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

Tonight I continued design work on my G-seat.

I decided I'd been thinking about one of the cues incorrectly and most of my work up until this point was a wasted effort and I was slowly working my way back to a build near identical to Bergison's. I worked on drawing a different setup for several hours and working that drawing out made me decide I was correct in the first place, I had just been taking it a tiny bit too far.

Have to sleep now, but I decided to make some notes before bed so I didn't forget the thoughts I had, and it turned into a very productive brain storming session. Funny how much writing things down helps me figure them out. I'm happy to say I think my setup will be at least a little bit unique. =)

good project! lot of thinking
I'm intrigued: what is the animated picture in post #19? where did you find this?

Thanks =)

The GIF in post 19 is a drawing of my design for the mechanism for tilting the bottom seat cushion left and right like the ACME DMS while still having the heave functionality I created that was inspired by SeatTime's designs. I drew it with Fusion 360 and animated it with the same. It's my first project making a 3D drawing with moving parts. =)

I'm still working on that drawing to build the full seat. I still have to finalize a design for the mechanism moving the seat back cushion and how to mount and connect the actuators to everything. It's driving me crazy with so many parts but this is a lot better than wasting materials figuring it out! I've been able to try a lot of design concepts with 3D drawings which allowed me to animate them and see if the design works etc. =) It's a fantastic tool, kudos to Autodesk for allowing hobbyists to use it for free!

great job! it looks so realistic

one point in the design is to prevent play and backlash in the mechanism

Looking at your mechanism for moving the paddles at the same time in the z direction made me think of this mechanism. If you can get enough lift out of it you might be able to have a more compact setup for the same motion. Just thought I would bring it to your attention as I just found it myself and thought it was pretty cool.

@cfischer : looks good!
but what are the specs?
can you share a link please?

I mainly meant to share the concept so someone could make it themselves with the motors and angles that suit their project. I haven't even looked at the specs on that actuator. I found it through google image searches.

https://www.google.com/search?safe=active&rlz=1C1CHBF_enUS876US876&tbm=isch&sxsrf=ACYBGNT80O2uSw3Co71BWXe-fjR_U_RgsQ:1575520668160&sa=1&ei=nInoXfGuCYi0tQXV4JTIDA&q=wedge+lift+stage&oq=wedge+lift+stage&gs_l=img.3..35i39.56649.61403..61543...2.0..0.66.1097.18......0....1..gws-wiz-img.......0i67j0j0i8i30j0i24j0i30.4ugcapklM-Q&ved=0ahUKEwjx7Ivu153mAhUIWq0KHVUwBckQ4dUDCAc&uact=5

One could make this pretty easily with some cheap mgn12 bearings and a ball screw from ebay. Slap a 80ST servo motor on there and I bet it could lift your car tire pretty quickly.

thanks for the hint with the terms "wedge lift stage" for a non native speaker like me!

that's a pity that the speed is 12mm/sec under a 20 kg load
it's really slow but as I read in the specs, the main goal of this lift is micrometric positionning

Once again, the more I work on the design and experiment the more I find myself moving back toward a design almost the same as either Bergison's or ACME's.

This is going to be a bit of a rambling thought process so skip it, or strap in!

I completed a design layout with all the features I'd decided to combine and with the animated drawing decided it was too convoluted, with too many redundant features. I think this is largely because of my decision to build the seat using only three actuators. If I can't move all these things independently there's a lot less advantage to having them.

As for the two reference designs, I'm asking myself "why" the ACME G-seat is designed the way that it is. I think this is a good question since I suspect that a very large amount of money was spent on R&D in developing this system. The main conclusion I come to is that the design loses all it's benefits when you cut out the ability to operate the different features independently, so that sometimes two surfaces will move together and other times they move opposite.

So, how best to achieve similar results with only three actuators.

Design goals:

1: Ability to feel the angle of a helicopter during pick-up and landing. Picking up you want to feel with great accuracy the direction the helicopter is tilting so you can compensate. This requires a pure "roll" cue, with the seat bottom and back rolling the same way and very minimal movement of the pilot within the seat.

