007 G-Seat: Ultra Compact/Servo Based

Hi @Spit40,
about my springs that I mounted below the heave flaps: They definitely helped the motors to move the flaps while my 90kg were resting on the seat. But with my 25kg daughter on the seat the springs were too strong for the motors.
So my conclusion is that the second approach that I am currently working on will go without springs. Instead I will try to mount the flaps in a way that they don‘t have to lift me, but just apply pressure.

Yes i'd agree with spreading the load over multiple smaller servos, though for me, even if I wasn't constrained by the compact design i'm aiming for, i wouldn't go back to those smaller servos. These robot servos are reassuringly rugged and seem more suited to the forces needed for a g seat.

That's great feedback thanks. I was starting to doubt the validity of spring assisted heave. What do you recommend in terms of spring strength? I did some testing and concluded i need about 30kg of 'force' on the heave flaps, but i'm not sure what strength of spring i would want to use.

@noorbeast @MarkusB i've just been researching springs but also trying to get my head around the physics. Feedback on both aspects would be very welcome. I thought if i got some mountain bike type strength springs under the heave flap it would give the kind of assistance i need. I was looking at this sort of thing. Rated 5Nmm so it would take ~20kg to compress 39mm from its extended length of 76mm to solid/collapsed length of 37mm.

http://www.entexstocksprings.co.uk/index.php?route=product/product&product_id=1669

Let's say that 39mm of travel was adequate the servo would have to apply 20kg force to fully close the flap.

Prior tests show that to reach full extension with my weight i need 30kg.

At 10mm of lift the spring is compressed 29mm so equivalent to 15kg - does that work like 15kg of assist?

At full extension there is no assist, so i still need 30kg from the servo, so how does this help? If i design it so i still have assist at max extension, then i need more servo power to close the flap defeating the purpose. If i don't rely on the servo to close the flap but just use body weight then even when fully closed i feel a significant upward force (relative to the part of my body which isn't on the flap) which i don't want to feel. Perhaps the seat flaps need to go full length to use spring assist?

This is exactly what I did on my g-seat.

https://www.xsimulator.net/community/threads/diy-g-seat.9848/page-2#post-130228

I thought it isn't necessary for the flaps to hold your weight so went with this approach. Works great.

Things are looking up

The 260kgcm arrived today along with very sturdy servo arms. I've been testing their actual torque. I do get a full range of the 75 degrees of movement I need up to 100kgcm or so. When I increased to 130kgcm they didn't quite manage it, so I'd say they are about 110kgcm true. That should be enough for me - with a lever arm of 5cm they'll lift over 20kg which I think will do the job.

The blue heatsink is something I added myself.

First test - cautiously optimistic

I now have the 260kgcm set up for heave and the 110kgcm for surge. The 260kgcm seem like they might be up to the job. I needed to cut away more of the dense foam seat base as that was adding quite a bit of resistance to the heave flaps. Due to the amount of force these single servos are applying it needs a pretty heft servo arm though, so there's some work to be done to swap the ones I have and modify placement.

I hope for a proper report after the weekend.

Secret Hidden 007 G-Seat Finished (Well, you know...)

I finished the g-seat this weekend. By that I mean it is all wired up, all 4 servos are working and under control of Arduino and responding to game telemetry. I'm still tuning a bit - trying for example to get the surge flaps and heave flaps to give an even kind of pressure when on a roll. Mechanically I also intend to reposition the surge flaps at some point as they don't come up high enough, so despite a good pressure in my back it isn't as convincing as I think it could be. I'll make the surge flaps extend higher and will see if I can get a bend in there to allow them to extend more to the edges of the seat. When I started this there were a few scenarios I hoped the seat could provide a believable physical feedback for:

1. Landings - I wanted to feel the difference between a light or heavy touchdown. In fact I wanted heavy landings to be off-putting to encourage me to fly better
2. Turbulence - I wanted to feel it if there was turbulence
3. G forces on take off, banking and strong acceleration or climbs

I would say that I am getting some of all of this now and I hope for more realistic sensations as I tune it better. I dialled up the turbulence last night and was really pleased to see the effect of being buffeted about, so much so that I had to quit the game when I could smell a servo overheating. That's another little improvement - I've just ordered a pair of these for the heave motors. The motors have a vent slot as they're designed to spin fast and can get hot, but when used in a servo this built in cooling mechanism can't take effect. One good thing about this servo model is that if I do burn out a motor I can replace it for about £8 as they are standard 550 size. In fact I plan to experiment with swapping the motors to see if I can get better performance. Something like this for example.

