my new motion sim with VR

[steve2112]

 

i just got this working today and i'm not used to flying in a motion sim so still getting used to it. anyhow, it works great, feels very realistic in oculus VR, the VR view is perfect, no artifacts from having the sensor mounted on the moving rig.

 

i plan to make a better video when i have time, showing how its built and how the software is configured.

 

it would be better if you could see my view also, i'll set up the monitor better next time.

 

this whole motion sys cost me about $500 not including racing seat.

 

UPDATE:

 

so i've had a bit of time to fly around and a couple of things i notice. first off, awesome! well worth the effort to build this thing.

 

i'm flying my vegas air race which i made a video about a while ago

 

you can see on the mosim video, when i go vertical climb, at some point the roll starts reacting strange. i'm sure this is DCS that's detecting roll in a near vertical is not very specific, but the motion sim doesn't know that, so it shakes me around, which isn't very realistic.

 

however, when i come over the top of the loop and pull a positive G to head straight down, it feels very realistic, so overall, pretty good.

 

the other thing is when i'm doing a tight banking turn, the mosim is banking (rolling) as its supposed to, but in real life of course, you wouldn't feel that, you would feel g forces pushing you down hard into the seat.

 

now obviously, i have no way to simulate 5 G of downward force, so i guess there is no answer for this problem.

 

just wonder if any setup advice? maybe instead of using roll in the axis assignment, i use something else?

 

and the other thing is, you know if you've been on a boat for a while and you get on dry land and you feel like you are still moving? we'll i'm sitting at my desk now and i feel like i'm still flying!

Edited November 20, 2016 by steve2112

[Einherjer]

Nice and easy setup! Good to see other DIY motion sims.

 

For rolling- what software are you using and what plugin?

In simtools you can use the "alternate DCS plugin" which gives you roll acceleration on the extra 2 output, so only movement as long as you are rolling ingame - steady bank angle is then a neutral mosim

[steve2112]

i'm using standard DCS pluging, about to try the other one now

[MacThai_75]

Looks great Steve!!

And your findings about it being able to use VR is exactly what I have been posting that I would expect, and aligns with what other people that actually have a motion rig and are flying VR with it, say.

 

WRT "getting shaken about" ... if you mean the smooth direction change you want the rig to make when going over the top without it trying to rapidly go in multiple directions at the same time ... I think somebody posted a fix for the simtools DCS plugin in one of these threads.

 

Great flying in Vegas, and I love the "past 11" quote .... just as long as that hard maneuver doesn't tap you on the spine. ;)

Edited November 20, 2016 by MacThai_75

[steve2112]

i'm trying to get the other plugin working, i installed it already but its not doing anything. i need to get on the xsimulator forum and ask why

 

but i started moding the exports.lua myself. that's the thing in DCS that sends flight data to the mosim. i already reduced the problem at high pitch angles, and i think there is are ways to make other improvements in that.

 

macthai, did you get your postal problems figured out yet? is your system running?

 

steve

[MacThai_75]

@steve2112 ... nope, my second box is still listed as being in NY USPS customs. :(

 

I'm told that I shouldn't lose hope, and that is will 'eventually' get delivered, but can take a long time (multi-months) in some cases.

 

I did find an email for shipment inquiries with customs, and emailed them for a status update, but I'm told by the local postmaster that they won't provide any info beyond the same that the tracking number reports.

The auto-response I did get said it can take up to 3 business weeks for them to reply, so .... we'll see.

 

In the meantime I'm living vicariously through you guys that have your motion rigs up and running. :)

 

P.S. It's good to read the you're getting the "over the top" in rolls and loops issue sorted out.

That's the big difference in flight and racing.

 

I'm sure it is solvable with some extra tweaking that will null the movement as the aircraft approaches vertical, and for some predetermined degrees of aircraft movement it will enforce a "vertical dead band" holding the rig at the travel limit in the last direction it was moving before the aircraft approached that vertical dead band.

Then it needs to keep looking for the aircraft's return to the seat's travel limit in the other direction(s) and once it 'approaches' that opposite travel limit within a preset amount of degrees, the seat should smoothly move in the maximum opposite direction(s) to intercept the aircraft's movement, and follow the rest of the roll/loop in the new direction(s) ... without getting confused and chucking the rider about as a roll or loop or worse ... a rolling loop sends it into fits as it goes past vertical or stays inverted while actually crossing back and forth through 'exact' vertical.

