Showroom Grigory's 6DOF #2 -- compact and no linear actuators, 220V servos

The video above is without motion cancellation. I still cannot figure out how to attach Vive tracker so that it is able to correctly capture the movements while being isolated from harsh vibrations.
I tried fixing the tracker to the top of the seat back, just behind my head, both rigidly or with some dampening from several layers of thick sticky tape — no luck, the bumps make the viewpoint drift and jump around randomly.

Grigory,

Can you tell me how you got the servo's to work on one arduino?

We want to do the same but ask our selfs how to do it. One com port for the arduino and 3 motors or more to get diferent feedback.

I hope to hear from you,

Greetings Eric Rijgwart from Holland.

Hi Eric,

In my case Arduino is just passing step/dir commands to the servo drivers, it does not get any feedback from them. Motor control is fully managed by the drivers.

Would be careful before mounting vive trackers I heard that can be damaged doing this

You heard this about Lighthouses, not Vive trackers, which are meant to be mounted on moving things.

Typed before thinking lol yeah that's what I meant and nice build looks really snappy

Since the tracker is threaded for a camera clamp you could try inserting a rubber isolator:



Unfortunately I haven't done much testing with this so I don't know if it will solve the problem for sure.

I've also considered various ways of adding more mass to the tracker side of the isolator. I tried putting the tracker on a threaded rod with a heavy weight in the middle, but never dared to drive with it because it seemed like the isolator was too flexible and I really didn't want that chunk of lead hitting me in the head

I have some less-dangerous ideas about suspending a weighted ring around the isolator but haven't tried it yet.

I experimented with putting the tracker on this thing using different types of rubber balls with different stiffness:
(https://www.thingiverse.com/thing:2370136)

No luck unfortunately. It either does not filter out the small vibrations from road texture, or is too flexible and even exaggerates larger movements.

I just use the rig without motion compensation now without any discomfort. Since I have a GSeat mounted on my 6DOF, I do not really need large cockpit movements.

Have you tried a less snappier tune in simtools? your rig is the fastest I've seen so far maybe it's too much to process or too much for the tracker to accurately read

No, I just love how the rig moves now I even disabled the seat's buttkickers because I can feel all the road texture just from the rig.

I am now running the rig with a bit smaller range of movement than what it was in the video above, and it feels very natural. GSeat takes care of longer forces and no big tilts are needed. I may eventually add a belt tentioner to improve braking feel, but acceleration and turning is already fine.

Have you tried mounting the tracker in a location on your rig that moves the least? Like down low by the connecting rods or at your center of rotation. I understand that heave effects the whole rig but it seems like putting it on the top of the chair gives it the longest lever arm to have the most violent motion.

Or maybe just strap it to your chest go pro style.

Hi Grigory, I have been following your sim builds for some time and have a few questions

lots of people go from gearbox/levers to a linear actuator/ball screw design

i see you have gone full circle and back to levers

does a sim really need 500mm of movement with long linear actuators

how do the 2 designs compare and which is your favourite ?

it seems very similar to this design -

My impression is that it is exactly the road texture/heave that disrupts the tracking.

That is an interesting idea to test. I expect, however, that there would be issues with obstructing the line of sight from the lighthouse to tracker with my hands when driving. The chest will probably need to be stationary relative to the cockpit. For example, I wonder if breathing would cause some funny effects.

I have just now thought that I should try putting a pillow on the seat back and then attaching the tracker on top.

Yes, I did a full circle indeed And not only in terms of actuators, but also with the whole rig type: 6DOF->Seat mover with traction loss->6DOF.

I do not think my rigs can be used to objectively compare the different approaches, as the manufacturing quality of my builds improved with each iteration. At least I hope it did Nonetheless, I can explain why I did what I did.

The first attempt with lever-based 6DOF failed because of the terrible backlash in my worm gearboxes. It was just not usable.


The 6DOF with linear actuators worked mostly okay, but it certainly had a number of issues.



First problem was the weight of the steel cockpit and relatively low speed of my motors. Theoretical maximum speed of my linear acutators was 250 mm/s unloaded. Back then the movements seemed acceptable, but now I understand that all effects were very, very dampened. No road texture and all car suspensions felt the same -- soft and spongy. The second problem was the size of the rig: it occupied a circle of around 2.5 meters and I had to use a stool to climb into it. So when I decided that I wanted a sim in my city appartment, relocating that rig from the weekend house was not an option.

My first "city" sim was a seat mover with traction loss (3DOF).


