Prototype Pitch & Roll Full Rotation

Hi all, I hope I get this new thread thing right, I am a forum newbie. Plenty of reading but I never had anything worthwhile to add until now. I hope it works. Here is the start of my motion sim project.

​

I hope to build more into it soon, but this was the first step. I would like my final version to have 5 DOF, which can be split up into 2 different rigs in one; a 3 DOF chair rig (which handles sway, heave and surge) built into a larger 2 DOF full rotational rig (pitch and roll).

This is my first attempt at such a machine, and I know I am probably biting off more than I can chew, but I love a good challenge and I always dream big. That is why I am starting off with a small prototype first, to test design ideas and develop control logic.

The rotational prototype I am displaying here is what I have completed so far, and the chair prototype is mostly drafted up.

All of the non mechanical parts were designed and 3D printed in my shed. The motors are both steppers (easier for me to control and write code for) driven by 1/16th stepper drivers. The roll motor on the inside is a Nema 17 47N.cm and the pitch motor is a Nema 23 185N.cm. Both are controlled by an Arduino Uno. I tried to learn how to use the accel.stepper code, but my codefu is quite weak, so I wrote my own. It works Ok, but it could be better. No fancy PID or anything.

​

At the moment the code accepts a string from the PC which indicates a desired position for each axis. The Arduino then decides which direction is the shortest (Clockwise or Counter-clockwise) and then calculates an acceleration - cruise - deceleration path to the set point.

Power is transferred to the inner gimbal by way of hand made slip rings. These were by far the hardest part to design. If anyone is interested I can share more details/designs.

So far to control it I have been sending strings of data by hand (via the serial monitor in the Arduino IDE) and have yet to integrate it into the awesome software you guys have made. All in good time. The video I have uploaded was done with a camera in one hand, and typing into the console with the other. Pretty crude I know but it does the job.



So far I don't have very many specs for the full scale sim as far as hardware goes, but I'm working all that out as I go. Nothing will be purchased until I am confident that what I am getting will both work as intended and and be cost effective. Until then there is no harm in playing around. I intend to use this with all sorts of flight and space based simulators. I am not really into racing, but then again I haven't really tried too many titles.

Hi, you certainly like a challenge for your first sim (go hard or go home ) and it was certainly a good idea to build a working model. I notice that you are also in the industry, so hopefully you can leverage off that (Eg. find second hand parts) to bring down the obvious high costs of building such a rig. Welcome and all the best in your project, I look forward to any updates.

Thanks @SeatTime for the reply. Yeah I do hope to use industry contacts to make things cheaper. I have gotten some advice for industrial servos from a friend, and he knows others who can give good prices. As for second hand gear I am always on the lookout. My shed is so full of equipment for those "just in case I need it" sort of jobs. I just threw out an old automatic sliding door mechanism because it took up too much space and I just didn't have the time to rework the controls to something I could use. I will post some updates shortly. I will update as I get along, but things will probably progress slowly. Can't rush these things .

Hi @dontcarefilmer, I may be able to assist you somewhat with required specs for a full scale version of your model if it’s any help. I would really like to see this built because I had just about everything worked out for a full scale version of this design about 5 years ago. The only difference with my design was I planned on mounting the roll motor on the outside frame instead of inside the roll cage so I didn’t need to run high current thru a second slip ring.

I think you will find that this will be much more difficult (and expensive goes without saying) to turn into a full-scale version than it was to make a model of. I spent over $1000 acquiring motors, gearboxes, gears, and power supplies on ebay only to abandon the idea myself. But this was mainly due to the fact that I don’t weld and, in the end, I didn’t think wood would be suitable for a frame and roll cage. Also, I hope you have high ceilings because you will probably need them to be taller than 8 feet (that’s the standard minimum ceiling height here in the US). And one thing I don’t particularly like about this design is, to keep the pitch axis in balance, you will inevitably need to position the pilot somewhat behind the pivot point, therefore moving the pilot downward when he should be going upward or not at all and vise versa. Since I didn’t built it I can’t say if it will make much difference in how it feels though.