It will need to be smart enough to continue the max opposite limit, if the roll/loop is stopped after crossing vertical and then the pilot returns to vertical going back towards the original direction, so it just doesn't exhibit the same "vertical issue" but at the travel limit after it has imposed the dead band routine.

 

Whatever patch is used, it will almost certainly need to know what your particular rig's travel limit (in rotational degrees) is for both pitch and roll, for it to be optimal.

 

As far as sensing G's ... be sure to read up on the G-Seat.

When my rig gets here I plan to build (or modify my existing seat) by adding the G-Seat's G functions 'only' ... and use it for the seat on the motion rig for added realism.

Edited November 21, 2016 by MacThai_75

[Einherjer]

The snapping with high pitch is done with the alternative plugin

For rolling: as said above , this plugin has the extra 2 feature which sends you a banking for the platform just as long as you roll , no matter how far (works with smoothing in simtools)

 

And it has the normal bank feature- so your platform will follow your Planes movement- but by default will also snap to the opposite direction when you reach 180°

 

It will be fixed in a future release- until that you can use this workaround:

 

-----------------

 

SimTools, Cockpit Motion Control Discussion Thread

 

Hi all,

 

 

 

If you have a seat-mover (I don't think it's an issue with a G-Seat) or have browsed the DCS plugin threads on X-Sim or SimTools, then you might be familiar with the roll bug.

 

 

 

The issue: At some point while rolling left, the motion platform will be pegged at full left deflection. Then just after crossing 180 while inverted, the motion platform will jerk to full right deflection to complete the roll to 0 degrees. Very distracting, not at all realistic.

 

 

 

RiftFlyer has worked very hard and proposed an interesting solution that calculates bank acceleration (as far as I can tell). However, the Alternative DCS Plugin RiftFlyer authored resulted in the motion platform violently/randomly gyrating along its roll axis (I think it may have been related to not converting degrees back to radians before outputting to the socket. I had the same result / issue at first).

 

 

 

Anyway, I wired up a fix tonight that I'm very pleased with. You can get a sense of the problem / solution approach in this JSBin: https://jsbin.com/qohelupovi/edit?js,console

 

 

 

Here's basically what it does:

 

 

 

 

 

Input

 

0

 

45 -45

 

90 -90

 

135 -135

 

180

 

 

 

 

 

Desired output

 

0

 

45 -45

 

90 -90

 

45 -45

 

0

 

 

 

 

 

I got it working in export.lua tonight:

 

 

 

 

 

function rad(deg)

 

return deg*(math.pi/180)

 

end

 

 

 

function deg(rad)

 

return rad/(math.pi/180)

 

end

 

 

 

 

 

local synRoll = deg(bank)

 

 

 

if synRoll < -90 then

 

synRoll = rad(-(synRoll + 180))

 

elseif synRoll > 90 then

 

synRoll = rad(90 - (synRoll - 90))

 

else

 

synRoll = bank

 

end

 

 

 

 

 

I certainly hope this can help some folks with seat movers who are experiencing the roll bug. Happy to answer any questions related to this.

 

https://r.tapatalk.com/shareLink?url=https%3A%2F%2Fforums%2Eeagle%2Eru%2Fshowthread%2Ephp%3Fp%3D2912408&share_tid=173966&share_fid=74365&share_type=t&share_pid=2912408

 

 

-----------------

 

Works great

[steve2112]

T

 

 

Here's basically what it does:

 

 

 

 

 

Input

 

0

 

45 -45

 

90 -90

 

135 -135

 

180

 

 

 

 

 

Desired output

 

0

 

45 -45

 

90 -90

 

45 -45

 

0

 

 

 

 

 

I got it working in export.lua tonight:

 

 

 

 

 

function rad(deg)

 

return deg*(math.pi/180)

 

end

 

 

 

function deg(rad)

 

return rad/(math.pi/180)

 

end

 

 

 

 

 

local synRoll = deg(bank)

 

 

 

if synRoll < -90 then

 

synRoll = rad(-(synRoll + 180))

 

elseif synRoll > 90 then

 

synRoll = rad(90 - (synRoll - 90))

 

else

 

synRoll = bank

 

end

 

 

i was thinking exactly the same thing.

