Trip's G-seat project (primarily for flight)

Alright, I've got it all working. You guys know what that means right? Time to decide it's no good and needs a redesign. LOL

I'm worried about the tapping noises when the paddles come back down against the rollers. I'm going to see if I can make them stay in contact at all times via spring tension but this could turn into another engineering nightmare. Back to the drawings I go.

Update: Modifying the roller assemblies to stay up against the paddles at all times was as easy as I expected but I'm feeling like things are getting unnecessarily complicated again after adding yet another eight springs and four sliding assemblies. The more of that kind of thing I build the more likely I am to have trouble with it. Time to attempt a K.I.S.S. rendition. I'm enjoying working on this and totally redesigning it repeatedly has been a great way to learn how to use Fusion 360, but I need to get it done so I can focus on other less fun but more important tasks!

What's been causing a lot of the complexity here is something that someone who knows 3DOF well might have a better solution to.

My seat is going to be rolling left and right for sway cues. If the seat is rolled to one side and I get a heave cue, the heave cue isn't going to be straight up, it's going to be slanted to the side I'm rolled over to.

I decided that was wrong and that is what lead me down the path I've been on thus far. Well, that logic was flawed. I'm accustomed to thinking in terms of MOTION where the heave axis in a G-seat is mostly about pressure. Add that to the physics of coordinated turns and such in aircraft and I'm now quite confident that this is a non-issue. If I can make the heave axis relative to the seat instead of to the floor all of this becomes a hell of a lot simpler.

exactly the point! and that is a good way to prevent parasitic movements.
In a 2DOF, you have a movement followed by a resulting persistent angle for a while. Movement and angle are contradictory in terms of simulation feedback.

Gseat could provide gforces without the drawback of contradictory informations.

Keep going!

This is obviously very unfinished, but before bed I wanted to post the new concept to see if you guys have any thoughts.

This shows only the movement I plan to achieve with the panels, not the mechanicals for how to make it happen but I'm pretty sure I know how to do it.

This is a tricky thing to figure out, I understand why SeatTime has been at it for years!

Thank you for bringing my mind back to this Mat. I really think I should focus more on minimizing that contradiction in the roll axis with the body loading stuff.

I'm now thinking about the ACME system again, where they move the bottom cushion in the direction of the actual roll , but the seat back moves your body the opposite way for the G-force. Combine that with body loading also moving you in the direction of the G-force and that might be the way to go. Damn! Back to the Acme system seeming to be the best way to go again! Just when I start going back toward SeatTime and Bergison methods...... Also let's not forget that SeatTime was talking about the bottom paddle sway cue moving the driver a bit in the wrong direction and that he may drastically reduce, eliminate, or even reverse it. That's exactly what we are talking about here. So now I'm going to look at going full Acme, but with body loading and larger movements. Maybe I'll keep real pitch in the mix too.

I really want snap rolls to be as good as possible which means getting a violent reaction with minimal contradictory/false cues.

Another thing... based on the reading I did about vestibular stimulation tricking the brain into thinking there was a big roll axis movement is the one thing it's apparently very easy to do. Doing anything else is a very difficult task, but it so happens that with the heave resolved by the eleveator, roll is the one thing really bugging me. I think experiments may be in order, but maybe not until after the G-seat is built.

Yep, that's the idea .

Well, it's almost done. Again. Pretty satisfied this time. I've been up for over 24 hours so I probably won't have anything to post until tomorrow. Basically I have to fix one joint that broke, make the linkages to sway the back cushion and roll the bottom cushion, and place the already made roller assemblies in the right spots. This one is quite simple in design so it should be pretty quick to get the motion studies done and then animations are just a couple mouse clicks away. I was glad to hear at least a couple of folks are enjoying the animations. I know I'm kinda going a little crazy with it it's my shiny new toy. It really is a whole lot better than trying to explain or draw diagrams though.

At this point it's basically an ACME DMS reduced to 3DOF plus SeatTime style heave.

Ok, so I've got the bottom part done and this time I'm satisfied. It should be pretty quiet and effective. I even got the heave actuator installed!

PS- There will be linear guide rails keeping the heave assembly straight up and down so in theory there should not be any load on the heave actuator except straight up and down. Worrying about this a bit just reminded me that the original plan was to use a 2nd class lever the same basic way SeatTime did. I'll change it next chance I get to work on this which may not be tomorrow. Why I didn't do that in this drawing I don't know. Possibly lack of sleep... I'm up to 36 hours awake now. Time to watch the last two episodes of High Castle and go to bed!