While I may have over designed, I purchased two 24vdc motors, a 1.3hp 3600rpm motor for roll, and a ¾hp 1800rpm motor for pitch. The roll motor would be attached to a 25:1 reduction gearbox and the pitch to a 60:1. I determined that I would need around 30 rpm max for roll and 8 for pitch if I recall correctly. Therefore, for final reduction, I planed on using a chain drive between my gearboxes and a large drive sprocket, one on the roll cage, and the second on the pitch frame. The drive sprockets I purchased where 60 tooth with a 100mm hole in the center. They have 8 boltholes around the perimeter of the inner hole for attachment to the frame and the large hole thru the center allows for passage of a slip ring and axle. In fact, here is a picture of one.



Then, I would use smaller sprockets on the gearboxes to achieve the proper final reduction for roll and pitch.

The main issue with this design, the way I see it, other than size, is the necessity of using slip rings to transmit power. Since there is good change of getting interference on a circuit, it’s not wise to transmit video and the like on one to my knowledge. Therefore, I planned on putting my computer in the roll cage with me. Have you made any plans yet concerning this? This may require a SSD instead of a hard drive. Do you know how much current your homemade slip rings can handle or will you purchase some instead for your full scale version? How many circuits will you need on your slip rings in order to power everything? Where will you be mounting your motors power supplies, motor drivers and arduino relative to your computer and how will you able to control them from it. What kind of axles and bearings will you use? These are a few of the questions you will need to determine answers for.

Wow, what an awesome reply @BlazinH. Thanks for the info, it will probably take me a while to digest it all.

I'm sorry to hear that you didn't finish your full rotation design. Did you end up building something similar? Did you end up reusing the equipment that you already purchased?

As for my full scale I haven't quite settled on motor sizes just yet as it will depend on the weight of the frames. My brother and I are planning to build the frames out of aluminium. He is a blacksmith by trade (yes they still exist ) and so he is quite good at welding.

Based on my own experience I planned to use 415 Volt AC motors to achieve motion. This would reduce the current flowing through the slip rings. I may use industrial inverters for control. In my field these drivers are typically located on or as close as possible to the motor itself.

May I ask why you opted for 24V motors? My experience with motor selection is rather limited and I would like to learn more.

As for more specs on my plan, I can post those when I get home from work.
Cheers.

Ok, some more info about what I currently have planned for my full scale:

Most of my plans to date have been designed from an electricians point of view, so details like the bearings and shaft sizes are yet to be determined. I expect those to be relatively straight forward once I know my expected weight and desired acceleration (probably my naivety showing a bit here). Same goes for gear ratios, desired kilowatt ratings on the motors, and mechanical power transmission styles. I like your idea with the sprocket and chain system to move the rig around. For my prototype I opted for meshed gears only because of simplicity, and belts seemed to be out of the question. How did you determine the number of teeth per sprocket, is there a standard set of sizes to choose from, and are the teeth rated to a certain amount torque? These are some of the things I really need to research more.

As for slip rings, ideally they will transmit 3 phase power, neutral and earth to the centre rig, plus a certain amount of control cabling. The control cables will need to drive/control the pitch motor on the outside, plus return any feedback. This control system will likely be a form of bus network to accommodate both motor control and feedback. I haven't quite decided yet but I was leaning towards modbus over ethernet, or a version of canbus. Both are quite robust and tolerate the expected noise on the slip ring system. It all depends on how well I can build a larger slip ring, which I hope to be easier than building a small one (more room to work with, less fiddly). The current requirements are yet to be determined, but I reckon I can pass up to 30 Amps without too much hassle. Heat will be the biggest hurdle.

Inside is where I will house the roll motor and inverter (or driver), feedback devices, the main controller (arduino/ PLC/ other microcontroller), the PC with all peripherals, and of course the passenger. The PC will need to have an SSD as @BlazinH suggested, and I am planning to use the oculus rift as well.

Outside will be the pitch motor and inverter (or driver), feedback devices, and the main power supply.

I believe there are many different ways to accomplish the end goal (FUN!), but this way is much more comfortable for me as it draws on my experience as an industrial electrician. I am interested to see some of the variations on control out there, that's what interests me the most.