 

that is exactly what i'm in the process of implementing now. when i get it working and make a new video to show it in action

 

here's my roll code, its working very well

-- returns roll back to zero as it passes +/- 90 degrees to avoing the -179 to 179 flip that occures in a full roll

if bank > 90/57.295779 then

bank = 180/57.295779 - bank

dbg_file:write(string.format("FIX +ROLL: p=%.3f r=%.3f ", pitch, bank))

elseif bank < -90/57.295779 then

bank = -(180/57.295779 + bank)

dbg_file:write(string.format("FIX -ROLL: p=%.3f r=%.3f ", pitch, bank))

end

 

steve

Edited November 21, 2016 by steve2112

[steve2112]

hi everybody, i've fixed all the jumping and jerking around problems at extreme pitch and roll (for DCS)

 

if you want to try it, install the standard DCS plugin, not the one from riftflyer, then go to your DCS scripts directory, mine is at C:\Users\me\Saved Games\DCS.openalpha\Scripts

 

exit your export.lua and change the body of AfterNextFrame to this

 

*******************************

AfterNextFrame=function(self)

 

local altRad = LoGetAltitudeAboveGroundLevel()

local pitch, roll, yaw = LoGetADIPitchBankYaw()

local accel = LoGetAccelerationUnits()

 

-- convert to degrees to make the math easier

pitch = pitch * 57.295779;

roll = roll * 57.295779;

 

--dbg_file:write(string.format( "p=%.1f r=%.1f ", pitch, roll))

 

-- STEVE reduce roll by when pitch > 45deg

if pitch > 45 then

local multiplier = 1- (pitch-45)/30

if multiplier < 0 then

multiplier = 0;

end

roll = roll*multiplier

--dbg_file:write(string.format("PITCH>45: p=%.1f r=%.1f m=%.3f ", pitch, roll, multiplier))

end

 

-- STEVE returns roll back to zero as it passes +/- 90 degrees to avoiding the -179 to 179 flip that occurs in a full roll

if roll > 90 then

roll = 180 - roll

--dbg_file:write(string.format("ROLL>90: p=%.1f r=%.1f ", pitch, roll))

elseif roll < -90 then

roll = -(180 + roll)

--dbg_file:write(string.format("ROLL<-90: p=%.1f r=%.1f ", pitch, roll))

end

 

--dbg_file:write(string.format("\n"))

 

-- reduce forces on ground

if altRad < 3 then

accel.x = accel.x * 0.25

accel.y = accel.y * 0.25

accel.z = accel.z * 0.25

end

 

my_send = socket.protect(function()

if c then

-- socket.try(c1:send(string.format("%.3f; %.3f; %.3f; %.3f; %.3f; %.3f;\n", pitch*1000.0, bank*1000.0, yaw*1000.0, accel.x*1000.0, accel.y*1000.0, accel.z*1000.0)))

socket.try(c:send(string.format("%.4f; %.4f; %.4f; %.4f; %.4f; %.4f;\n", pitch/57.295779, roll/57.295779, yaw, accel.x, accel.y, accel.z)))

 

end

end)

my_send()

 

end,

 

*******************************

 

if you want to run the debug printouts, add this to the Start function

dbg_file = io.open(lfs.writedir()..[[Logs\dbg.txt]], "w")

dbg_file:write(string.format("start ok\n"))

 

so what this does, is when pitch goes over 45 degrees it attenuates the roll factor, which as you can seem from the log file below, goes a bit nuts when pitch is over 60+ degrees. this makes loops much smoother

 

then the next bit fixes the flip that occurs when doing a 360 degree rolls. when roll goes past +/- 90 deg, it rolls it back to zero at 180, so there is a smooth transition form + to - roll numbers

 

with a VR headset on, both fixes feel quite natural. i've done about 40 hours of aerobatic flying in real life so i have an idea how it feels

 

enjoy

 

steve

dbg.txt

[MacThai_75]

^^^^ yup, that's the stuff!! :)

Both roll 'and' pitch needed that fix.

 

Now if the damn second half of my DKP motion rig would get here, I'd be a happy camper.

 

I had meant to also mention that I was expecting you to do a clone of the DKP, but I see you went in a entirely different direction in your implementation of the components.

When you get a chance ... I think more details of your rig would be of great interest to many here.

 

Glad you got it working, and I've DL'd your pitch roll fix.

[steve2112]

yes, i felt from day 1 that having a big range, especially on pitch was important. all the sims i've seen on the tube seem to have pitch +/- about 5-10 degrees at the most. i have about +/-25deg usable right now and it feels great. when you are in a near vertical climb/dive you really feel it. i'd love to go more.

