Looks like we broke the T-72B3's Fire Control System, in some specific (and possibly rare) situations [FIXED?]

[Ssnake]

We'll be fixing it, of course, but bug #11178 will probably spoil the experience for you. I'm sorry, but it seems to have happened at the last minute.

[El_Chacho]

Thanks Ssnake,

What is bug #11178?

Cheers,

 

[Ssnake]

Under some circumstances (but apparently not all the time) the activation of the automated target tracking can result in erratic turret movement. It may not happen to you at all.

[El_Chacho]

Thanks! I'm thoroughly enjoying the new toy.

 

 

[Mirzayev]

Ah, so that's not expected behavior. 

[Gibsonm]

7 hours ago, Mirzayev said:

Ah, so that's not expected behavior. 

 

Russian build quality?

[Volcano]

This thread may not have been entirely warranted, because it seems to happen only in specific situations - so no need for there to be any mass hysteria here.  😉

 

But now that it has been mentioned (vaguely), I will follow up and provide an internal detailed description of the bug (see below), to limit "false positives" of observations of legitimate behavior (the T-72B3 has plenty of real life quirks represented). 

So, to be clear, here is a detailed description of this behavior to look for:

 

"

(In certain, specific situations) when a target is lased, the sight jumps far below the target (1 to 2 target heights) with auto tracking on (and possibly off too). Then when the user tries to bring the sight back onto target by elevating), he cannot. The movements of the gun/sight then becomes exaggerated, like 2 to 3 times the normal speed - and any left/right movement of the input brings about a seemingly opposite movement of the turret, as if it is fighting against the input (like when you have lead being calculated on [a single axis type] stabilized sight and you traverse rapidly from side to side). The AI also cannot aim the gun when this starts to happen; if you jump out  to F8 view (to let the AI shoot the target), you can instead see the gun swinging back and forth left and right, and the AI doesn't shoot. 

 

Also, if you are in the gunner position when this happens, the AI commander seemingly becomes completely inactive - doesn't  put the gun onto target, etc.

 

After going through all this, the user will eventually press P or O key to disable/reset the FCS and auto-tracking. When this happens, it goes away, until the user lases the target again.  So the only way to engage the target when this occurs is through the GAS (the old T-72 sight).

"

 

 If anyone encounters this described behavior then please post a test scenario here.  We have discovered this issue post 4.3 release and are having trouble reliably reproducing it... 😔

 

(The best way to get a test scenario might be to pause the game when you see it, then save in progress, and upload that here with a note on what vehicle on what side its currently being observed).

 

I should stress that there are many unknowns here, whether auto-tracking is actually a cause or not, for one thing, or whether it is a Network Session only bug or not (probably not). 

 

Normally we don't put out unknown information like this, but here we are. So, as I said, the main thing that would help at this point would be a test scenario where someone is able to reproduce the behavior (or possibly a save in progress of a scenario where someone sees it occurring right then, with a mention of what vehicle it is happening on).

 

 

[Volcano]

Looks like we may have fixed this already for the next update. Or at least something very similar. 😅

 

Still, if you observe this and you are able to provide a test scenario (or can do a save in progress scenario of it) then please post here anyway, as it might help us make sure its the same thing.

[Badger]

Is it working as expected in regards to the difference in slew/track sensitivity compared to day sight?..day sight tracking /slewing is much smoother and controlled where as the TIS is a lot more sensitive. Its hard to manually track in TIS and often causes over compensating as well as getting more pronounced in higher magnifications. Shouldn't they have the same slew rate/sensitivity of inputs ?

Edited August 10, 2022 by Badger

[Volcano]

Yes that works right as far as I can tell - the reason there is that the TIS has different and higher magnifications than the day sight, so it can appear to move a lot faster since your FOV is narrower. This is not related to the issue described above.

[Badger]

I tried to engage a target manually in max zoom using TIS mode ...it was damn near impossible. kept loosing the target due to input over compensation. It's too erratic IMO it needs to be toned down.

Edited August 10, 2022 by Badger

[Volcano]

You may be observing what was described above. Hard to say. You would have to observe it after the change to know.

 

But let's say that if you aren't observing the bug above (because the sight would be moving all over the place erratically no matter what FOV you are in); in general - just because you are zoomed in to maximum magnification doesn't mean that the turret and gun of the T-72B3 should be less responsive. The T-72B3's FCS doesn't know any better (as far as we currently know), so what you will get then is like looking through a telescope trying to track a moving object - any input at that level of magnification is going to feel exaggerated. The highest zoom magnification on the T-72B3 is really intended to be used on long range stationary targets, or in conjunction with the auto-tracking feature (that's the main reason the latter function exists IIRC). 

