reported G Onset and G Visual Effects

[NineLine]

14 hours ago, DummyCatz said:

A bug related to g-response and pitch instability is fixed in 2.9, which could also affect g-onset if more pitch dampening is applied:

Using the example of going from 9g to 1g by releasing the stick, in DCS 2.7 and later 2.8 (with partial fix) this looked like going from 9g to 0.5g rather quickly, undershooting below 1g and then slowly creeps back to 1g, especially at speeds between 400 to 500 kts, displaying a lack of dynamic stability. Now it looks like 9g -> 2g -> 1g as a 1st-order g-response, which is a correct response type as IRL is. But the bug fix introduces pitch creep, which is to say, upon releasing the stick the nose tends to continue creep a bit more degrees until it stops.

Thanks for the report you made, but it is off-topic here and under investigation in that report. Thanks. 

[FusRoPotato]

Unfortunately, I haven't been able to find any sources that describe a possible source for it. The signal and servo input runs at a very high frequency of 300hz, which suggest there probably isn't one, but there's no way to know what else it would come from. 

I've measured it using export lua and commanded input scripting (to avoid any possible issue with physical joystick delays). It floats between 50 and 60 ms when the framerate is locked to 120, while no other aircraft I've recorded shows this including the F-18. It's a small delay that most might not be able to "feel", but some of us can. This is essentially 6 frames of delay at 120, but I haven't done a run yet at 60fps to see what that's like. If it's based on frame count rather than time, it could be much worse for people who have poor frame rates.

Being really heavy into estimation and filtering techniques, I'd put my guess on a filtering mistake somewhere in the joystick translation scheme. It may be waiting for a history or 3-4 input ticks before deciding how fast and far the input curve should move. 

IMO, it would probably be best to undo anything complicated with the stick interpretation and allow something more direct.

Edited November 16, 2023 by FusRoPotato

[Sacarino111]

Hi.

Definitively yes, there is lag, but I found it happens when on the ground; I tryed both grounded and airborne and found that in the air the response is not lagged. I did some testing with other planes and the F 18 also has lag, but smaller. I talk about roll; pitch has really small lag.

I did cross check with the control monitor ((RCTRL+ENT), and deffinitivelly, it was the visual mouvement the one with lag; the input monitor was coherent wit the stick deflection.

I also did try with CAT I and III and the response was the same speed but less throw (the ailerons move less deflection).

Anyhow, as airborne I can't feel/see no lag, I find the manouverability optimal (as always).

Saludos.

Saca111

[fapador]

On 11/3/2023 at 12:13 AM, Nedum said:

I watched the video. It looks to me as if the new FM is slower. The circle is completed in the same time. With the new FM there is less speed left and the blackout comes faster and takes longer.

The video shows exactly what I thought after the first few tests. I lose more speed and the blackout comes quicker and lasts longer. And I feel like the rolling axe is very itchy.

I need to test it again, but I also feel like I black out before I even get to the 9G.

Am I the only one who thinks this is the opposite of better? I don't see anything there that is better.

I understand your doubts,.... but trust me the new g-onset is much more realistic and far more correct than the previous. I have experience on aircraft manufacturers sim and I can say that pulling g's in DCS as of now is almost exactly the same.

Personally, I am happy with the update. ED has nailed  it and I am quite a tough guy to be satisfied with such things.

I have to mention though, in order to be exact and fair that there are other aspects of the viper fm that can be still improved further, but that is for another topic.

 

Edited November 17, 2023 by fapador

[Buzz313th]

How can people comment on the realistic FM of the Viper unless they have flown one?  

 

To get a feeling of roughly what 9G's might feel like, go down to your local airport and buy a demo hour in an Acro Aircraft.  A "Cheap" hour would be in a Decathalon.  It is rated to +6G -4G.  Take that ride and ask the pilot to show you +5 - +6G at the bottom of a loop.  Tell me if you could keep your head up and stop staring at your own crotch.    If you can find something with a bigger engine, Pitts, Eagle, Extra..  Then you can experience longer duration at +6 and possibly higher.    Then come back and tell us how many seconds of +9G a Viper Pilot pulls. 

Edited November 17, 2023 by Buzz313th

[darkman222]

Before you get too excited here. This thread is about the FM. And we have charts for that. After them ED is modelling the F16.

The second thing, which is off topic here, is in another ongoing thread about the g modelling. Short version: The reason why you or me don't fly F16 because we possibly would not be qualified to do so. F16 pilots are demanded to withstand a certain G profile. Otherwise they simply would not be allowed in the cockpit. And the simulated DCS pilot is not qualified. We are not talking about vague opinions. We talk about a minimum set amount of time of 9G for the pilot by the military. And not about the g tolerance of a civilian on a rollercoaster.

 

Because the simulated DCS pilot is a limiting factor for the F16 flight model is the reason for this thread going off topic.

