reported Roll axis Incredibly sluggish.

[fapador]

As the title says I find the viper's roll axis  incredibly sluggish even at empty loads. Years ago I had a go at the Lockheed Martin Viper Simulator and I remember the high roll inertia distinctively,  DCS F16 is very far from what I remember. I tried different  curve settings but all failed to satisfy me 100%. I cannot pull sharp instantaneous  wings levelling, a maneuver   I have seen live on many Aerobatic shows like in the Video I post. According to what I have read on  the book Vipers in the Storm  Captain Keith Rosenkranz  states that F16 In fact requires opposite stick movement when banking sharp in order to stop roll due to high roll inertia.  For this I am creating this post. He also states that at high speeds F16 FLCS becomes incredibly sensitive to the slightest movement something also not currently present.

Anyone out there having the patience (I lack) willing to perform tests? What are your  opinions ?

I am using a t16000 full linear setup

Look at 3:56, 9:15 and 10:34

 

 

Edited April 2, 2022 by fapador

[Tholozor]

What position is your CAT stores config switch set to?

[fapador]

I

 

20 minutes ago, Tholozor said:

What position is your CAT stores config switch set to?

 

[Dragon1-1]

There's something up with the roll axis on the Viper. A lag or deadzone of some sort, not possible to adjust in the controls menu. It seems to be a controls issue, not an FM issue, but I might be wrong. More noticeable in CAT III than in CAT I, but present in both. Roll rate is fine, but that deadzone makes it feel really bad.

[fapador]

Yeah Roll rate (Max) might be good, I never questioned that. However the roll axis behavior is way off. Might be the way ED has chosen to replicate real non moving stick implementation to common joysticks. It just doesn't feel correct though.

3 minutes ago, Dragon1-1 said:

There's something up with the roll axis on the Viper. A lag or deadzone of some sort, not possible to adjust in the controls menu. It seems to be a controls issue, not an FM issue, but I might be wrong. More noticeable in CAT III than in CAT I, but present in both. Roll rate is fine, but that deadzone makes it feel really bad.

 

[Dragon1-1]

It's almost certainly a misbegotten attempt at translating how the real Viper stick works. If so, it doesn't really work, force sensing can't really be "emulated" like this. Anyone with an actual force sensing base can setup things like breakout force and the like in software, meanwhile everyone else should be given a response that works with the majority of sticks on the market.

[janitha2]

yeah the roll rate is very slow compared to any video aaand other sims

[BIGNEWY]

19 hours ago, fapador said:

As the title says I find the viper's roll axis  incredibly sluggish even at empty loads. Years ago I had a go at the Lockheed Martin Viper Simulator and I remember the high roll inertia distinctively,  DCS F16 is very far from what I remember. I tried different  curve settings but all failed to satisfy me 100%. I cannot pull sharp instantaneous  wings levelling, a maneuver   I have seen live on many Aerobatic shows like in the Video I post. According to what I have read on  the book Vipers in the Storm  Captain Keith Rosenkranz  states that F16 In fact requires opposite stick movement when banking sharp in order to stop roll due to high roll inertia.  For this I am creating this post. He also states that at high speeds F16 FLCS becomes incredibly sensitive to the slightest movement something also not currently present.

Anyone out there having the patience (I lack) willing to perform tests? What are your  opinions ?

I am using a t16000 full linear setup

Look at 3:56, 9:15 and 10:34

 

 

 

Hi do you have any evidence other than a video? and can you include your track replays from your tests. 

thanks

[ExA4K]

On 4/3/2022 at 4:39 AM, fapador said:

As the title says I find the viper's roll axis  incredibly sluggish even at empty loads. Years ago I had a go at the Lockheed Martin Viper Simulator and I remember the high roll inertia distinctively,  DCS F16 is very far from what I remember. I tried different  curve settings but all failed to satisfy me 100%. I cannot pull sharp instantaneous  wings levelling, a maneuver   I have seen live on many Aerobatic shows like in the Video I post. According to what I have read on  the book Vipers in the Storm  Captain Keith Rosenkranz  states that F16 In fact requires opposite stick movement when banking sharp in order to stop roll due to high roll inertia.  For this I am creating this post. He also states that at high speeds F16 FLCS becomes incredibly sensitive to the slightest movement something also not currently present.