 

But I am also not saying there isn't another issue there either. Just a general description that the higher the magnification, the harder its going to be to move the sight around, and this this might be getting confused with the bug behavior described above.

[Higgs]

On 8/11/2022 at 12:34 AM, Volcano said:

You may be observing what was described above. Hard to say. You would have to observe it after the change to know.

 

But let's say that if you aren't observing the bug above (because the sight would be moving all over the place erratically no matter what FOV you are in); in general - just because you are zoomed in to maximum magnification doesn't mean that the turret and gun of the T-72B3 should be less responsive. The T-72B3's FCS doesn't know any better (as far as we currently know), so what you will get then is like looking through a telescope trying to track a moving object - any input at that level of magnification is going to feel exaggerated. The highest zoom magnification on the T-72B3 is really intended to be used on long range stationary targets, or in conjunction with the auto-tracking feature (that's the main reason the latter function exists IIRC). 

 

But I am also not saying there isn't another issue there either. Just a general description that the higher the magnification, the harder its going to be to move the sight around, and this this might be getting confused with the bug behavior described above.

While I understand what you're saying, I've noticed something similar with other vehicles in the past. For example on the Leopard 2. The gunner (NFV) is far less responsive then the TC controlling the gun. (Betriebsart: KW). At least with my joystick, the turret is significantly more sensitiv to slight adjustments than it is in the gunners seat, making small adjustments nearly impossible.
The same thing is true for the T-72B3 thermal sight. The turret is far more sensitive when using the thermal sight, than it is using the daysight. Regardless of magnification.

 

The same is true when using the mouse. On the gunners seat, there's a little area in the middle, in which keeping the mouse will result in no turret movement.

Using the TC this area is much smaller. I guess for some reason this translates into the sensitivity of the joystick.

[Volcano]

The problem with these general observations that are thrown out is that it requires that someone look at it and test, which costs time.  But at the current moment, this all just appears to be opinion related and not in any way related to the original topic of the thread.  So either someone should open a new discussion about this perceived difference in input at different magnifications levels, or not.

 

But regarding to what was described with different sensitivity of input at high and low magnification, sorry, I just don't see it. The easiest way to prove this is by looking through the GPS with a constant input applied to the controller (joystick or mouse) while looking at something in the terrain as reference, then toggling between FOV.  If the inputs were more or less sensitive at different magnifications and FOV, then the traverse movement would not be at an identical rate between the FOV level when magnification is changed with constant controller input. 

 

I tested this with both joystick and using a mouse (probably the best test) with mouse vector turned on.

 

But like I said, if this is some genuine concern then it needs to be discussed in a separate thread, but at the moment it was put forth as some kind of "by the way" comment, that we did not dismiss -- it was tested, and the observation and explanation was provided. 

 

Anything beyond that relating now to TC and gunner inputs, and different FOVs, and now we are in a whole different discussion. Because of the fact that now we are on an entirely different subject to what this threat was originally about, I will close the thread to avoid confusion.

 

 

[Volcano]

Good news (hopefully):

 

So we dug down further into the comments with the T-72B3 possibly being more sensitive to joystick/mouse input while looking at the thermal sight view and...

 

It seems that there may have been something going on which caused some additional oddities that might have contributed to this behavior. At least on the surface this should not be true, but a fixes to the T-72B3 FCS seems like it may have also taken care of this as well. Kudos to those who spotted it (Badger and others).  Take a look at it in the patch and see if things have improved. 👍

[Volcano]

Just FYI,

 

So it seems there is still some strange behavior here in Network Session. For the past weeks we have been looking into this, and have fixed (hopefully) all remaining issues recently, for the next patch (post 4.363).

 

In the mean time, it was expressed that a way to work around this problem is ...

 

In Network Session, if you observe the T-72B3 FCS doing some "crazy" (like acting jittery, or as if there is some kind of conflict going on) then in the gunners position change the FCS into "emergency" [. key] and then back to "normal" [, key] and then the issue *should* resolve itself. If you jump out of the tank and back in, you might have to do it again.

 

It seems that the HOST and CLIENT were getting out of sync with an FCS setting, and doing this toggle between normal/emergency forces a message to be sent across the network.