adjusted for later update F-16 Rolling Left during High G turns and Normal Loops

[PanicMan15]

DCS Version: Standalone v2.9.4.53627
NO MODS
CLEAN F-16

Hello all. I have a new bug report that is a pretty big one that has gone unnoticed for quite some time but has become a big issue with the latest update.

After dealing with a slight rolling left issue for the longest time within the aerobatic community, the new DCS update came out and was excited to try it. After doing some initial flying it was found that the F-16 has a TERRIBLE rolling left tendency when performing loops or even higher G level turns. It has made formation flying or even general BFM near impossible without putting almost half to full right stick input to correct when flying.

So let's lay down some basics that have been tried to troubleshoot this issue.
1. All of this was performed with 3 different sticks to try to rule out any sort of sensor issues.
2. There was 4 different machines tested on to see if it was machine/software specific. All machines had 0 mods and have tried repair/reinstall
3. DCS curves/settings were looked at and configured with no changes to default settings.
4. Wind was turned off for all testing.

With that cleared up, let's go over the parameters of the maneuvers performed in the test to help replicate the issue.
LOOP PARAMETERS:
- 410-430 kts calibrated entry speed
- 87-89% RPM
- 4G pull into the loop
- Reduce the pull to "float" over the top at 0.8 to 1.0G
- Back in with the pull to bring back to starting altitude

MAX G TURN:
- Level at 430-450 kts calibrated entry speed
- Roll and Pull to max deflection while engaging afterburner
- When speed starts decaying, release pull to hold 400kts

With the parameters set. We can now go over what is actually happening with the aircraft. During the loop, the rolling left starts almost immediately but is very subtle. As the aircraft approaches around 70 degrees nose high and around 320 kts, the aircraft makes a hard left roll. I started the loop heading 180 magnetic and during this loop completely upside down I should see around heading 0 magnetic. Instead I saw a deviation of 23 degrees magnetic due to the left roll. Pulling through the back side my aircraft has rolled so much I am at 45 degrees AOB on the backside of the loop. This is unacceptable for flying formation in loops.

Now onto the max g turn. This one is less aggressive but the left roll is still there. After a 360 degree turn, I ended up 15 degrees nose high from complete level and gained 1000 feet in altitude.

Please note all of these test were done with no correction with any sort of right roll input. When right roll input was introduced. During the highest rate of roll, near full right roll command was needed.
 

CONCLUSION

After all testing was concluded it was time to look into the cause of the problem. Since this has been an issue since the F-16 first came out, I can assume it isn't related to any new updates that have been done previous except the latest update with the yaw rate boost. Here's what we know it isn't. Gun Weight. If it WAS the gun weight we would be seeing it at all regimes of flight since the weight is a constant and doesn't shift. There has been previous talk that the weight of the gun is the cause of this and nothing can be done but it has gotten worse since the last update. Here's what it could be. LEF schedule incorrect. If the LEF was deploying asymmetrically it could explain the exponential increase in drag on the left wing causing the bigger roll rates. It could also be the rudder trying to compensate for something. This would make more sense since the issue got worse with a rudder change in the latest update. Either way, I don't have access to the code to look at this so it is up to you ED, to find this issue. I have given all of the information I can on this. In future, I urge you ED to PLEASE do further testing with flight models before you hand them out. I understand you crowd source us to help bug report but this is something that is un-flyable and most of the aerobatic community can't update because of it.

Thanks, Panic.

Track has been attached for debugging.

F-16 Rolling bug.trk

[Raven (Elysian Angel)]

Which control stick do you use? The real F-16 limits available roll when you pull back on the side stick depending on AoA. In the Realsimulator software you can emulate that.

15 hours ago, PanicMan15 said:

Since this has been an issue since the F-16 first came out

I never noticed it, but then again I use a FSSB R3L.

[PanicMan15]

3 hours ago, Raven (Elysian Angel) said:

Which control stick do you use? The real F-16 limits available roll when you pull back on the side stick depending on AoA. In the Realsimulator software you can emulate that.

I never noticed it, but then again I use a FSSB R3L.

Three sticks were tested. T16000, Virpil WarBRD, Winwing MFSSB. All of them had the same result. I urge you to re-read the post cause roll rate is NOT the issue. There is a rolling left issue that happens when no input is commanded left. Also would recommend flying those parameters I posted and seeing if you can replicate it yourself. The track is there for reference use as well.

Edited April 14 by PanicMan15

[3shyammenon]

I hope this gets looked into soon! Wasn't nearly as serious of a bug as is it now.