Anyone out there having the patience (I lack) willing to perform tests? What are your  opinions ?

I am using a t16000 full linear setup

Look at 3:56, 9:15 and 10:34

 

 

 

Interesting that a Lockheed Martin test pilot states the contrary for having to check against the roll inertia (like a conventional control system):

https://www.codeonemagazine.com/article.html?item_id=174

Quote

 

The F-16's very good ride at high airspeed and low altitude is one result of this self-checking feature. As soon as the F-16 is disturbed by any type of turbulence, the flight control system has a correction in, almost before you can think about it.

This self-correcting feature is why you see the horizontal stabilizer moving around so much when the airplane is taxiing. The flight control system is not getting any input from you, but is feeling the aircraft move as you taxi across all the bumps on the taxi route. So what you see is the flight control system trying to smooth out the taxiway. This is also why you don’t have to put in any check command to stop the roll rate as you try to do any number of precision point rolls.

One minor drawback of the self-checking feature shows up in what has been described as roll ratcheting. You will recall earlier that I talked about how different the flight control system is, compared to what you’ve been using. The ability to do smooth rolls requires some concentration on your part until you become completely familiar with this different airplane. What’s happening is that you’re putting in some amount of roll command. Since the roll acceleration of the F-16 is so good, you make the subconscious decision that if you’re rolling this fast and this quick then in a couple of seconds you will be rolling at nine million rpm.

The natural tendency is to want to slow the roll rate. With a conventional flight control system, we simply decrease the amount of stick deflection. In order to accomplish this, we relax pressure on the stick and allow the self-centering forces to move the stick closer to center (i.e., less aileron deflection), thus slowing the roll rate to what we want; then apply sufficient pressure to keep the stick at the new position. This relaxing of the pressure will normally go to zero momentarily, and with the F-16 this is sufficient for the self-checking feature to stop the roll rate completely. (Remember — you don’t have direct control over the amount or direction of the control surface deflection.) The roll rate deceleration is ALSO rapid, so your body and hand tend to couple with the aircraft motion and probably make stick inputs that weren’t intended.

The end result is some pretty sloppy rolls until we get used to the system. What you need to do is (1) learn to adjust the roll rate with subtle pressure changes to the stick and (2) get away from the stick position cues you’ve been used to using. Once you can get yourself tuned to using finite pressure changes to control the roll rate, you’ll be able to make smooth roll inputs. This is so despite a force-per-roll-rate slope that isn’t constant. There are two distinct changes in the slope of the curve. This is to make sure that the airplane isn’t too sensitive for small inputs, and that the force required for max inputs is not too high. Those devilish engineers also used two different roll time constants for small and large roll inputs. All this is nice to know, but if you simply pay attention to the amount of force you’re using on the stick, you’ll be able to do very nice rolls with the F-16.

By now I’m sure all of you are asking why it’s necessary to use such a markedly different flight control system. Well this self-checking feature is really one of the main reasons this flight control system is in the F-16. It allows an aircraft design that uses new and different aerodynamic principles.

 

 

[LJQCN101]

FYI apart from the built-in software breakout in the DFLCS, the roll rate command is also gone through a first order lag filter, which will induce lag and reduce sensitivity of the input.

The transfer function is 5/(s+5), which will change to 2.5/(s+2.5) depends on conditions (i.e. at higher dynamic pressures).

An example from SimuLink using a 324 deg/s roll rate command as input:

It would take around 1 sec to go from 0 to 324 with a transfer function of 5/(s+5).

[jayst0r]

its hard to tell in the video, i tried to playback at 25% but it seems he stompt the rudder really quick to initiate the roll.