2: Best possible cues for sudden roll maneuvers in high performance aerobatic and combat aircraft.
This requires the pilot feel violently thrown in the seat to the opposite direction of the maneuver. I want to feel myself being moved in the seat for this. That requires the bottom and back move opposite directions, or given my three actuator limit I think I can achieve this by rolling the bottom of the seat but making the back cushion stationary.

3: Simulate pitch angle on the ground as with roll, and simulate pitch rotational rate changes in the air with an actual pitch motion.

4: Simulate large forward accelerations positive and negative including aircraft carrier steam catapult launches and arrestor wire landings. This requires very fast and violent cues.

To achieve both goals for the roll/sway axis I think I will have the rear cushion either roll with the bottom or just the cushion sway with the bottom roll (like ACME) up until a given point. Beyond that point the bottom continues to tilt but the back either doesn't move any farther or moves the opposite direction. The first part designed for subtle roll cues on the ground, the rest for bigger lateral-G cues. This is a problem to think hard about.

For the pitch/surge axis I think things are a little bit simpler and that Bergison's system will do very well by itself without the need for any significant modification. For a catapult launch the seat violently tilts back, moving the pilot's eyepoint in the proper direction and the back paddles push forward as with SeatTime's designs to increase perceived force of the body being pressed back into the cushion. I've found the mechanics of this can be quickly complicated by complex designs intended to solve the roll problem described above.

I think that Bergison's design of rolling the entire seat as one (and applying pressure with paddles) will provide the best cue for violent lateral G maneuvers and is also perfect for the subtle roll angle on the ground.

I'm realizing more and more that Bergison's design is damn good. Really the only thing I feel it's missing is the feeling of friction of your body moving across the cushion since the cushion is moving with you instead of the reality where you are moving against it.

To solve that I'm thinking about artificially creating a feeling of friction as the pilot's back moves laterally.

More thoughts developing out of that:

I'm going with the fighter jet here, since it's the tougher thing to simulate.

If the jet rolling left violently is simulated by the seat rolling right violently the initial onset forces are reversed. This is the idea behind the ACME design. With that design the seat bottom moves with the aircraft, and the seat back moves you the opposite direction. A violent left roll has the bottom roll left and the back sway right.

That is similar to what SeatTime was recently talking about with the bottom paddles in his G-seat actually moving the driver in the wrong direction during sway cues. He said that he may drastically reduce that paddle moving up, or even consider making it go the opposite way. The ACME seat actually does have it go the opposite way. I'm not sure about it feeling right after the initial onset, but again we are talking about a product the military uses heavily that is the result of many years of development and R&D.

IDEA!: Hydraulic damper on the bottom seat cushion causes the bottom cushion to delay the "roll" motion, giving a better onset cue, but then a spring brings the cushion over to the "outside" of the roll for the sustained lateral G-force cue. Brilliant? Maybe!

It would be better to have the bottom cushion actually move opposite initially and then go over... could I make that happen with only one actuator, and have it not interfere with the roll cues for helicopter pick-up? Need to think more about that for a minute.

I think I might be onto something here with this use of dampers, and it so happens I've got three or so spare ones on hand. =) I've got a whole new concept to think through now. Various ways I might use a dampened cue, and various situations where it might cause a false cue.

I think the dampers will be set fairly soft, so the delay it creates is pretty brief and basically doesn't happen during slow movements of the seat. The dampers I have are adjustable.

Imo 1 and 3 are easily and best handled by 3 dof motion (minimum). That's what I used years ago with FSX and it felt really good with helicopter flight and it was a blast. As soon as you lift off you have to start working the stick hard in order to keep it level and under control. When you land, unless you're really on top of things, you will feel varying degrees of bump and skid when you make contact with the ground or carrier. The only thing missing was yaw cueing.

For 2, I will be releasing details on my G-seat possibly later today that will fill the void for sudden roll maneuvers. But its designed to work in conjunction with a motion rig so it may not be right for you.