[oldcrusty]

Sorry for sticking my Hornet's nose in here but it reminds me of my attempts to fly a Hornet with an improperly positioned side stick. Even when correctly angled I had to focus on pulling the stick perfectly straight and w/o angling it also left.

I know, I know you guys have all the axes displayed on the screen so there's probably something else going on...  it just popped in my head  

[Raven (Elysian Angel)]

54 minutes ago, PanicMan15 said:

There is a rolling left issue that happens when no input is commanded left.

Not on my end there isn't. I played your track and see what you mean, but when I restart it and take control immediately to recreate your conditions, everything works as it should.

1 hour ago, PanicMan15 said:

roll rate is NOT the issue

I wasn't talking about roll rate. I was talking about roll being inhibited altogether at higher AOA.

[PanicMan15]

1 hour ago, Raven (Elysian Angel) said:

Not on my end there isn't. I played your track and see what you mean, but when I restart it and take control immediately to recreate your conditions, everything works as it should.

I wasn't talking about roll rate. I was talking about roll being inhibited altogether at higher AOA.

Were you correcting the roll with right roll? I'm curious to see your test because this was tried multiple times over different test with other people that saw the same thing.

1 hour ago, oldcrusty said:

Sorry for sticking my Hornet's nose in here but it reminds me of my attempts to fly a Hornet with an improperly positioned side stick. Even when correctly angled I had to focus on pulling the stick perfectly straight and w/o angling it also left.

I know, I know you guys have all the axes displayed on the screen so there's probably something else going on...  it just popped in my head  

That was made sure to be taken into account. All of the test had the stick indicator on and was made sure it was straight. Feel free to download my track and watch with the indicator enabled.

I have attached a tacview track with some testing done later with roll correction and non-roll correction done.

Tacview-20240413-141948-DCS.zip.acmi

[Raven (Elysian Angel)]

Just now, PanicMan15 said:

Were you correcting the roll with right roll?

No I wasn’t because on my end (when I have control) there is nothing to correct. I simply pulled back on the stick.

Even doing half-Cuban-8s is dead easy. In your mission (no wind) I start at heading 000 and end up exactly at 180 at roll-out (not 179, not 181 but exactly 180).

I’m not claiming there is no problem when you don’t have appropriate hardware. I agree with many people that there should be a Special Option that “translates” the FCLS for conventional hardware - if that’s even possible. 

I also don’t know how Winwing’s software works (as apparently the Orion 2 is based on the real F-16 stick just as the R3L is), but the Realsimulator software claims it works as the real F-16 stick does - and based on what I see in DCS I see no reason to doubt that.

[PanicMan15]

35 minutes ago, Raven (Elysian Angel) said:

No I wasn’t because on my end (when I have control) there is nothing to correct. I simply pulled back on the stick.

Even doing half-Cuban-8s is dead easy. In your mission (no wind) I start at heading 000 and end up exactly at 180 at roll-out (not 179, not 181 but exactly 180).

I’m not claiming there is no problem when you don’t have appropriate hardware. I agree with many people that there should be a Special Option that “translates” the FCLS for conventional hardware - if that’s even possible. 

I also don’t know how Winwing’s software works (as apparently the Orion 2 is based on the real F-16 stick just as the R3L is), but the Realsimulator software claims it works as the real F-16 stick does - and based on what I see in DCS I see no reason to doubt that.

You are actually bringing up a really good point. There has been talks that the F-16 in DCS is made for a FSSB. It could very well be something to do with hardware differences between the R3L and other sticks. I have a friend who has one and will test it out. Thanks!

[Raven (Elysian Angel)]

It is no secret that ED used R3Ls when developing the F-16’s FM. Wags has one, and he used it in his YT tutorials.

7 minutes ago, PanicMan15 said:

hardware differences

Software too possibly: either configuration differences between Realsimulator and Winwing or programming behind the scenes.

[PanicMan15]

4 minutes ago, Raven (Elysian Angel) said:

It is no secret that ED used R3Ls when developing the F-16’s FM. Wags has one, and he used it in his YT tutorials.

Software too possibly: either configuration differences between Realsimulator and Winwing or programming behind the scenes.

There were 3 different sticks used in the test, Winwing FSSB, T16000 and a Virpil WarBRD. It wouldn't make sense for all 3 to have software config issues.

[Raven (Elysian Angel)]

14 minutes ago, PanicMan15 said:

It wouldn't make sense for all 3 to have software config issues.

Sorry if I wasn’t clear but that’s not what I meant.

There is an option in the Realsimulator software that emulates behaviour that is coded in the real F-16 flight control software. As in: it was specifically designed for it and it’s actually what I was referring to in my first reply in this thread.

