6DoF Seat Mover with DC worm gear motors

Hi everyone,

I'm looking at putting together a small 6DoF sim and would appreciate any feedback that might improve the design. The end goal is to produce a motion sim that will work well for VR so I'm excluding monitor mounting at this stage, but will include HOTAS mounts for flight sims and possibly a keyboard/steering wheel tray. I have some woodworking and electrical engineering experience but not a lot of metalwork/welding so I'm planning to make as much of the frame as possible out of wood, with essential load-bearing parts out of steel or aluminium depending on availability. A basic stewart platform design consisting of geared motors turning arms and universal-jointed linkages connected to a flat platform is what I have in mind.

At this stage the Dimension Engineering Kangaroo + Sabertooth 2x32 driver paired with the Motion Dynamics 200W Worm Drive DC motors looks like best balance of cost and performance. (Sorry my account is too new to post links but these have been linked all over the forums)

As the motors have up to 20Nm of torque this setup should provide up to 40kg support force per motor if a small arm length (50mm) is used, or 240kg in total. Extending the arm to 100mm reduces this to 20kg per motor (120kg total) so I figured staying at 50mm should provide enough support for dynamic forces, uneven loading, etc. Napkin math shows that with a 500mm base width this should allow for ~10degrees of roll in each axis which I think will be enough.

I am planning to buy a single kangaroo+sabertooth+motor and perform some testing to validate these numbers before going ahead and purchasing all 6 motors, but I thought I should ask the community for some input first. Since I'm heading back home to NZ for a few weeks next week I've got about a month before I want to start testing hardware.

My main questions are:
- Will the motors I've selected (200W, 20nM, 180RPM) actually be powerful enough to move and support a seat+person?
- Is ~10degrees of roll/pitch/yaw enough?

Basically I would like to hear from some more experienced builders if I'm just wasting my time with the smaller motors and really need more power in order to achieve 6DoF motion. As I'm located in Australia the shipping costs of these from Motion Dynamics isn't an issue but it seems like the next step up in size will increase the cost per motor from ~$100 to more like ~$400. That's not entirely unacceptable but isn't ideal.

Thanks in advance!

Wish I could help, but your proposed design and mine are very different. A note on motor ratings, most are not meant to be run even close to this rating continuously (without allot of help to mange heat). Eg. the small motor below is rated at 350W while the big one is 320W (but it is an industrial motor - continuous rating).

@SeatTime Thanks for that - I certainly haven't factored any thermal considerations yet. I has assumed I would eventually need cooling but that's quite a difference in motor size too. I'm happy to post on your thread instead of here but can I ask if you considered brushless motors for your linear actuators?

skim through some of this info on motors. It may help.
http://www.engineeringtoolbox.com/nema-insulation-classes-d_734.html
this is very infomative: http://www.electrical-knowhow.com/2012/05/classification-of-electric-motors.html
hope this helps

Brushless motors? Building the actuators was going to be challenging enough for me, so tried to keep the rest simple and something that I knew worked and was compatible with the driver cards I was familiar with, as building a 6DOF is quite an investment in time and money. Happy to let someone else trail blaze brushless motors.

Thanks for the input everyone - I think I'll definitely need to do some experimenting with the hardware.

@SeatTime I'm really interested in the hobby R/C motors that you can get relatively cheaply (eg Turnigy RotoMax - 2960W 80A, US$130) with insane power specs but haven't seen anyone use one for position control yet so I'm not sure it's viable. Maybe worth experimenting on with smaller motors in the meantime...

'Cheap'? don't forget about the need for a gear box, power supplies and drivers at ??$. Any rate - good luck in your research.

No one has mentioned yet that the sabertooth/kangaroo combination has proven to be problematic when used on motion simulators thus far. While there seems to be better results obtained by using encoders instead of pots with the kangaroo, there is still the issue of using the kangaroos auto-tuning feature on 6 motors for a 6dof when you can only tune 2 motors at a time.

Here's more discussion on the matter in this thread
http://www.xsimulator.net/community/threads/wannabeaflyers-6-dof-with-24v-dc-gearmotor.5106/page-3

@BlazinH Thanks for the info - I have read the thread and noted that encoders would be essential.

I plan to treat them as servomotors controlled via arduino as I have some experience with r/c hobby servo control. My understanding of the kangaroos is that they remember tuning parameters between power cycles so should only need tuning once. But that aside - what issues are people having when trying to train a single actuator at a time? Stewart platforms can move freely when only a single actuator length is changed so while it would be awkward and time-consuming I don't see it being impossible. Happy to be proven wrong though - and of course I'm expecting all sorts of unforseen roadblocks to pop up once I've actually got the hardware in my hands but I do believe in planning in advance as much as possible.

Hi @Khanage my project is on hold short term ( waiting for Replacement kangaroos ) a mistake I made with wiring, fried 1 channel on 2 of my 3 kangaroo controllers, So far the main 2x32 Sabertooths seem OK so just placed and order for 2 new replacements kangaroos and sending them back for maybe warranty replacement ( if they feel generous ) at £22 each it will be just as cost effective to buy new, and if the others are returned they will be spares , the tuning process can be a pain , even though if, done correctly the controller write the values to EEprom so they are there untiol changed ( or in my case till doufuss here wires things in reverse polarity ) .. mind you the new DE describe software has a really nice touch where you can see the encoder Output and the limit switch state Live on screen so that make for a great enhancement during tuning ..

one thing I still cant get my head around is that DE states tuning needs to be done with a decent load on the actuators , now this is fine if your actuators don't backdrive and are able to hold position without power , the problem I have with my sim is ,, trying to hold the Actuator at the mid point , which is required for succesfull autotuning .. Real hassle and bloody dangerous because with the platform loaded the Autotuning process is quite violent.

These issues aside , I also found the Encoder option seemingly a better route for this type of design, time will tell once I get my replacement Kangaroos , But as of the last round of testing, I was getting very promising Results which verify the mechanical design and now that I have fitted Gas struts ( gas Springs) im hoping i'll be able to increase the CTC length to 75 - 100mm and so improve the operation angles of the platform to circa 15 degrees for each Pitch and Roll, and the other DOFs will improve by default .

Yeah guys I know, I ramble on but just keen to keep you guys posted as im sure i'll have good news soon :-0 ( famous last words LOL )