Time for pictures:

As a reminder this is how it looks externally (bought for £70 off ebay):


Underneath I now have 260kgcm servos powering the heave flaps attached to some folded 2mm steel plate that just spans the main crossbars of the seat base. The blue finned heat sinks are about to be replaced by fan assisted heat sinks as the base gets pretty warm if you work it hard. Reassuringly its the motors that get hottest not the circuit board as motors are cheap to replace. There's very little room to spare with the servo arm and lever. I had to shim the servo back precisely to give the arm space and used countersunk screw heads to minimise the space it used. I also had to cut away the seat's handle and mechanism for adjusting forward position on the seat rails.


The servos came with strong servo arms that I very much recommend. I tried some other flange couplings that took an 8mm shaft but they just couldn't cope with the forces and slipped. The arms have a D shaped slot so can't slip.


Here's the back. I constructed the surge flap assembly attached to a steel plate before I attached it to the steel plate seat back, using the extra plate as a template for drilling holes through the seat back. The extra plate also provided reinforcement as the seat back was quite flexible. The depth of the motors and cage is about 65mm so hides quite neatly under the back flap of vinyl. These servos are 110kgcm and feel quite adequate for the job.



Lifting off the seat cushion now. I had to cut up the seat cushion into two parts. The rear is not lifted by flaps so you feel the lift through your thighs only. I had to cut into the front part too, removing a 50mm wide channel from the top and bottom, to give it as much flex as possible rather than have it resisting the surge flaps. There's a stop plate on the picture below to prevent the rear seat base from sliding forwarding and getting lifted by the surge flaps.



In the end I had about 4.5cm of leverage on the surge flaps which produced around 7.5cm of lift at the very edges. I did a lot of tweaking of the limits of movement, eventually setting the neutral position for both heave and surge flaps to about 20% extension so that there was a more immediate feeling when they acted rather than them just using up slack.


I was pleased with the addition of the 2.8mm 9-way car mini connector kit as I had to remove and replace the seat so many times during the build. For wiring I'm going with 1.5mm 2 core for the 24v (only about 0.5volt drop over 2m) and some 6 core alarm wire for the 4 servos+grnd+5v.

Improvements
Apart from the tweaks I mentioned, I do still intend to explore options for more power. More power is always better. With this design there isn't much scope for a seatbelt tensioner, though other seats could accommodate that better.

That is just awesome!! Such a neat and compact build, you'd never guess it looking at the seat what magic it holds

Cheers mate!

its working for now? interesting on how to put a transductor on seat direclty... effects is good?

Yes, all working. I haven't melted any motors yet. I had a moment when it seemed dead yesterday, but a connector wiggle sorted that. I don't think those 2.8mm connectors are good enough for 24v with a few amps.

I'm happy with the transducers. I had them on the seat back at one point and got lots of amplification from that, but had to move them. I mount the seat on rubber isolators to try and localise/maximise the effect. I'm using a 100w amp but have 200w on order to beef it up a bit.

@Archie - in an FS2 post you said "You can use Heave to achieve what you describe. If you put small values in the Tuning centre but dial down the % in the Axis for heave you can get the subtle feeling of the aircraft "settling down" after a climb / dive. The bonus also, is you get a nice "thunk" when you land and taxi. "

I'm now wondering if I can combine both a quick strong pitch up from the seat mover with the flap lift of my G-Seat to strengthen the "thunk" on landing, as its not quite as thunky as I'd like and I don't think my G-Seat motors alone will really be up to that. The problem I have is that your approach involves narrowing the tuning range so that the DOF goes rapidly from min-to-max which is mapped to a rapid but short range lift of the seat mover. If I did that I'd lose the more gradual heave range which is producing great results with the G-seat as I tighten into turns and on take off.

There's only one tuning centre option though isn't there or is there some functionality to SimTools (@yobuddy ?) I have yet to discover?

@Spit40 - Given your setup you may actually find you can use Surge a bit better for that effect. Playing with Pitch on a flight Sim may well introduce some unwanted "nervousness" from the Sim on approach etc.