To answer your initial questions @dontcarefilmer :

No, I didn’t. Besides not being a welder, after much thought, I decided I must have heave also. Unfortunately, this design is not heave friendly. This is what I ended up building instead.


However, I no longer use this simulator. It was designed for flight but I modified it to be faster for race as in the video. But for race, you really don’t need to move the monitors, so I built an even faster simulator for race that uses detached triple 61 inch screens atm. One day I may un-modify my old simulator for flight again. It still works as is but is not as smooth of movement as it once was due to faster gear ratios used for race.

Not yet. I thought about selling some of it but I’m a pack rat and hate to give stuff up because I still could have a use for it some day.

Well, at the time, I didn’t know anything about controlling ac motors with VFD’s and brushed dc motors are quite easy to control. And I scored a couple of new/old stock 24vdc motors as well as some gearboxes on ebay for a reasonable price too. Those large drive sprockets cost me $90 a piece though but that’s partially due to the fact they are new and made from stainless steel (original use for them was in a poultry processing machine). Also, I’m not keen on the idea of using high voltage ac for a project like this myself. We only have 230vac standard here for household use but +400? It appears you know what your doing with electricity but still, it could be deadly! You may want to consider putting your VFD’s etc. on the outside instead of the inside for safety reasons.

I’m not an expert on chain drive systems but since the large drive sprockets I purchased were size 35 that was what I was going to use. That’s probably beefier than necessary though. The reduction ratio is easy to figure; # of teeth on large sprocket : # of teeth on small sprocket. For example, my roll motor is 3600rpm / 25:1 gearbox or 144 rpm. To reduce it to 30 rpm maximum I need a final chain drive reduction of 144 / 30 = 4.8:1. So with my 60 tooth large sprocket I would need a 60 / 4.8 = 12 or 13 tooth small sprocket to put on my gearboxes output shaft.

If your interested here is a link to the slip rings I was planning on using. http://www.ebay.com/itm/Slip-Ring-w...020?pt=LH_DefaultDomain_0&hash=item3f305b56d4.

If your design is well balanced, it won’t take a lot of power to keep it moving. But you still need some fairly large motors if you want to have a fast acceleration rate for aircraft like fighters, racers, and acrobatic planes.

I look forward to seeing more details from you.

@BlazinH : I like your simulator!
Why did you say it is not heave friendly? It looks really bumpy and synchronized with game

That video of yours @BlazinH is pretty sweet, such a large screen on a moving platform. I'd be worried about breaking it . I'm quite a ways off what you have accomplished, but i'll get there one day. Well done!

Just to address some of your electrical concerns/comments:
In my industry we usually install the VSD's close the motor to reduce the length of shielded cable required. Shielded cable is required (as far as I know it is law) here to reduce the harmful radio frequencies which are generated on the load cable. To best of my knowledge, the radio waves can interfere with pacemakers when the motor draws enough current and the frequency of the drive is high, so shielding is the only option. I'm not sure how well this will translate through a slip ring assembly, I've never installed one in the field in this fashion before. If the VSD is for a very large motor (~say 5kW / 6.7HP) then we install it in a local control panel.

Here in Australia we too mostly only have 240V for households, but there are many places which have 3 phase air conditioners, and so have access to 415V. Because I tend to work with industrial machines, I have been leaning towards the higher voltage motors for my design. The risk associated with this kind of voltage is overrated in my opinion (maybe that's a bit biased though). There are many good products available to protect the people in and around these environments.

Thanks for the ebay link, I tried to find units like that in Australia but without much luck (including ebay). But to be fair I was only trying to build the prototype at the time, and so my mind and wallet were focusing an slightly different models. I will check these guys out further for the full size sim. No way can I build a similar slip ring for that kind of price, definitely worth buying some.

I really need to study up on the mechanical aspect to this type motion sim. You spoke of needing 30RPM and 8RPM for your initial flight design, may I ask how those values were calculated? I have no reference for how fast my rig will need to go, and calculating the required torque seems to be a headache, any help would be appreciated . I noticed there was a "Motor Actuators and Drivers" forum on this website, so I think I will go for a nice long browse to see what I can learn.

Ok, enough of my rambling.
Cheers.