For 4, I will also be releasing information on my new g system for the head which is perfect for this. And this reminds me, I need to add some language to release me of liability from neck injuries when people like you want violence, lol.

LOL too true. I never think in terms of liability when just having conversations and showing things we've done but you are correct we write things all the time that could be considered instructional to some degree. Fortunately for me my net worth would be less than the cost of their lawyer, blood from a stone and all that. =P I'd suggest adding it as a signature so it's attached to all your posts past and future.

I'll look forward to seeing what ideas you share. I haven't talked about harness tightening, head loading, and probably some other things that I definitely intend to include but they will be present.

Bergison's design is basically a g-seat with 3DOF built into it. Having the 3DOF move only the seat has the disadvantage of limiting the range of motion you can use, but the advantage of moving the pilot in the correct direction within the cockpit or car. Of course as always the ideal solution is "D" All of the above. =D

Well, despite my best efforts it looks like my G-seat will be very similar to Bergison's in terms of function. The only thing I'm adding is the seat bottom and seat back are separate and can move at different rates, and a G-vest type feature.

Basically the rest of the features I had in mind either didn't make sense or would require additional actuators for independent movement to be useful.

Design features as of this writing:

• Seat bottom and seat back both "roll", mechanically separate with my current thought to have the seat back move at a mild ratio so it rolls a bit less/slower than the bottom. I will also experiment with applying mild damping to the seat bottom or seat back roll movements to see if either one or the other provides improved sway cues.
• When the seat rolls it causes the "low" side paddles to come up and apply pressure. Essentially they don't move much while everything else does, you get pushed over onto them which should be very good.
• Seat bottom and seat back both "pitch". Seat back is hinged so it can also tilt forward and back independently of the bottom. My current intended use for this is to allow it to tilt forward beyond the range of the full seat pitch in negative surge accelerations.
• When the seat pitches back it causes the back paddles to push forward and apply pressure for positive surge cue.
• Entire seat with pitch/roll assembly moves on heave axis.
• When the seat heaves down it causes the bottom paddles to raise for positive heave cue.
• Harness tighteners for surge/heave cues.
• G-vest to apply pressure to the front of the torso in high-G maneuvers and to assist with the sway cues.
• Head loading to simulate G-force pushing the pilot's head around.
• Possibly a ratio applied to the G-vest sway and head sway so it leads the seat tilt a bit as the cue gets bigger.

At this time I don't see any advantage in having the cushions sway, as it will be redundant to the roll motion unless I have separate actuators.

Well, I think I've finally got the design nailed down. A bit less elaborate than I had planned.

The actuators will be mounted vertically behind the seat, attached to the inner frame that goes up and down.

I've got a pretty good idea of the linkages I'm going to use but won't know for sure that I've got it sorted until I draw it all. This drawing doesn't show the actuators, actuator linkage, belt tensioner, G-vest, or cosmetic/safety housings. The friction points for the paddles will almost certainly get rollers or some other solution. The different colored parts are there just to make it easier to see what's moving.

The shape is a near copy of the ACME DMS G-seats, but there are quite a few major differences. Functionally it's more akin to Bergison's rig.

I will probably take another look at actuating the paddles with actual linkages. I'd rather do it that way but it will add quite a lot of mechanical bits and I expect things will get crowded pretty fast. I think that will allow me a lot more control over the sway paddle movement vs. heave/surge paddle movement, and hopefully make it possible to actuate the paddles from nearer their load center of gravity.


(I drew and animated this image with Fusion 360)

New version. Added the ball transfer rollers that SeatTime posted about and made some changes all around.

I've only got the bottom done so far, but the back will get a similar treatment. I think I've got it figured out in my head but a lot of times when I think that and draw it in 3D I find my idea is flawed.

The roller assemblies are just floating in midair right now, I just have to run mounts straight up from the floor to them.

Hi Trip,

your 3D animations are a great way of illustrating what you have in mind.

In your most recent animation, the backrest is not participating in the roll movement anymore, as it still was in the previous version. Is there a reason for that?

