the gyro compass somewhat irrelevant.

[JBDCS]

My guess would be no, not like that in reality, but since I don't really have any knowledge of it it's just a guess.

[kablamoman]

13 hours ago, JBDCS said:

Yes the R.I. Compass Repeater is certainly the easier instrument to read but one important note is that that one does read as if the compass it repeats from is fixed in the airframe and not gimballed at all, so with that one it's even more important to make sure you're flying dead level when reading the compass because without a gimbal the reading will be affected a lot by pitch and bank.

For example in my test results detailed earlier in this thread for the P8M at various headings on the ground, with the aircraft pointed east or west there is a couple of degrees error due to the pitch up being more than what the compass can handle.  For the repeater compass, this same test gives an error of more than 20 degrees in heading at 11 degrees pitch, which is what would be expected if it's fixed in the airframe.

I've tried adding a max_pitch_bank line to that compass setup file, but it seems to have no effect, so looks like it's not that simple to set up.

This is a bit strange when the DCS flight manual says during engine warmup to check that the P8M and repeater compasses agree and to set the gyrocompass from them, since on the ground as it is neither of them will usually read correctly and there will usually be a difference between their readings, often quite large.

There's no information in the manual about the compass device which the repeater reads from, so no idea whether this is realistic or not, but someone will perhaps know more.

FYI, it seems to be a bug with all remote compass installations for the WW2 modules.

Although I'm not sure which model of compass is used in the Mosquito I am reasonably certain it shouldn't affect the p51 or the axis birds in the same way you detailed (but it does at the moment).

Some additional info from this thread where I did a bit of research on the remote indicating compass in the mustang: https://forum.dcs.world/topic/264224-cannot-calibrate-gyro-with-mag-compass/

Edited December 23, 2023 by kablamoman

[Bozon]

I can’t find the thread where this was discussed, but what my fading memory tells me is that the Mosquito used the same remote indicating compass as the RAF bombers, whose sensing part was a pretty bulky gimbaled unit. I remember something about the mosquito ground sitting angle being over the gimbal limit by a hair.

Now this is pure conjecture - since it was made for big fat bombers, the gimbal limits may have been quite tight, which is fine for the bombers. Maybe we could ask the ground crew to lift the tail by a few inches so we could get RI reading on the ground? 

[JBDCS]

10 hours ago, kablamoman said:

Some additional info from this thread where I did a bit of research on the remote indicating compass in the mustang: https://forum.dcs.world/topic/264224-cannot-calibrate-gyro-with-mag-compass/

Thanks for the thread link kablamoman, I hadn't seen that thread but as you say it seems to be describing exactly the same problem in the Mustang (just from a quick look so far -  I'll give it a proper read sometime soon).

[Lau]

@JBDCS

Hi,

As promised, I did have a look and tested the Manchester compass deviation figures that you provided, just to realize that with such little deviations, there is not much to be noticed from a pilot perspective in game. I would gladly use your spreadsheet work to implement higher deviations figures like the ones I submitted to you on the 4th of December. 

Best,

Lau