The seat back isn't put together at all yet but it should be pretty easy. All it's getting is heave coming right off the same actuator that's there, surge from the frame behind, and sway via Bowden cables most likely.



I can't get the nice animation rendering to work with this version. I keep getting an error, which is most likely the result of the huge number of bits of erased crap and unattached sketches that have been piling up from each mistake I make. I'm getting a lot better though, last time I started a sketch from scratch I was getting very few. This is not that drawing though, this is an edit of the one I've been working on for days.

Just out of curiosity I put in an inquiry about the cost of an ACME DMS G-seat since their advertising is loaded with "cost effective" claims. They indicated that a minimum budget of at least $50,000 for the G-seat is appropriate! D=

I had guessed $20,ooo to $30,000 and thought that was crazy. And I thought $3,000 or so for a simple four paddle SimXperience G-seat was bad! Speaking of which, in verifying that I remembered the price correctly I just happened upon this on their site.... they called their G-seat "6DOF" and followed it with this image.

LMAO Apparently "at rest" is a DOF, left and right is two DOF's so we could build a three DOF with one motor! We've been doing it all wrong!

yeah, it's stupid marketing shit calling this seat a "6DOF"...

it's a nice 3DOF gseat though

I agree it's probably about the best option for someone who is not a DIY type and build quality looks nice.

Hi guys. I'm looking forward to digging back in on this project now that December is behind us.

I started over from scratch again on the drawing and I've got the bottom portion done (mechanically). Hopefully I'll have another 3D model to show you within a week or so and then it will be time to start actually building it.

More pictures soon!

Ok so I decided to get the elevator more or less figured out before I go any farther with the G-seat.

I suddenly thought about the fact that my load of 317 kg is going to be resting on, and shoved upward by two little ballscrew nuts! I figured I'd better check and make sure that's not going to break the bearings! If I'm looking at this right, the dynamic load capacity per 1610 ballscrew is 710 KGF. So 1420 KGF total. That's a heck of a lot but I've been warned many times how dynamic loads can get astronomically huge with accelerations. (In retrospect, I should have realized how little acceleration we really use on motion sims.)

Edit: I found a calculator, I just had to up google-fu game. That made me realize that my one figure was speed... er velocity, not acceleration, so I can do this! I think...

So yeah, problem is I don't know how long it takes to accelerate my load to 417 mm/s. More maths. Bah.

Ok so lets say it takes two seconds to accelerate my load from stopped to the max of 417 mm/s.. finally I've got all the correct numbers to plug into everyone's favorite equation!

OK so if I did this right I've got my answers.

0 - max in two seconds= 326.06 KGF

9.8 m/s/s + .208 m/s/s divided by 9.8 m/s/s = 1.021 times 317 kg load = 326.6

0 - max in half of one second= 343.9 KGF



I've spent the last thirty minutes trying to (re)learn how to calculate a dynamic load, but it's very difficult since I don't remember even the most basic things from high school 22-26 years ago.

I've got my acceleration at 0.416 m/s (416 mm/s)
and my load at
317 KG

I need the answer in KGF (Kilogram Force)

I know I multiply them, but I'm not at all confident I'm doing this right. I think the .416 m/s is actually .416 m/s squared, is that correct?

317 kg x .416 m/s = ~132 KGF?

That can't be right! I figured I'd show my "work" so you guys know I'm not just being lazy about this.

I'm just trying to make sure that just two 1610 ballscrews/nuts can handle pushing the weight of my platform up at 416 mm/s. I was afraid the bearings might break! I imagine they can, but I want to be sure. If they can't I'll need to add spring assistance I guess.

Thanks Noorbeast. I'm confused by that result though, I thought it would be way higher. If you feel like it, show me the numbers you used so I can try to follow. I'm still searching the internet and trying to force my brain to work LOL.

I got 326.06 if it takes 2 seconds to reach max velocity. Still seems too low. I guess maybe two seconds is much too long. Or I've still got the whole thing scrambled!