If those 3 alternatives were not specifically coded to emulate that behaviour as well, that would explain the uncommanded roll you are talking about. I don’t think you can blame those 3 manufacturers for that though: consumer hardware is generally designed to be as compatible as possible and to work the same way regardless of the aircraft you’re virtually flying.

But with the FSSB R3L that’s different. Sure there are options that make it more customisable (you can use the index finger switch as a hat for example while the real F-16 doesn’t work that way) but you can disable that feature if you want to use it as a real F-16 stick.

That is the point I am making: the FSSB R3L both on a hardware and software level is specifically designed to mimic the real F-16 stick and how it works in the actual aircraft.

Edited April 14 by Raven (Elysian Angel)
Minor grammar fix

[Jackjack171]

49 minutes ago, Raven (Elysian Angel) said:

Sorry if I wasn’t clear but that’s not what I meant.

There is an option in the Realsimulator software that emulates behaviour that is coded in the real F-16 flight control software. As in: it was specifically designed for it and it’s actually what I was referring to in my first reply in this thread.

If those 3 alternatives were not specifically coded to emulate that behaviour as well, that would explain the uncommanded roll you are talking about. I don’t think you can blame those 3 manufacturers for that though: consumer hardware is generally designed to be as compatible as possible and to work the same way regardless of the aircraft you’re virtually flying.

But with the FSSB R3L that’s different. Sure there are options that make it more customisable (you can use the index finger switch as a hat for example while the real F-16 doesn’t work that way) but you can disable that feature if you want to use it as a real F-16 stick.

That is the point I am making: the FSSB R3L both on a hardware and software level is specifically designed to mimic the real F-16 stick and how it works in the actual aircraft.

 

May I ask, did you purchase the grip or the base or the entire setup? I've been wanting this setup for a while but it's always out-of-stock. I currently have the WinWing F-16EX and it's almost overly sensitive or not enough. 

Didn't meant to derail the current topic of a roll problem.

Thanks in advance.

[Raven (Elysian Angel)]

4 minutes ago, Jackjack171 said:

Didn't meant to derail the current topic of a roll problem.

This is as good a place as any to ask that question

I bought both the grip and base. I have an older version though, that doesn’t deflect at all.

The current version of the R3L allows for the same 1/4” deflection the real F-16 stick does.

[Jackjack171]

2 minutes ago, Raven (Elysian Angel) said:

This is as good a place as any to ask that question

I bought both the grip and base. I have an older version though, that doesn’t deflect at all.

The current version of the R3L allows for the same 1/4” deflection the real F-16 stick does.

Cool. Thank you! 
Would the angle adapter base be a thing or is that overkill? None of the grips are available so I'm thinking of purchasing the base and dusting off my TMWH grip. I understand what you said about the programming. Probably why they are never in stock. Must be good equipment!

[Raven (Elysian Angel)]

11 minutes ago, Jackjack171 said:

Must be good equipment

Yes it is. It takes a while and some back-and-forth experimentation but once it’s set up correctly it’s a brilliant piece of kit.

I’d recommend the adapter yes if you can’t get hold of a Realsimulator grip. It simply makes it more ergonomic.

If you end up with one feel free to drop me PM here on the forum and I’ll send you a screenshot of my settings so you can have a solid base to start from - everyone has their own preferences primarily as to how much force is needed for full “deflection”.

The software explains all the settings very well and there are YT guides as well.

[Jackjack171]

48 minutes ago, Raven (Elysian Angel) said:

 

If you end up with one feel free to drop me PM here on the forum and I’ll send you a screenshot of my settings so you can have a solid base to start from - everyone has their own preferences primarily as to how much force is needed for full “deflection”.

 

Will do. And thank you. I really appreciate it!

[Glide]

Could it be the rudder?  Slightly out of position, one foot heavier than the other? 

[PanicMan15]

1 hour ago, Glide said:

Could it be the rudder?  Slightly out of position, one foot heavier than the other? 

Feel free to download and watch the track. No input but Y axis was used.

2 hours ago, Raven (Elysian Angel) said:

Yes it is. It takes a while and some back-and-forth experimentation but once it’s set up correctly it’s a brilliant piece of kit.

I’d recommend the adapter yes if you can’t get hold of a Realsimulator grip. It simply makes it more ergonomic.

If you end up with one feel free to drop me PM here on the forum and I’ll send you a screenshot of my settings so you can have a solid base to start from - everyone has their own preferences primarily as to how much force is needed for full “deflection”.

The software explains all the settings very well and there are YT guides as well.

Realsimulator FSSB is doing the same thing. If you could, please show me your test with indicator on.