Look at the posted G profile in this thread and feel free to try it in DCS. And post your opinion in it if you like.

 

Edited November 18, 2023 by darkman222

[NineLine]

Locked as this is reported.

Thank you for your feedback on items like G onset rates and G visual effects. We are quite aware of these items, thank you, and we are investigating.

[Noctrach]

Observation:
It's measurable in-game that no matter the controller, OS or hardware configuration, there's a ~60ms delay between an input on the F-16 and the FLCS response. You can see this across a slew of different controls types, but it's especially demonstrable with very twitchy joysticks or good low-latency keyboards.

This same latency can not be reproduced on any other airplane

Question, without judgement:
Is this intentional behaviour for this jet, and can you shed light on what causes it? (i.e. Why is this a feature unique to the F-16)

[DummyCatz]

On 2/24/2023 at 4:50 AM, 85th_Maverick said:

Not in roll and yaw, only in pitch for the Su-27! And speaking of witch, why does the Su-33 have a quick elevators response in contrast to the low elevators rate of the Su-27 if you happen to have in depth knowledge about the differences (why would there be any) between the Su-27's FCS and the Su-33's FCS?

I'm not sure about the Su-33, but the lag filter in Su-27 is to reduce sensitivity in pitch control especially at higher speeds, as it poses a danger in pitch control over-sensitivity. The same can be said for the F-16 as there's a 'pitch prefilter' applied after the pitch 'stick command limiter', and there's also a known issue of pitch control over-sensitivity at transonic speeds in the F-16.

According to Figure 3.1 in the paper "F-16 Simulator for Man-in-the-Loop Testing of Aircraft Control Systems", https://apps.dtic.mil/sti/citations/ADA189675, the pitch prefilter is a first-order low-pass filter (a.k.a lag filter so yes it creates lag), with a time constant of 1/N14. The gain N14 is variable, scheduled with dynamic pressure. With a dynamic pressure below 250 psf (~= 266KCAS), the time constant is 1/8.3 sec, which means the shaped g-command will be reaching 63.2% of the commanded g value in 1/8.3 seconds, given a step input. With a dynamic pressure greater than 400 psf (~= 585KCAS), the time constant is 1/6 sec, lowering the control sensitivity at higher speeds.

This essentially shapes the desired first-order g-response of the aircraft, with no overshoots nor oscillations. This may explain both the lag and the g-onset.

An illustration of first-order response given a step input:

The location of pitch prefilter as shown in Figure 3.1, https://apps.dtic.mil/sti/citations/ADA189675

[NineLine]

Dear all,

Due to your concerns, the team has reviewed G-Onset and is currently correct based on our review. Please see the chart and explanation below. (Also note we are actively adjusting the F-16C visual G effects. It will be available in a later update).

Points have been added for easier understanding. You can see two lines on the graph, the black one is G's as a function of time and the blue is the control input (pitch) as a function of time. You can also see points on the chart. Two on each line. These points show time and value at these points.

Let's look at the control input line first. The first point at 6.35 seconds and a value of 0, this is the last point before you pull the stick back. We then placed a point on the G line at the same time (6.35 seconds, 1.0107 G). Then we find the point where the G value is 6Gs, check the bottom axis and this value is 7.358 seconds.

So 7.358 sec minus 6.35 sec equals 1.008 seconds, this value is the time to 6Gs when pulling back on the stick. 

As always, if you have further concerns, please include a track, and make sure you are not at too high of an altitude, too heavy or at CAT III, as well as being too fast or too slow. Also, note that having your controls set properly (remove any curve), will impact this as well. Under correct conditions, our Viper can certainly achieve 6G within one second.

[Buzz313th]

22 hours ago, NineLine said:

Dear all,

Due to your concerns, the team has reviewed G-Onset and is currently correct based on our review. Please see the chart and explanation below. (Also note we are actively adjusting the F-16C visual G effects. It will be available in a later update).

Points have been added for easier understanding. You can see two lines on the graph, the black one is G's as a function of time and the blue is the control input (pitch) as a function of time. You can also see points on the chart. Two on each line. These points show time and value at these points.

Let's look at the control input line first. The first point at 6.35 seconds and a value of 0, this is the last point before you pull the stick back. We then placed a point on the G line at the same time (6.35 seconds, 1.0107 G). Then we find the point where the G value is 6Gs, check the bottom axis and this value is 7.358 seconds.

So 7.358 sec minus 6.35 sec equals 1.008 seconds, this value is the time to 6Gs when pulling back on the stick. 

As always, if you have further concerns, please include a track, and make sure you are not at too high of an altitude, too heavy or at CAT III, as well as being too fast or too slow. Also, note that having your controls set properly (remove any curve), will impact this as well. Under correct conditions, our Viper can certainly achieve 6G within one second.

What was the starting config of the jet?