[fapador]

The result is an F-16 that achieves 324 degrees per second maximum roll-rate command within the first ninety degrees

 

Instead, consider this: the F-16 is as fast to ninety degrees of bank as just about anything you'll run across; and although there are areas of the envelope where the computer limits the F-16 to less than 100 degrees per second, you still have nearly twice the roll rate available, under similar conditions, as any adversary you may meet.

 

Very interesting review and site from a test pilot that sadly lost his life in a F16 . However note the above about roll rate. Currently DCS F-18 is much more sensitive and thus more agile in roll axis than F-16 in the same full linear setup , I can bank DCS F-18 much faster in ninety degrees than F-16, it doesnt make sense.

9 hours ago, ExA4K said:

Interesting that a Lockheed Martin test pilot states the contrary for having to check against the roll inertia (like a conventional control system):

https://www.codeonemagazine.com/article.html?item_id=174

 

7 hours ago, LJQCN101 said:

FYI apart from the built-in software breakout in the DFLCS, the roll rate command is also gone through a first order lag filter, which will induce lag and reduce sensitivity of the input.

The transfer function is 5/(s+5), which will change to 2.5/(s+2.5) depends on conditions (i.e. at higher dynamic pressures).

An example from SimuLink using a 324 deg/s roll rate command as input:

Yes the lag is very much sensible in FBW systems. However, note that it doesn't reduce the commanded roll rate or the sensitivity thus the control surfaces amount of deflection (in a way that DCS does) it just delays for some ms the command from happening. Its hard to describe if you have never  had a go  on a FBW plane.

 

The term reduce sensitivity is incorrect IMAO. It just makes the command happen at a later time (few ms). 

Edited April 4, 2022 by fapador

[LJQCN101]

13 hours ago, fapador said:

Yes the lag is very much sensible in FBW systems. However, note that it doesn't reduce the commanded roll rate or the sensitivity thus the control surfaces amount of deflection (in a way that DCS does) it just delays for some ms the command from happening. Its hard to describe if you have never  had a go  on a FBW plane.

 

The term reduce sensitivity is incorrect IMAO. It just makes the command happen at a later time (few ms). 

 

I agree it may not be the best/correct term out there (if you consider the meaning in signal processing). A better term would be reduce input noise or jitter. Surely it does not reduce the maximum command with the transfer function. Roll rate command limiter is another block in the diagram.

The FCS of F18 doesn't have such designs. It has a 5 Hz notch filter applied to the stick lateral position sensor (Filter R2) but that's it. I'm yet to spot any lag filters in the path.

So the command can be sent to the aileron without any prominent delays.

 

On the other hand, I myself is a X65F user and I set 22 lbs for pitch and 13 lbs for roll, with the stick well fixed to the desk. Compared to T16000M that I also use as a backup, the time for me to reach the maximum force/stick deflection from neutral is a bit longer with a X65 than a T16000M, because of the much larger stick force. The end result is that I can feel more lag in the T16000M when flying the Viper because I can yank the stick so quick, the instantaneous lag between the filtered command and non-filtered command is very significant. But with the X65F, it's like you're moving the stick not as quickly, and the lag between the filtered command and non-filtered command is not so significant.

Edited April 5, 2022 by LJQCN101

[fapador]

 

13 hours ago, LJQCN101 said:

I agree it may not be the best/correct term out there (if you consider the meaning in signal processing). A better term would be reduce input noise or jitter. Surely it does not reduce the maximum command with the transfer function. Roll rate command limiter is another block in the diagram.

The FCS of F18 doesn't have such designs. It has a 5 Hz notch filter applied to the stick lateral position sensor (Filter R2) but that's it. I'm yet to spot any lag filters in the path.

So the command can be sent to the aileron without any prominent delays.