 

i know some of the people on xsim were saying don't bother with that, but i never believe the conventional wisdom, i always like to check stuff out for myself and in this case i'm sure i was right. i also think i was right about the 6:1 gear system. its simple math really, that motor we are all using has far too high a shaft speed to drive the actuators directly, but some know it alls on xsim like to make out that i'm "trying to get something for nothing" by gearing the shaft speed down to a more practical value, when in fact what i've done is matched the speed and torque requirements of the system to the motor. by not gearing, people are overloading the motor and effectively getting a lower speed by stressing the motor.

 

also, roll angle isn't so important i think. more than 25 degrees and you are starting to fall out of the seat, so i'm happy to stay where i am on that, but i am thinking about ways to get even more pitch.

 

i plan to do a detailed video on both the mechanics and the software setup

 

here it is with the roll and pitch fixes i added

no more jerky movement at extreme pitch and roll angles, and now i can increase the range of motion safely

 

 

steve

Edited November 21, 2016 by steve2112

[MacThai_75]

Steve, I think most of the guys on xsim are racers and don't understand, or are not concerned with the special requirements of a flight sim Vs a racing sim when it comes to the extra travel ... even on a course that has something like the corkscrew at Laguna Seca.

 

But irrespective of why the discounted your ideas, I'm glad to see you stuck to your guns (pun intended :smilewink: ) and got it built they way you wanted ... and then dialed it in to work that well. :thumbup:

 

I am interested to see what happens if you fly inverted for a short course, and do a spiraling climb into an immelman with a long inverted delay before you roll upright again on top.

 

I think that will be the real test of the plugin.

[steve2112]

flying inverted will be exactly like flying level the way i have it set as anything close to roll=180 will generate a roll of close to zero. i don't see any way around this other than one of those full 360deg 2DOF rigs which look awesome but a major building commitment, and not something i have the space or time for.

 

i agree about car racing sims, a few degrees is probably enough and they are concerned about surge (acceleration/braking) and sway (side ways acceleration). except for going up and down steep hills, pitch isn't important for them and a car can't climb a vertical hill, so yes you are right, my only real interest is aerobatic flight sims. even a multi-million $ commercial flight mosim can't pitch and roll like my system can.

 

as many people have said, its almost impossible to truly simulate the real feeling of a aerobatic flight. the only thing that comes close is that huge arm thing in an aircraft hanger size room. so give my constraints, i think my fun/$ ratio is quite high

Edited November 22, 2016 by steve2112

[MacThai_75]

Yeah ... and unlike a 360º 6DoF rig .... no need for a barf bag! ;)

 

IMHO the most important test is the one were you ask yourself if it's fun irrespective of the price you paid.

You've definitely answered that question, and put it's low cost onto that answer ... like icing on a cake. :)

 

Once again, thanks for the file!

I'm sure I won't be the only one happy to have it. :thumbup:

[Einherjer]

Nice to see it running!

I tested my platform on DCS and dirt rally

 

Can confirm that a racing sim needs just a few degrees to feel good (full movement doesn't feel right here) but for flying you need more! Mine is about 15° to each side and it feels good at all !

Pulling up after a dive is a great sensation with traveling the full 30° of pitching plus the oculus

[steve2112]

i just added the .lua file here

this is a major improvement over the standard DCS plugin

Export.lua

[MacThai_75]

^^^ :thumbup:

 

Thanks Steve! :)

[steve2112]

here is the description of how i built it and set it up

 

its working great now so time for some flying, until my next project!

 

steve

[MacThai_75]

Steve, great viddy and very informative.

If I have my dates right about when you were just starting this project, you got that all accomplished in under 2 months, and actually closer to only one month!

Very well done. :)

 

WRT:

The mouse ... do you think a cheap track ball (if only for set-up) would work as well as the mouse wheel being added to the program?

 

The spring noise ... it might be eliminated by using a silicon rubber tube either around or inside of the coils ... that's cut a bit longer than their most stretched length.

 

Camera aim changing ... although you don't seem to have this problem, I've read that on race sims that the CV1 camera can be jostled enough to cause it to point to a different place.

Some guys even report using duct tape to hold it's pointing.