Surge (unless you're in a Jet) is not really a much-used force (at least on my 2dof) - I only really feel it on take off, but could certainly up the numbers to give a bit of jolt as you touchdown and ground friction kicks in ??

Something to tinker with anyway...

The surge is such fun (and doesn't strain the motors anything like heave does) that i've dialled it up quite a bit so i feel it in the new FS2 biplane on takeoff for example. I also notice it when pitching forward. I'll look out for the landing friction effect.

As for seat movement on landing it occurred to me that 10% seat up on heave would still be useful on landing as its the one time that heave goes very rapidly from min to max even if the range isn't too narrow.

I think i may start a thread to consolidate learnings in mapping flight physics to motion platforms. All these little tips are spread across dozens of threads and its not really plugin specific or project specific.

Someone on a flight sim forum asked what was involved in making this G-Seat. Here's the initial braindump I gave.

1. Find a suitable seat. There's the model I used or its a question of a checklist of things to watch for when using another. TBH without checking them out and half pulling them apart I wouldn't know how other seats are constructed and whether their frames are suited to it. I was keen not to rip off all the upholstery which made things quite a bit (much!) more fiddly
2. Shopping list
   1. 2 x ASMC-04A 110kgcm servos & 2 x ASME-03A 260kgcm servos ~ £240 total
   2. ASME Servo arms ~£40, though I got 2 of them with the ASMC-03A, so then only £20
   3. Heat sinks/fans ~£20
   4. Steel - not sure, I got mine free as offcuts from a local guy
   5. Folding the steel - not sure. Local guy did it for me
   6. Bearing arms/rods ~ £20
   7. Hinges/miscellaneous hardware and screws ~ £20
   8. Power supply ~ £40
   9. Arduino circuit board ~ £8
   10. A box for PSU - use whatever you have
   11. Miscellaneous wires and switches, electrical connectors ~ £20
   12. Software to interface to FS2 - Need to get a SimTools licence ~$70 I think
3. Code for the Arduino to take SimTools / FS2 telemetry and output to the servos - I have that, but would need tweaking depending on the seat
4. Designs
   1. These will vary according to the seat but I have various notes on placement/angles
5. Skills/Tools
   1. Good drills for drilling lots of holes in steel. Bench drill is handy.
   2. Ability to drill precisely in the right places - that was a learning curve for me
   3. Ability to do straight cuts in steel or a way of getting it done
6. Something to mount the seat to

Step 1: Choosing a suitable seat

I've been asked how to get a seat like the one I used to build this. I got very lucky there as I wasn't planning to build a G seat when I bought this cheap seat off ebay. At the time I was frustrated I couldn't get a cheap JEGS seat in the UK, so bought a second hand Cobra Classic seat that had been in a Triumph Spitfire for £70. I see one for sale here

http://www.minispares-online.co.uk/...st-basketweave-classic-seat-for-classic-mini/

The important features (for this build) are as follows

1. The seat base was flat made of cross sections of steel about 2cm square as per the picture. I could easily bolt the flaps to the top and the servo cage underneath
2. The cushion lifted off with some simple unclipping.
3. The cushion was a single piece of dense foam that I cut in half (front and back) then I cut into the front half down the middle to allow it to flex on each side. All the pieces then fitted back into the vinyl cover
4. The seatback was accessed from the front by removing some wire twists that held the vinyl to some steel rods, so I could get under the vinyl and slide surge flaps up
5. The rear of the seat had very easy access. There was a flap held by a clip that I could simply release

Here's the base mounted on my 3DOF. The previous posts show it with seat cushion on.


If you want to build one of these, then starting with exactly this seat is obviously a safe bet. But if you can't get hold of one of these, I don't know what else would be suitable.

Does anyone with experience of vehicle seats have any ideas?

Solving the hot motor issue

These heave motors are working hard given their size but I think I've solved any overheating risk by attaching one of these to each of them.



This is how it now looks


I had the strangest experience when I first strapped these things on. They wouldn't work. After trying everything, checking voltages etc etc I finally worked out that the strong magnets in the servo motors were preventing the tiny fan motors from working. I found that by rotating the fan motors around to a particular place on the servo I could get them to function.

For power, they use 5v but rather than power them off the Arduino I use a **very** cheap buck converter (top left of my picture, seriously it was £1.60 inc. postage) to take a feed off the 24v power and reduce it to 5v. This means that that the fans stop when the servos are powered off rather than going constantly via the USB power.