 

On the other hand, I myself is a X65F user and I set 22 lbs for pitch and 13 lbs for roll, with the stick well fixed to the desk. Compared to T16000M that I also use as a backup, the time for me to reach the maximum force/stick deflection from neutral is a bit longer with a X65 than a T16000M, because of the much larger stick force. The end result is that I can feel more lag in the T16000M when flying the Viper because I can yank the stick so quick, the instantaneous lag between the filtered command and non-filtered command is very significant. But with the X65F, it's like you're moving the stick not as quickly, and the lag between the filtered command and non-filtered command is not so significant.

Edited 8 hours ago by LJQCN101

I am currently using a curve at roll axis to improve things to my personal taste. My curve is sophisticatedly setup 'd  in a way to better  emulate lag  "artificially" and increase responsiveness in the near center zone. This is just a sneaky workaround I am working to  further  improve ,but you can give it a go . 

It makes things more realistic than default IMO but its still not Perfect compared to a real F16 sidestick.

PS: It might make AR a little harder and possibly  easier to overcontrol but it's closer to what I would expect. Feel free to switch back to your previous setup and note your current settings in case you don't like. 

Edited April 5, 2022 by fapador

[Mikaa]

I wonder if it would be easier for ED to implement a modified curve for non-FSSB users (which is the majority of us), similar to the central trimmer mode options for Helos (or F-5). That way those who have a replica Viper Force Sense stick can use realistic the breakout forces and gains, but the rest of us with physically moving sticks can still have a similarly snappy Viper out of the box. Just a thought. 

 

Edited April 5, 2022 by Mikaa
words

[SCPanda]

On 4/3/2022 at 8:13 AM, fapador said:

Yeah Roll rate (Max) might be good, I never questioned that. However the roll axis behavior is way off. Might be the way ED has chosen to replicate real non moving stick implementation to common joysticks. It just doesn't feel correct though.

 

 

On 4/3/2022 at 8:19 AM, Dragon1-1 said:

It's almost certainly a misbegotten attempt at translating how the real Viper stick works. If so, it doesn't really work, force sensing can't really be "emulated" like this. Anyone with an actual force sensing base can setup things like breakout force and the like in software, meanwhile everyone else should be given a response that works with the majority of sticks on the market.

 

9 hours ago, Mikaa said:

I wonder if it would be easier for ED to implement a modified curve for non-FSSB users (which is the majority of us), similar to the central trimmer mode options for Helos (or F-5). That way those who have a replica Viper Force Sense stick can use realistic the breakout forces and gains, but the rest of us with physically moving sticks can still have a similarly snappy Viper out of the box. Just a thought. 

 

 

As a FSSB user, I'm telling you guys that force sensing does not make things any better. 

Even with the force sensing stick, flying with 0 curves in DCS, the jet still feels very sluggish in roll. Feels like I am flying a 747. I am not even using a high force settings with my FSSB. My max pitch force is 6.25 lbs, and my max roll force is 3.25 lbs (default is 8.5 lbs and 4.33 lbs). In the other F-16 sim (BMS), roll feels much more responsive. 

To describe the feeling, in the other F-16 sim, if I want to achieve X degrees of roll rate, I only have to apply 50% of my total force setting in roll to my force sensing stick, in DCS, however, I have to apply about 80-90% of force to achieve the same X degrees of roll rate. 

This issue in DCS is NOT the problem of your hardware. Buying a FSSB R3L base WILL NOT help. 

Still, like the tag said, we have to provide some evidence to ED that DCS F-16's FLCS is very insensitive when receiving roll inputs from the pilot, whether the input is from a force sensing stick or not. 

Edit: In case you wonder, I use same the force setting in DCS and the other F-16 sim. 

Edited April 6, 2022 by SCPanda

[Mikaa]

1 hour ago, SCPanda said:

 

 

As a FSSB user, I'm telling you guys that force sensing does not make things any better. 

Even with the force sensing stick, flying with 0 curves in DCS, the jet still feels very sluggish in roll. Feels like I am flying a 747. I am not even using a high force settings with my FSSB. My max pitch force is 6.25 lbs, and my max roll force is 3.25 lbs (default is 8.5 lbs and 4.33 lbs). In the other F-16 sim (BMS), roll feels much more responsive. 

