Discussion: Issues with 6DOF

Fully agree!

Maybe ask on their forums why they don't export the gravity component. I'm sure they must use it internally, because otherwise it would be a mystery to me how they'd produce a credible racing simulation.

Those forums are to.... let's say ... avoidable for questions that might put in danger their simulation quality...

Easier to calculate and add it than ask for them.
Having rotation angles, it's just getting the vector components in each direction and add them to the other accelerations.
Just need to see the type of axis used... What a mess.

Perfect! A line that's pleasure to read!

Dirty

@Dirty , I think I understand, but I want to confirm.

I didn't think about the simulator telemetry including the continuous 1G accleration away from Earth's surface that is required to remain aloft.

So if I'm understanding this correctly I will get the same cues I described in my four-point roll example, but without any need to use the Euler Angles which have obvious issues.

For my example, the pilot rolls the (simulated) aircraft 90 degrees left bank and holds it there, flying as straight as possible. If, under these circumstances, the aircraft is not losing altitude it means that it is generating 1G of "lift" perpendicular to Earth's surface. (presumably with the rudder). So as far as the simulator is concerned this is a "sway" axis lateral acceleration.

Now this is a sustained lateral acceleration, and we already have our tilt coordination set up so that this lateral acceleration will be cued by rolling the platform left a bit.

That's so obvious after it's been pointed out! LMAO And following my example as the aircraft rolls past 90 degrees left toward fully inverted the lateral acceleration will decrease because a larger and larger portion of it is becoming a (negative) heave acceleration. Upon getting to 180 degrees (inverted) there is now no lateral acceleration at all.

So now rolls are no problem, but there might still be a conversation to be had about heave axis cues!

Ok for once I'm going to force myself to not give a lengthy explanation and get right to my conclusion:

We should have available a Tilt-Coordination channel for sustained heave axis accelerations.

Sustained positive heave accelerations (with 1G as the baseline instead of zero) would result in simulator pitch up, sustained negative accelerations would result in pitch down. I think we should have this option. If you don't want to use it, just set it to zero tilt coordination cue on this axis.

The main reason I'm saying this is that it should basically perform the same function for climbs/dives/loops as the roll cue does for sustained sway.

I'm not sure if you will all immediately agree or if I need to argue my case. =) I can demonstrate that this should be somewhat effective if anyone cares to argue!

The pilot should definitely be feeling forces throughout the loop, but they will not be extreme because one of the main forces involved is pushing down toward your feet like a barrel roll, this simulates "gravity" throughout the maneuver.

Here we go, time for a loop!

1. Aircraft is flying straight and level.
No significant cues.

2. Pilot pulls back the stick to initiate the loop.
High frequency pitch up cue for change in angular velocity. This cue washes out.
High frequency heave cue for change in vertical acceleration. This cue washes out.

3. Pilot continues the loop.
Sustained vertical acceleration continues, tilt coordination heave cue generates nose up pitch cue which does not wash out. This should feel good, and is consistent with real video of water in a glass during a loop. This cue will taper off without washout as the aircraft loses speed from the steep climb.

Once the aircraft reaches the top of the loop I'm not 100% certain how things will work out, but in theory the end result of the angular velocity change, sustained negative heave, and decrease/increase in speed will result in a "pitch up" pose throughout, but at certain points the pitch up cue will be somewhat diminished. I believe this should be about as close to the correct cues as we could hope for.

Note: I do not intend for a large pitch angle of the simulator platform during the loop, with all contributing forces some in opposite directions the platform should only wind up with a very modest pitch angle. Maybe 5-10 degrees nose up would be my guess.

So @pmvcda @hexpod @Dirty , what do you guys think? I very much want to be able to play with tilt coordination channel on the heave axis so I hope you guys include it! =D

Use pitch for sustained heave acceleration? I don't think it's correct.
Need to think more about it.
Surge and pitch yes, and we are already making it.
Now heave. If we have a constant 9.8m/s^2 down, it means, we would have a constant pitch down. Doesn't make sense to me.