It worked just fine and I especially liked the braking and acceleration forces that you feel with your hands when you the seat moves you closer or further away from the wheel. I think that I may eventually add another bowden cable to my current 6DOF to pull me forward when braking.
The only thing that I was missing was heave. Small road texture was simulated by the shakers, but bumps and potholes were not there. That was why I decided to once again re-build the rig and add the missing DOFs.

I could not figure out how to make a linear acutator based 6DOF that would fit within the footprint of my 3DOF, so levers were the only option.

As a result, my current rig is by all parameters (except the price ) much better than any of my previous attempts:

• It is compact. It is just slightly larger than a stationary driving rig would be.
• The cockpit is made of aluminum and is very light while being rigid. I believe that my Accuforce and pedals weigh almost two times more than the cockpit.
• The 220V servos are FAST! The average linear speed within the working range of 100° is almost 1400mm/s with 120mm levers, unloaded. Unloaded speed in th 40° range (that covers most of the moves when driving) is over 1530 mm/s. I do not know what the loaded speed is, but as can be seen in the video that I posted earlier, it is still very high. I even stopped using the seat buttckickers for the road texture, it all comes from the actuators. There is no latency that I can feel.
• Servos are also powerful. Even the front ones that take most of the load from the overhanging legs+direct drive wheel+pedals only get slightly warm after couple of hours of action.
• Movements are practically linear. I have a self-made software converting DOF values from Simtools to cockpit coordinates and then to lever positions, so that all the moves are geometrically correct.
• GSeat. It is amazing how much this adds to immersion. With a GSeat, you do not need all that much of a cockpit movement and so latency can be further decreased. Also I do not think you can even get the same strength of acceleration feeling with just tilting the cockpit as you can with a GSeat. I am really grateful to @SeatTime for the idea with bowden cables. Without it, it would not have been possible to make the seat so light.

thank you for taking the time to reply to my questions, I like the way you have not given up and have been driven to build a better sim

I really want to build another sim and 6dof seems the way to go

Ball screw linear actuators are expensive to buy and the diy ones that really work are not that much cheaper - i think for a linear actuator a belt or chain drive may be a faster solution with less noise

I am sold on the gearmotors and levers approach - i need to decide do i go DC or servo motors.

My 2dof uses DC motors from a wheelchair and has no backlash or noise at all, however it appears i have been lucky and got a good set first time out

what do you think of this rig - do you think it has enough movement for realism -- https://www.xsimulator.net/community/threads/my-first-3dof-wip-1kw-ac-servo-arduino.8906/
is 3dof enough ?



(you may like this thread https://www.xsimulator.net/communit...s-linear-actuators-vs-3dof.11785/#post-163179)

After trying both powerful 24V motors (https://hobby.npcrobotics.com/store/npc-t74/) and 220V brushless servos I would say that 220V servos are the way to go. The cost of a servo+controller+ultra low backlash platenary gearbox , if all is sourced from China, is comparable to the price of of these NPC-T74 with sabertooths, kangaroos, power supplies, batteries and encoders. Control and the whole setup is so much simpler with servos. I think it is possible to get the similar power and responsiveness from DC, but it would mean huge currents. A single NPC-T74 would go above 200A when stalled. I gave these motors up because they started melting their own factory installed wires. With AC servos I do not have any overheating problem whatsoever.

The range of angular movement of this sim is impressive, so I think it will be great for flight sims.

For driving, I think that the best combination is a GSeat to simulate longer forces and a rig that can actually reproduce short sway and surge impulses. The disadvantage of this 3DOF is that it cannot actually move the cockpit linearly left-right or forward-backward, so the rapid sway/surge moves with low amplitude cannot be reproduced. 6DOF easily does that. I would also miss the traction loss that is not available in such 3DOF.

It seems that the motor-gearbox combination that is used there may be much slower than what I am using on my rig. If I read the servo label correctly, this is a 1500 RPM model. With 1:100 gearbox that would give a maximum of 15RPM at the lever. I have 2500RPM with 1:20 gearbox = 125RPM. Quite a difference. I personally think that fast reactions are much more important for immersion than the range of movement.

Grigory - what is your opinion of these gearmotors ?

https://www.ebay.co.uk/p/48v-DC-500...-Controller/665947544?iid=152435171812&chn=ps

I do not think that 500w with 6:1 gearbox will be enough. Even if the indicated torque of 1.7Nm is before the gearbox, the resulting 10.2Nm will only provide 10kg push at the end of 10cm lever.

Very nice Sim. I have been looking at your power system. Am I correct that the torque would equal 80 nm ??