To describe the feeling, in the other F-16 sim, if I want to achieve X degrees of roll rate, I only have to apply 50% of my total force setting in roll to my force sensing stick, in DCS, however, I have to apply about 80-90% of force to achieve the same X degrees of roll rate. 

This issue in DCS is NOT the problem of your hardware. Buying a FSSB R3L base WILL NOT help. 

Still, like the tag said, we have to provide some evidence to ED that DCS F-16's FLCS is very insensitive when receiving roll inputs from the pilot, whether the input is from a force sensing stick or not. 

Edit: In case you wonder, I use same the force setting in DCS and the other F-16 sim. 

 

Thanks for your feedback. The plot thickens…

I also fly the other sim, and have had no issues with my TM gear in terms of roll response. Hopefully someone well versed in the FLCS can shed some light on what might be causing this. 

 

just spitballing, but could it be that some of the FBW gain filters are being accidentally applied more than once? Either way, I’m out of my depth for this aircraft, however it anecdotally seems a bit sluggish when roll is initially commanded. 

[SCPanda]

23 minutes ago, Mikaa said:

Thanks for your feedback. The plot thickens…

I also fly the other sim, and have had no issues with my TM gear in terms of roll response. Hopefully someone well versed in the FLCS can shed some light on what might be causing this. 

 

just spitballing, but could it be that some of the FBW gain filters are being accidentally applied more than once? Either way, I’m out of my depth for this aircraft, however it anecdotally seems a bit sluggish when roll is initially commanded. 

As far as I know, the other sim modeled F-16's FLCS and force sensing stick very well. I don't know whether the other sim or DCS is correct, and I cannot compare it to DCS since ED does not allow it. 

[LJQCN101]

Okay I did a simple test to run through some figures according to the DFLCS block diagram.

Some preparations: (text in bold is for TL;DR)

1. I'm using active pause to stop the roll rate feedback from intervening the result, so that it's only about the stick input and the flaperon output.

2. I'm also choosing an high enough dynamic pressure (airspeed greater than 250 KCAS or Qc > 210 psf) and a low AOA (less than 4.2 degrees), and also elevator position less than 15 degrees TEU, to prevent the Roll Rate Command Limiter from reducing our maximum roll rate command. So the maximum roll rate command can stay at 324 deg/s.

 

By freezing any other parameters, the lateral control law can be simplified as this: (roll rate command - roll rate feedback) * 0.12, apply a -21.5 to 21.5 limit, and then apply a 1.5 deg mechanical bias, so that the final output to flaperon ISA is 20 TEU to 23 TED.

So essentially, you're able to command a maximum flaperon deflection with a 21.5/0.12=179.2 deg/s roll rate command. By checking the DFLCS built-in roll command gradient curve, it corresponds to 12.5 lbs stick force, which is 71% of the full stick force range (17.57 lbs).

Now we're getting a reference point to perform a test, to see whether the flaperon reaches its full deflection at 71% stick input.

 

Test result:

So no, I'm using a stick input at around 80% to reach the maximum flaperon deflection. It means either the roll rate command value is a bit low or the 0.12 gain is questionable.

 

EDIT: rechecked the DFLCS curve, 179.2 deg/s actually corresponds to 12.5 lbs stick force, but the test result still stands.

roll rate command test.trk

Edited April 7, 2022 by LJQCN101

[fapador]

5 hours ago, LJQCN101 said:

Okay I did a simple test to run through some figures according to the DFLCS block diagram.

Some preparations: (text in bold is for TL;DR)

1. I'm using active pause to stop the roll rate feedback from intervening the result, so that it's only about the stick input and the flaperon output.

2. I'm also choosing an high enough dynamic pressure (airspeed greater than 250 KCAS or Qc > 210 psf) and a low AOA (less than 4.2 degrees), and also elevator position less than 15 degrees TEU, to prevent the Roll Rate Command Limiter from reducing our maximum roll rate command. So the maximum roll rate command can stay at 324 deg/s.