But I would suspect if it even 'is' a problem (possibly only appearing over time on a smoother moving flight rig, as the it's friction might become more loose) ... using a piece of foam pipe insulation around the camera's mount stem, with it's length cut longer then the stem so it gets compressed between the camera's body and stem's base ... that should prevent it from changing it's pointing, and keep you at the same level of view in the pit.

If your particular camera's tilt mount is stiff enough not to allow the camera to change it's aim during rough .... turbulence :smilewink: ... then this would not be needed.

 

Increased angle ... can you just buy another double flanged U joint and bolt their flanges together to sort of make a joint like this 80º "high angle" CV joint?

 

 

Thanks for posting the viddy!

I'm sure it will be as interesting and informative to to others as it is to me. :)

Edited November 24, 2016 by MacThai_75

[steve2112]

great ideas, thanks

 

most of the time build was waiting for china post (i know you know how that feels!). i would say the mechanical work was about 4 days, the electronics, 1/2 a day and software setup and tuning, 2 days. i should add that info to the video.

 

of course, had i seen my video before had, i could have cut that in half

 

steve

[MacThai_75]

Wow, that's much faster work than I could ever do! :)

 

If my second box never gets here, I'll just buy the motors and electronics, and follow your lead.

 

The box that did get here has most of the base frame, and I never really planed on using much of the upper frame, preferring to modify my seat (to a G-Seat) and use the control mounting system from my stationary rig.

 

So if my second box is with Elvis :smilewink: ... I'll use the $600 DK refunded me (out of my $1245 deposit) and buy what's needed.

 

I also plan to add a third axis for yaw, by building a T shaped base with a pivot at the bottom of the T, and add casters to my rig so it's rear section can float across the top of the T.

 

Then with my entire rig sitting on the T with the pivot point at the front, the back will then swing side to side across the T, driven by a lead-screw motor mounted between my rig and the T, for Yaw movement.

 

Sort of like a home built version of this $1300 accessory for the SimX 2 DoF rig ...

Edited November 24, 2016 by MacThai_75

[Einherjer]

my new motion sim with VR

 

Two months for a motion sim is quite realistic, mine was ready in under two months too - the longest time takes the tuning of the movement for different games.

After Dcs i added a plugin for dirt rally and needed to tune everything for it , it requires way smaller movements than DCS .

 

 

MacThai , good luck with your build! What i can see from the forums is , that the DIY platforms are often better and more accurate than some expensive commercial products

[MacThai_75]

Thanks Einherjer!

 

There is still a chance that my second box might arrive, but it's looking more like it'is lost in the USPS/Customs system. :(

 

But not to worry, if it never makes it I'll start on my home built rig after the holidays (that will employ the parts I 'did' receive along with some of my stationary rig) and like you say ... probably end up with a more satisfying end results. :)

 

I really think the incorporation of the 3'rd yaw axis, and adding only the G force components of a G-Seat mounted on the 3DoF rig, will be the best of all worlds in a compact rig, that along with the motion sensation of a VR HMD ... will provide near 6DoF immersion.

[steve2112]

Thanks Einherjer!

 

I really think the incorporation of the 3'rd yaw axis, and adding only the G force components of a G-Seat mounted on the 3DoF rig, will be the best of all worlds in a compact rig, that along with the motion sensation of a VR HMD ... will provide near 6DoF immersion.

 

does that G seat thing really work? i mean feel like real G forces?

[MacThai_75]

From what I've read it should be about as good at synthesizing the feel of pulling G-Forces as a non-360º travel motion rig with VR is at synthesizing full roll and pitch ... if that makes sense.

 

But there are people this forum that have built the G-Seat ... and more on it and those people can be found here ...

https://forums.eagle.ru/showthread.php?t=122577&highlight=Bergison+gseat (which has more links in the thread).

 

Here's a viddy of the creator with it ...

 

But I don't think if I mount one on a motion rig (by modding my existing seat with single "push/pull plates"(instead of flaps) in the seat bottom 'and' in the seat seat back) that it would require the all of the G-Seat functions.

 

Just the motorized seat belt snugging and relaxing system, with the single seat push/pull plate in the seat bottom and back being motor driven in both directions (up/down and back/forth) all working in concert, should be enough to synth both positive and negative Gs well enough to pull it all off, as well as to add what's needed to feel some heave ... with the motion rig providing pitch, roll, and yaw, and the VR's semi-motion sensation filling in the gaps. :)

Edited November 25, 2016 by MacThai_75