007 G-Seat: Ultra Compact/Servo Based

Hope you find it useful. I like building with real physical stuff, but my brain is better suited to code, as i tend to work towards a solution incrementally. When i get code wrong i don't have to throw stuff away and start again cutting to size, redrilling etc Do...Loop

A productive weekend. I had some steel plate cut to size and made some cages for the servos, spent many happy hours drilling holes, and took the plunge to cut up my foam seat base to enable it to flex with the flap movement.

Overall conclusion: Not bad. I need to do some tuning as I can get a fairly nice sense of pressure but it doesn't kick in soon enough. If I fly a stunt plane and make tight turns its quite nice, but that won't be the norm. One effect I was hoping for that these servos just aren't capable of is the thump that comes with a hard landing. They aren't quick enough for that.
Overall then I want more power so I have two of these on order. 24v and almost double the kgcm. The KingMax servos can be put on the seat back plate for surge. Also I'll test the foam base without the vinyl wrap and try padding the front of the seat to make it deeper. I want to minimise the movement-to-sensation time.

I've also been thinking about -ve flap movement. I suppose if neutral is slightly high then it can drop, so I'll experiment with that. It won't be sustainable but as an instantaneous sinking feeling in windy conditions it might be nice.

Your experiments and shared experiences are really interesting, thanks for that.
When my rig was still in one peace, I also used the negative movements of my heave paddles and liked it, because it felt like releasing pressure. (I also had the paddles a bit angled in their neutral position.)

About the servos you ordered: Just out of interest, why did you choose these instead of the SUPER500 ones you mentioned in another thread? They also looked quite promising.

@MarkusB - Thanks, I'll try that -ve heave. I switched to these servos for 3 reasons:

1. I'm tight for space under the seat with my rig and 2x Super500's won't quite fit
2. The 110kgcm is actually slightly "stronger" - to match the speed of rotation the super 500's would need 4-5x more leverage, which dilutes the force
3. The open nature of these will be good for cooling

It's also nice that they're cheaper

Just to give an update on this servo, they arrived from China very quickly. And they are reassuringly rugged. Unfortunately one is a duff and the supplier is being difficult about it. The one that works handled my tests pretty well though. The design seems to be something that has evolved over several generations and comes in some different flavours with different gearing. This YouTube demo is quite impressive.

Dang dude, thats a Beast of a little motor. ....not super fast...but lots of torque.

Part Two - Surge

As I am currently waiting for a replacement board for one of my 110kgcm servos which I'm now using for the heave function, I have repurposed my 65kgcm KingMax servos to be used for surge. I've certainly made things difficult for myself too. The original goal was to add G-Seat functionality to a compact racing seat with minimal visible external gubbins. Ideally I wanted it to look just like the normal seat but with secret G-Seat functionality built in. That proved just about possible for heave, by lifting the seat cushion and hanging some servos underneath, but with surge I have all the upholstery contend with and if I tore that off it would never go back on right. Now I see why people build a separate seat or work with a bare bones aluminium Kirkey style seat. Still I ain't giving up.

I realised I was never going to mount the flaps and servos directly to the steel seat back. Just not enough room to measure, position, drill and fix. I decided to:

1. build the flap/servo system onto a separate steel plate
2. dismantle it and use the plate as a template fixed to the rear of the seat to drill through where needed
3. Move the plate to the front of the seat back and reassemble the flaps to that with the servos behind the seat back

It would involve some drilling through the vinyl but I got away with very little in the end. The hardest bit was cutting through the seat back to allow enough space for the bearing rods to move. In fact the reason I don't have completed pictures is because I'm still cutting away at bits of the seat back that are catching the bearing rods and servo arms, but I'm 90% there, and I'm very pleased by how discreet it all is. I've yet to find out what it feels like of course, which I'm rather nervous about.

I love it! Well I did for about 5 minutes.

Now I have £140 of burned out servos. My conclusion is simple - Small sealed RC servos should not be used for motion platforms. I disassembled one of the dead servos and the motor inside is so small. The motor still works so its the circuit board that has failed.

I do have some other learnings.