EDIT:


Arrows show the "pressure" of the seat we feel on the body.

Heave in a loop (blue):
0º Heave is just 9.8 (gravity)
45º Heave force increasing
90º Heave at maximum while climbing
180º Inverted, we have the minimum heave
from there, heave increases.
We change heave in the rig to feel those changes, with a filter that would have washout so we regain range to more heave. I'm not sure, but I think that LP(HP) might be a good solution, where the HP should allow a good range of frequencies.
I think we need the filter in such way, that we have heave in the rig until the heave force becomes equal to gravity, maybe around 100º. Then the washout should make it's work to give us enough range to start a new heave when we reach around 250º.

Surge in loop (green):
0º We have no surge, we are at constant speed to start loop
From there, as we climb, surge changes as we loose speed (rig goes back so we feel deceleration)
From 180º down, surge changes as we gain speed (rig goes front, we feel pressure on the back as speed increases).
Now that's that deceleration/acceleration (in longitudinal direction - surge) that we can apply to pitch.
Idea here is to use rig surge for changes (high pass, but smoothed) and send low pass to rig pitch.

Conclusion, why there's people doing 360º rigs?
Only advantage is if you are doing steady inverted flight.
For our rigs, inverted steady flight is the rig just in neutral position. There are no accelerations, except gravity, and that after a small amount of time is already washed out.
Your cueing is done just visually in your VR headset.

Looking at if from a conventional/rational standpoint, I'd also say that the cues you are generating from those persistent vertical Gs are not cueing anything useful to the user. At east not when looking at things from a "pure" 6DOF point of view.

However, I certainly do encourage anyone to get creative and play around ...and I think FlyPT Mover will allow you to use any source you want to drive any axis you want. Try it out

On the other hand, I can assure you that in an aircraft you actually have this behaviour already "through the back door", so to speak. It is a side-effect of ordinary flight dynamics: When you pull Gs you will increase angle off attack --> longitudinal acceleration component --> Pitch up through tilt coordination.
That is of course not present in cars or rotary-wings, which is also correct.

What I read from your post on a somewhat deeper level is this:
"G-forces are such an essential part of what makes flying special, so it's a pity we don't cue them to the user to replicate the experience".
Yes, it's a pity indeed! There is this device that can actually generate persistent G loads and there is a similar device in Amsterdam. But both pay dearly for this ability with a rigorous limitation on your view angle. When they generate G-loads you have to hold your head perfectly still and look only straight ahead. Thats why those 20 million $ sims have a field of view of less than 50°. It does one thing, and one thing only: Cue Gs.

If you want to cue persistent G-load, harness tighteners or inflatable seat cushions might give some cues. There are quite a few who use them out there. I think @apointner and @SeatTime if I recall correctly. They seem to like it. I'd love to try it some day.

Dirty

I very much like your drawing, well done!

Thanks for the responses guys. @pmvcda

Here is where I see the problem. During the upward part of the loop, because we can not apply a sustained heave cue the pilot is basically going to be feeling the green arrow cue as the main force.

But in reality I'm almost certain the main force the pilot would feel in the climb is the blue arrow. Because of that I don't think we should accept that it is ok to not have any cue for the blue arrow during the climb (after the simulator heave travel has been used).

Now here is my basis for using pitch tilt for sustained heave:

Sustained upward heave acceleration "pushes you down into your seat".
Platform Pitch up will make you feel "pushed against seat" We pretty much ignore 1G straight down, but by tilting so that it's not straight down but into the "corner" of the chair I think it will be perceived as being pushed into the seat.