And I notice that your heave travel is quite large. In a previous post you intended to have it 600-900 mm long. Is this still your plan?
I am asking because both the Bergison as well as the ACME seats are seat movers, which means that your feet remain on the ground.
This will probably increase the feeling of your body being lifted upwards or being pressed into the seat.

However, if the heave travel distance is too large and causing your legs dangling above the ground, this may feel unnatural. What do you think?

really pretty nice!
but I not sure if it will trigger the right sensation (needing going down to feel pressure on the side...)

keep going! it's difficult to invent

This is a blast to watch unfold!

Hi Markus, RacingMat, and CFischer, thanks for your kind words and input. =) The tools I've gained this year (3D printer, K40 laser, and Fusion 360 which is free) have changed everything for me. I feel like I can build almost anything! Being able to create these designs and see how the parts interact is fantastic.

Markus, it's not very obvious because the only part of the seat back you see is the bottom where it is very near the fulcrum but if you look very closely at the newest animation you will see that the seat back is actually moving in the roll axis, but at only 50% of the seat bottom roll rate. I can adjust this ratio, but the hope is that this will still move my body effectively and also provide that feeling of friction against the seat back as you move in the seat from the G-force. The G-vest will be used to also assist a little in pulling me over. Developing this seat has included some time spent yanking the steering wheel back and forth at around 20mph in the car to mentally analyze what I feel and that has lead to some of the ideas you are seeing me try out here. =D

The 600-900mm travel is referring to the heave travel of the "elevator" a.k.a. "lift" platform that the whole cockpit and the G-seat will be put atop complete with all vehicle controls. That platform will be only heave, no other motion is planned. The elevator should have about 900mm of travel, but how much of that I can use is limited by ceiling height! When I move the simulator to it's future location I'm hoping to be able to use 100% of that, but where it will be upon initialconstruction I think it will be 600mm at most.

I have been giving a lot of thought to the extents of the travel of the G-seat as well. My current thought is to build it with a little bit more travel than I think I could really use and then be able to adjust it down both with changes to bellcrank ratios etc. and software control of the actuators. Decreasing it should be far less work than increasing it.

RacingMat, I think the fact that the animation is showing pitch and heave movement at the same time as the roll movement is greatly exaggerating the appearance of vertical movement on the roll/sway axis. The theory is that my body will only move in realistic distances and directions. This is based on SeatTime's work. For example in an upward heave movement the center panel goes down while the sides go up, with the intended result being the eyepoint moves down a bit but you feel increased pressure. Since I doubt I'll be able to achieve that perfectly, I should be able to mix a bit of heave movement into the maths to compensate. I think roll/sway is the place I'm most likely to need this. I expect to find that my eyepoint is dropping noticeably, so I will try to add a little positive G-seat heave to the roll cues just enough to put my eyepoint where I want it.

More animations will be coming, hopefully 100% complete before too many more days. I'll also add animations showing individual DOF's. The focus up until this point has been coming up with a design where the various DOF's can all work at the same time. That's what the animations have been showing so far so the movement can be misleading.

I'm also getting to where I'm much more satisfied with the design being somewhat original now that I've started to get things sorted.

I think is the most significant innovation (which may or may not work out!) is the roll/sway cue. The design is based on my findings using my car for testing, combined with the ACME roll/sway cue system and traditional G-paddle setups. Here is the difference:

Other than that the only thing I think may be fairly original is the concept of rolling both the seat bottom, G-vest, harness, and seat back together all in the same direction but having the seat back and harness move slower (initial experiment will likely be 50%) than the bottom and vest. I'm doing this for two reasons. First for large roll DOF movements to have that sway cue feeling of sliding across the seat surface because the G-force is moving you but not the seat, and second to try to make fairly small and slow cues on this DOF feel more like roll cues than sway for helicopter pick-up and touchdown. Will it work? Opinions will likely vary because different people are more sensitive to different cues regardless of whether those cues are true or false. If I find it effective myself that's good enough. If not, I'll change it.