 

By freezing any other parameters, the lateral control law can be simplified as this: (roll rate command - roll rate feedback) * 0.12, apply a -21.5 to 21.5 limit, and then apply a 1.5 deg mechanical bias, so that the final output to flaperon ISA is 20 TEU to 23 TED.

So essentially, you're able to command a maximum flaperon deflection with a 21.5/0.12=179.2 deg/s roll rate command. By checking the DFLCS built-in roll command gradient curve, it corresponds to 11.6 lbs stick force, which is 66% of the full stick force range (17.57 lbs).

Now we're getting a reference point to perform a test, to see whether the flaperon reaches its full def

Once again we find ourselves doing the work for ED... This is also only CAT I mode, CAT3 might be wrong as well so does the landing mode or even AR.

Great test LJQCN101 and I personally thank you. Try  my suggested curve it will also need 66% to reach maximum deflection but it will happen earlier in joystick travel due to curve, somehow fooling you and giving a better sense. I tried to somehow match it  as best as I could from my memory on my go at LM sim. Its not perfect but its better than default.

 

PS: As for the general FM I also think there is also something else  a little off. I find that the jets nose drops too much with bank angle in most speed ranges, resulting in poor turn performance (most complain as they cant reach 9g in turns) but I think its due to the nose tendency dropping too much when banking in any weight configuration( I even tried unloading all the gun ammo to reduce nose weight but that only  improves it very little).

So maybe something is also little off there, if not aerodynamic wise possibly  Ix Inertia is off or the rudder provides too much lift pushing the nose down when banking (it becomes like an elevator).

Edited April 6, 2022 by fapador

[LJQCN101]

I started to wonder if ED actually uses the roll command gradient curve from NASA FLCS (TP1538) and Analog FLCS, instead of DFLCS. So I did the maths.

 

In NASA FLCS and Analog FLCS, the roll gradient curve is pretty different from the one in DFLCS:

1. Analog FLCS requires 13.6 lbs stick force to command 179.2 deg/s, and 17 lbs to command 308 deg/s. Note that the maximum roll rate command is 308.

2. DFLCS requires 12.5 lbs to command 179.2 deg/s (as in the above test), and 17.57 lbs to command 324 deg/s. Note that the maximum roll rate command is 324.

 

(In case I'm drunk, the known figures in Analog FLCS roll command gradient curve are: 11 lbs for 80 deg/s, and linearly extends to 17 lbs for 308 deg/s. DFLCS: 9 lbs for 80 deg/s, and linearly extends to 17.57 lbs for 324 deg/s.)

 

So in Analog FLCS, it would require 13.6/17 = 80% stick input to command 179.2 deg/s, while in DFLCS the figure is only 71%.

And if you check the test result again, this exactly matches what currently is in DCS.

 

Smaller stick inputs may have had a bigger impact if Analog FLCS is used instead of DFLCS. For instance, at 11 lbs stick force, Analog FLCS is only able to command 80 deg/s roll rate, while with DFLCS the figure is 137 deg/s. That's a 71% increase.

Edited April 7, 2022 by LJQCN101

[LJQCN101]

14 hours ago, fapador said:

Once again we find ourselves doing the work for ED... This is also only CAT I mode, CAT3 might be wrong as well so does the landing mode or even AR.

 

I think CAT III and Takeoff & landing gains are worth testing too. The DFLCS control logic are:

In CAT III, roll rate command is reduced by 134 deg/s, plus the reduction by the dynamic roll rate command limiter. So the maximum roll rate command is 324-134=190 deg/s.

In Takeoff & landing gains, CAT switch is ignored. The maximum roll rate command is 167 deg/s.

[SCPanda]

@Wags@BIGNEWY @NineLine

Please take a look. Thx

[BIGNEWY]

We have a report open for roll rate.

thanks