Sustained negative heave acceleration makes you "float up off the seat".
Platform Pitch down will make you feel like you are going to come out of the seat. (not perfect but better than no cue IMO). The green arrow is the direction the pilot will feel he is being pulled (because he is ignoring the slightly less than 1G now going straight down). Now note that with the platform tilted forward the green arrow (perceived force) is actually a somewhat upward angle, negative heave! Go ahead, tell me I'm a genius. =P

62990

Now in @pmvcda drawing, on the up side of the loop once the sim runs out of heave travel the cue is pitch down (because it is cue for negative surge, airplane slowing down), which will make the pilot feel somewhat weightless falling forward out of the seat , but this is incorrect overall effect caused by the missing sustained heave cue (the deceleration cue is actually an upward cue to some extent as per my green arrow above). The pilot should feel pushed down into the seat (from heave cue which is missing). So if we use pitch to push the pilot into the seat as in my drawing I think it will be better.

Even better, this cue will be additive to the cues at the end of the loop (or dive) when you are pulling out of the dive and G-force should be at it's maximum, enhancing that cue which is good!

What's more, I don't think my idea of tilt coordination for heave axis will cause false cues elsewhere. In a normal climb there is basically no sustained heave, so this cue will not occur and the pilot will simply feel the airplane decelerating (negative surge cue) as is proper.

In a banking turn the heave cue should only be momentary as the turn is initiated, so my proposed cue (which is for sustained heave only) should not occur.

You may find this interesting...

https://www.researchgate.net/public...otion_Simulator.pdf?origin=publication_detail

I find this interesting...
Adaptive filtering.... hmmm



Source:
https://www.researchgate.net/profil...hout-Algorithms.pdf?origin=publication_detail

An this one is maybe the most exhaustive and comprehensive:

https://www.researchgate.net/profil...ators-microform.pdf?origin=publication_detail

And so on into the night... ;-)

That does look interesting indeed.

But before we compxify we have to sort out the simple approach @Dirty is talking about

It's a matter of what you are trying to achieve. In general, normal people get more excited about a rig that turns you upside down and will pay more to try it!

Most of us are trying to get the most "correct" cues out of our motion sims, but I won't criticize someone for wanting to make their own sim to be more of a theme park experience. Also, if you are running it as a commercial business I'd say what's "correct" is what makes it most profitable, as long as you aren't advertising it as an authentic experience or training tool.

Yeah, it’s another application.

I don’t see any adventage of a 360 for General Aviation (Cessna, Airbus etc.). You never experience a sustained inverted flight.

Nevertheless for high performance aircrafts (acrobatics and combat stuff) it can be fun.

Can you explain what is missing in simtools so you can’t map your 360 deg. axis ?

I guess by realistic you mean close to reality.
I have tried 20 years ago a roll/pitch 360 and was not really convinced back than. For sure with an additional vertical axis it is for sure another level of experience.

To be clear, I am not putting anything on a scale of “what’s better” I will always applaud every creative approach of motion, so is yours.

What min/max settings did you used in tunning center for your eulers back then ? Are you sure you’ve explored all the possibilities?

I would give a try with the new 2.4. It should be out soon. Question of hours from now I guess...

In real life the forces felt in a moving aircraft can really fool the inner ear on what the aircraft is actually doing as we were not designed to fly. It is why early on many pilots drove their aircraft into the ground - especially at night when they could not see the real angle of the aircraft https://en.wikipedia.org/wiki/Sensory_illusions_in_aviation . Bit of a rabbit hole and we have been down this many times in the past. I have found for 'simulation', its best to use a bit of both rotation and G forces which is why I have ended up with a 6DOF with a G harness system. Eg in a turn I can roll the sim but will be held properly in place with the required forces by the G - Vest. A 360 degree sim is not required, but looks like fun to ride.

Just my thoughts from building and using sims for the last 6 years. It does not matter how realistic you may get your algorithms, our small sims cannot generate the required forces, so a good G system is paramount if you want to feel anything close to realism. Basically I rate G systems as the VR for the body.

Are we agree that in a coordinated flight up to 30 deg. You feel only the entry and exit accel and NOT the bank angle ??