1. The principle of using hinged flaps to simulate surge is sound and a very enjoyable addition to a seat mover. In fact I enjoyed it so much I stopped playing around in my Cessna and starting blasting around the skies in an F15 with afterburners so perhaps that's why I stressed this little servo so much
2. The flap needed to go up higher. It was a good effect in my lower and middle back but I didn't go up high enough so upper back was getting no sensation

The plan now is this:

1. Switch over to using my 110kgcm 24volt servos for Surge as these are open with heat sinks - see here
2. Get 2 more of these 24v servos for heave, but I might as well get the stronger 260kgcm - see here

I have photos of the finished thing with servos/flaps in-situ but there's no point showing these now. I do however still think I can build this with the stronger servos and fit it all into the same small space. There might just be a small bulge in the vinyl at the back.

Disappointed
I now have one heave flap fully assembled and have just tried it to see how powerful these things are. I'm still using the 110kgcm while i wait for the 260kgcm.

I was thoroughly underwhelmed. Very weak. I ran some tests which came out that a more realistic rating is about 60kgcm.

@MarkusB I am starting to wonder whether I really do need to restrict myself to the fast 0.12 sec/60deg servo model. I wanted the fastest possible response to get the most instant flap movement. Of course they don't really move that fast under load. I have designed the leverage to give me full range flap movement within 60° of rotation so that 0 to max is ideally achieved in 0.12 seconds. Incidentally the lever arm is 51 mm and lifts the flap 1/3 in from the edge. However if I went for the more highly geared version of this servo, it is rated at 380 KGCM but takes 0.5 seconds (same speed as the motor you were interested in) , perhaps that is actually quite adequate and the extra gearing should give comfortable margin even if the true torque is less than advertised. Trouble is I feel I have spent enough on servos now.

Hi @Spit40, I'm following your g-seat testing with great interest. Sad to read that your servos seem to be so much weaker than expected.

Did you ever measure the force/torque that needs to be applied to your flap levers? Some time ago I did some tests by using a scale.
In your case you would need to connect the scale with the flap lever. Then put the flap under real load (i. e. sit down onto the seat flap or lean against the backrest flap), and measure the weight by pulling the scale. Probably a second person needs to help with this.

And did you already measure the real stall torque of your new servos and compare it with the claimed torque? Just for checking if the difference between claimed and real torque is really that big. You could use the setup of the video that you posted on September 13th.

At the moment I am building a new rig from scratch, because my old one had too much weaknesses.
Regarding the g-seat component: For the backrest, my current approach are three 30kgcm servos per flap. I am still at the beginning of my constuction and have lots of non-simulator-related tasks at the moment, so progress will be slow.

About the 500kgcm servos: I am thinking of using them for my "movable backrest": The entire backrest including the flaps will be able to move up and down, which shall simulate my body being pressed downwards into or lifted upwards out of the seat (as part of the "heave" component). And instead of servos, I am also thinking of using bowden cables connected to a wheelchair motor. So I'm still not sure whether I will give these servos a try.

About the speed: I think that high speed is especially important for simulating street bumps. For movements of airplanes (which is my main focus) a slower servo may be sufficient.

With regard to required kgcm i didn't ever test that, but relied on estimates. I felt that there was plenty of headroom with these new higher power servos. If my new 260kgcm's don't do the job i think more measurements of required force would be good. As for real torque, the real torque is about 50-65 kgcm. I suspended a weight below the lever arn and made an arduino script to smoothly raise and lower the arm. 10kg x 5cm lever was going strong. 12.5kgcm x 5cm lever was struggling and couldn't go full range. It didn't stall though.... makes me wonder what is the definition of a 110kgcm servo, does it count if the servo achieves anything more than a stall?

A busy weekend. The heave/underside is now all built, mechanically at least. I'm just waiting for my new servos to be dropped in place of the old. I'm very pleased with the design. Its very low profile underneath, in fact I have about 2mm clearance between the motors and the centre bar of the 3DOF. There's very little space for the bearing rods too, but liberal use of countersunk screw heads allowed me to fit everything in and still use a thick gauge of steel for the lever arms. I'd really rather have gone 6mm with the bearing ends but space limits me to 5mm. Should be OK for flying.