9.8 m/s/s + .208 m/s/s divided by 9.8 m/s/s = 1.021 times 317 kg load = 326.6

Same math saying it only takes 0.5 seconds to reach max speed gave me = 343.9

If I'm doing this right that means accelerating to .416 m/s in half of a second is only 1.2 G's? That seems low but.. well now looking at it well under half a meter per second isn't very fast. About 1.5 kph. That seems reasonable. I think I got it!

Hey guys. I made a wall of text post explaining the details of why a visually enclosed cockpit probably won't work for most people, and how it will work for a sim like the one I'm building in this thread.

Most/all folks will want to just read the TL;DR =)
https://www.xsimulator.net/communit...n-compensation-and-enclosed-simulators.14150/

Small update, though the only progress recently has been made in my head. =)

WMR Enclosed Cockpit Tracking Test

I got a Reverb and tried my enclosed cockpit idea and the tracking didn't seem to be able to work at all in a small enclosed space. This was definitely not expected!

When the cockpit is done I might have to do a real test as it will have just a little little bit more space in some dimensions than the closet I tested in. Mostly, it might be able to track based on looking at my feet LOL. I put black duct tape X's on the walls for it to use for reference and made sure there was enough light but it seems like maybe it can only focus in on things that are more than 3 feet (915 mm) from the cameras. Sent back the Reverb much to the relief of my conscience.

Interestingly when I tried the same test with Oculus Quest (but less thoroughly so more subject to false results) it did appear to work. It's widely known Oculus has by far the best tracking of that type so it makes sense.

No way I'm going to Oculus though, at least not this generation. I now know that I definitely have to build with lighthouse tracking in mind.

Simulator Design Evolution

The biggest news is that I've decided to add a small surge and sway table to my design! Approximately -4/+4 inches for a total of eight inches on each axis. That's -100/+100 mm total 200 mm each axis for people on the logic system. This will make much better use of the servos I have that are massive overkill for the G-seat.

The surge and sway axes will simply be hooked up to the same actuators running those axes on the G-seat by linkage of some kind. This will give me an additional way to really sell the onset cues for snap rolls and carrier launches and landings etc.

I've also decided to go back to running the G-seat bottom cushion roll on the sway actuator. I feel more comfortable having the three motors powering the heave elevator instead of just two. Two can handle the load, but I was worried about the load being imbalanced. I think I'm fairly likely to add a seventh channel for that so I can adjust "roll" separately from sway in the software but I'll try it without that feature first.

Also, I'm adding a description of the mostly finalized design to the original post here. It's mostly copied from the conversation in that other thread.

Edit: Trying to fall asleep I remembered a couple inconvenient facts about the decision to switch back to three actuators for the heave elevator. First, I only ordered two of the 1500 mm ballscrews so I'd have to order a third. That's another $75 or more. Second, I remembered the specific reason I really wanted seat bottom roll tilt separate from the seat back sway. For sway cue I want the seat back sway and seat bottom roll to move together (for knife edge flight, tilting on the ground) and for roll rate I want them to move opposite, the bottom rolls the direction of the actual roll and seat back sways opposite to create the lateral g-force from the sudden change in rotational velocity. I edited the OP again to correct that, and also to add the active harness tensioner system which I had forgotten to put in there.

I either have to go back to only two heave actuators or commit to adding another actuator system to go to more than six actuators. I'll try the former first. If I have problems then I'll get another controller (Thanos) and at least one more servo and ballscrew. That's a lot of extra expense. If I do that I almost might as well get dedicated servos for the surge/sway table too which is even more expense. Down the slippery slope I slide. Again. =D

I think I'll be ok with just the six.

So up until now I've thought the VR "scent module" concept was a bit silly. I think we've got a very long way to go with VR before we get to the point where this would be a priority.

Now @J-1775 mentioned to me the scent of a real aircraft cockpit and now I changed my mind! =P Joking aside, it would be fantastic to capture that particular smell. I think the only effective way would be an enclosed real cockpit. Anyone know where there's an abandoned Huey? I need a bigger budget!

It's a little premature, but here is a little video of the idea for the platform. Now I just have to figure out how to mount the actuators so they can operate the surge/sway table and the G-seat both without interference and without getting in the way of G-seat heave. =/

Hey, for once it was easier than I expected to figure something out!

I think all I really have to do now before I can start building is finish the G-seat drawing to make sure there isn't anything I'm overlooking.

I might be cutting steel within a week or so. =)

And before that you may need to pile bricks for the hangar.

This seems to become a considerable piece of work and I wish you all the best for this project.