[Glide]

21 minutes ago, PanicMan15 said:

Feel free to download and watch the track. No input but Y axis was used.

Do you mind performing a simple test with just keyboard inputs?  I can see what you are describing in the ACMI file.

Edited April 15 by Glide

[PanicMan15]

1 hour ago, Glide said:

Do you mind performing a simple test with just keyboard inputs?  I can see what you are describing in the ACMI file.

 

Will do. Give me a second to record and upload the video.

[PanicMan15]

1 hour ago, Glide said:

Do you mind performing a simple test with just keyboard inputs?  I can see what you are describing in the ACMI file.

 

Here is the video with a simple loop using ONLY keyboard inputs

[PanicMan15]

So funny little thing. I reverted back and tested it again in 2.9.3.51704. Issue still persist there.

[NineLine]

This effect is from the modelling of the gyroscopic effect of the engine. We believe it is correct as is. Thanks all. 

[Teej]

20 hours ago, Raven (Elysian Angel) said:

Yes it is. It takes a while and some back-and-forth experimentation but once it’s set up correctly it’s a brilliant piece of kit.

I’d recommend the adapter yes if you can’t get hold of a Realsimulator grip. It simply makes it more ergonomic.

If you end up with one feel free to drop me PM here on the forum and I’ll send you a screenshot of my settings so you can have a solid base to start from - everyone has their own preferences primarily as to how much force is needed for full “deflection”.

The software explains all the settings very well and there are YT guides as well.

Hey, Raven...

I have an R3L and would be curious to see your settings, but let me make some comments here too....  

I am saying the following NOT as any kind of flex, just to put my comments in perspective.  

1: I am a (private) pilot - but not of any jet, let alone a Viper. I fly an RV-10. 

2: I have a good background in controls and control theory. ~ 15 years ago I designed and successfully implemented a "hold" function for radio-controlled helicopters. The idea was to flip a switch and be able to let the helicopter fly itself while you worked a camera on board. This was before drones/etc were "a thing". 

3: I wrote my own version of an aerobatic team tool ("Phantom Control") for FC3 SFM after the Berkuts' legacy Phantom no longer worked / no longer supported.

4: I've written a simple autopilot that will read data from a text file and fly the viper (or any plane, with some tweaks I'm sure) around a "flight plan" similar to what I would do in the real world with an autopilot coupled to a GPS navigator. Headings, alittude, airspeed etc.

5: I've written code that "linearizes" the ED Viper throttle so that it's a flat response across the range rather than being aggressive at low % and less responsive at higher end.

Again - not a flex, just commenting on my background.

Now...  What happens with joystick input is this. 

1: Your device (FSSB, Warthog, t16000, Virpil, whatever) reads the stick position and assigns it X and Y values. Let's say it's a 16-bit device. The device will then be reporting each axis to Windows as an integer in the range of -32767 to +32767.  

2: DCS scales those numbers to floating points ranging from -1.0 to +1.0.  

3: DCS process accordingly.

The only way DCS knows your device is an FSSB is from the text identifier it picks up from DirectInput. 

If in fact your FSSB is doing "something different" from any other stick and "dialing out" this left rolling tendency, then one has to consider it will vary based on airspeed, jet weight, symmetry of loading, etc. DCS would have to send that data to the FSSB and the FSSB would have to process your inputs and "do things" based on DCS data before sending its pitch/roll inputs in to Windows. I can state with an _extremely_ high level of confidence that this is not occurring. Could I be wrong? Sure. But I'd bet BIG money against it, if I were a gambler. I would be shocked beyond words if anything like that were happening.

Edit: Went into the software to look.  RCA turns down roll authority (saturation) when pitch is put in. Absolutely not the issue here. Jet rolls without any roll input.

So as I said - I have an FSSB R3L and would be glad/curious to try whatever settings you're using.

I will note that I believe Panic's video is exaggerated due to the fact that with the keyboard input, it's essentially trying to command ~ 4G even though over the top there's not enough smash, thus it's "stalling" a bit.     

I will agree with ED in that there are portions of the flight regime where the F-16 should have some undesired/unintended roll occurring. Whether it's inertial (asymmetric weight), reaction (to the turning of the engine) or precession (gyroscopic) I don't know. I do know the F-16 FLCS won't "auto trim" in roll like it does in pitch (attempting to maintain +1.0G unless input to the contrary). Also, many jets are "bent" a little bit requiring a unique roll/yaw trim for every jet and as with any sort of trim, you can only trim for one airspeed at a time.  

So...yeah.  Let's see what you've got!

Edited April 15 by Teej