The rating given by servos can sometimes mean the stall rating, even so a 260kgcm servo will have a rating of 52kg lifting force with a 5cm lever. The 110kgcm should theoretically stall out at 22kg with a 5cm lever. Also look at voltage ratings on the servo, sometimes the max force of the servo might be using a higher voltage than normally nominated.

@Alexey thanks for the clarification. I did expect the 110s to comfortably lift 12.5kg at 5cm and go full range. My idea of reasonable would be to struggle but still go full range at around 18kg and stall at > 22kg.

I read in an earlier post that chinese servos are often overrated. Are all servos often overrated or are the good brands reliable with their spec?

Just a letting off steam post...

I can't believe how frustrating it is trying to obtain these servos from China. After 2 weeks my 260kg ASME order just arrived - they sent another pair of 110kgcm ASMC. Now 2 months after ordering I have 1 working ASMC from the original order (performing at 50kgcm) and one lacking a circuit board and 2 new ASMCs which should be ASMEs.

I really think this particular motor design might be perfect for G Seats and if I ever work out how to source them reliably I'll be delighted to share this information with everyone. Right now.... arrggghhhh

Is it a seat or is it a G-Seat?

Finally some good news on my project. Although I still don't have the higher powered 260kgcm servos which I'll use for heave (the seller doesn't want to pay return postage on their shipping error) I have now rebuilt surge with the 110kgcms and they work great. I'm using a 3.5cm lever arm, which combined with about 75degrees movement and the placement of the bearing arm is giving about 6cm of movement on the surge flaps. The pressure feels pretty strong and the servos seem to be holding up well to some real life testing - at least I haven't burned out anything yet. And it all fits in so neatly you'd never know it was anything other than a normal seat, apart from a few drill holes through the vinyl. Just needs a bit of SimTools tuning but I'm very pleased with it.

Springs!

@MarkusB I've just reread one of your posts and reminded myself that you were putting springs under the heave flaps. Great idea. That way the servo doesn't need to do all the work - 1/2 the work maybe? If neutral is the mid point? Can anyone work out the forces and best settings for this? My design is so compact though that I'll have to see what kind of springs would do the job. Maybe just normal coil springs as I have a metal plate below each flap that the springs could press against. I have no idea about springs though and how to work out tensions and forces.

If I eventually get these 260kgcm servos and incorporate springs I should have ample power and speed. My hope is I can simulate a bumpy runway or in fact bumpy air. I think it should also be strong enough for a driver's G seat.

Update: Having doubts about this now. Do springs really reduce the work the servo has to do? Surely every ounce of force they aid in one direction has to be pushed against in the other?

Runs hot

I've been playing with the surge flaps for a few days now and they're working well. After a bit of use though I notice that my back is getting warm. It turns out the motors are running pretty hot. There's a heatsink on the controller board but its the motors themselves that get rather toasty. I measured the motor diameter as 36mm which I discovered corresponds to a 540 size motor, so I've got a couple of these heatsinks in the post to me now. It looks like hot 540 motors are pretty easy to deal with as I see all kinds of options on ebay, many with built in 5v powered fans.

https://www.ebay.co.uk/itm/RC-1-10-...nk-Heatsink-Tamiya-HSP-Car-Truck/263030741808

The other reassuring thing about discovering that the motor built into these servos is of a standard size is that its therefore feasible to upgrade them. The gear mechanism can be accessed by removing 4 screws so the whole thing is quite open. You just can't do this sort of tweaking with a sealed Turnigy style servo. I'm still hopeful that this servo model can provide a great and simple solution for a lot of people to build compact G-Seats.... Despite the fact I'm still arguing daily with my chinese supplier of these servos and have yet to receive the 260kgcm model. I've always been a glass half full kind of guy.

Understand were your coming from for your design, but don't know if I agree with this statement as a whole, if you keep within the servos limitations and spread the load over multiple servos. I had been running a bank of 8 x 15KGcm (sway) and 2 x 65Kgcm (harness) servos in my sim for over a year without a issue. I've only moved away from them due to the servo gear noise and wanting to go to a whole new level of simulated Gs. Used within their specifications, they can work well and can be simple, light and compact. Agree that purchasing from China can be a lottery - Eg. MM drivers.