240v motors a bit slow really for race, but good for flight sims. Maybe why we use 12v dc motors at about 150 rpm at shaft. With a 40mm CTC lever, it translates to about 300mm per second travel. The SCN actuator sims run at 200mm per second. Thats nice and snappy for race.]
However , for this prototype, no choice but 240v, so it will still do the job. Maybe add bass shakers for finer motion vibration.
Dave.

I'm all for 12V motors if I can get the power out of it, but my knowledge and experience with these types of motors is limited. Is there a standard range in the 12V series as far as speed and torque are concerned? Or are there too many different types? I've been a bit overwhelmed with all the different kinds of motors available. Is there a motor selection guide out there?

While my initial interest is for flight I don't want to limit myself to it, unless I don't have much choice with this style of motion sim.

I have an idea in my head about how a quality ride depends on how snappy the motors are as @bsft describes, but does anyone know how response times influence motor selection? Does it come down to the torque, the power, the driver or something else?

Cheers.

Thanks @RacingMat. What I meant was that a continuous rotation design is not heave friendly since a heave effect requires gravity to work in a simulator. Once you are past a certain angle off level, the heave motion will start giving incorrect feedback to the pilot.

Thank you for your kind words too @dontcarefilmer. You definitely need shielding on your cables to prevent interference with you feedback devices if nothing else. If you are comfortable using high voltage ac then using it is your prerogative. However, if you were to attempt to commercialize it, it would be much more difficult to pass safety requirements and more expensive to obtain product liability insurance than when using dc power.

Since the slip rings link I posted shows several sizes and numbers of circuits available, I assume that $115 is the price for the smallest one with the least amount of circuits on it. It is possible that the largest one with the most circuits could run 2, 3, or even more times the price shown. I never did contact them for pricing.

How did I calculate the roll and pitch rpms required you ask? I used a very scientific method! I fired up FSX and loaded the fastest rolling and looping aircraft (pylon racer) and timed how long it takes to do the tightest loops and the fastest rolls in it. Lol. If you have an abundance of power and torque available I suppose it would be safer the make it even a little faster though in order to make sure its fast enough.

Calculating the exact torque you will require may be a little difficult but it shouldn’t be very hard to get it close enough once you can supply the required data. I think I can help you out with that then. But it’s easy to calculate what torque you can achieve with a motor and gearing if its torque is listed on the motor. My 1.3hp 3600rpm roll motor has 22.8 in. lbs. of torque. With my 25:1 gearbox that multiplies the torque by 25 so 22.8 x 25 = 570 in. lbs. With a further reduction of 4.8:1 with my chain drive I would get 570 * 4.8 = 2736 in. lbs. torque total. Dividing that by 12 will convert it to ft. lbs. so 2736 / 12 = 228 ft. lbs torque @ 30 rpm maximum. What that means is that if I were to use a 1 foot lever or a 2 foot pulley or sprocket (1ft radius) at the end of the reduction, then it should have the required torque to lift up to 228 pounds. With a 6 inch lever or a 1 foot pulley, it would be double that or 456 ft. lbs. torque. With an 18 inch lever or a 3 foot pulley, it would be 2/3 what it would be at one foot or 152 ft. lbs torque. I don’t know exactly what unit of measurement nm is and how it compares with ft. lbs. but if needed we can convert between the two with an online calculator to figure out what you will need. But, with a well balanced simulator, you will not need to lift anything, but will only need to overcome the inertial force required to get it moving and then keep it moving. With this situation, 228 ft. lbs. torque is more than adequate!

And not to throw a wrench into things, but there is something else very important that has not been discussed with you yet! Beside the challenges of building the continuous rotation simulator itself, there is also the issue of “flip-over” in the roll and pitch axis data coming from the simulation itself and from the feedback devices also (unless you use an incremental encoder that is)! Since the roll and pitch data represents what you would see on an aircrafts interment panel, whenever you make a full rotation, the gages output (that’s in degrees) flips back to 0 or 360 instead of incrementing. I have already posted some code examples here somewhere on how to test for a flip over condition and how to deal with it (I will try to find it).

You can use hall effect pots but they flip over too and have a 1 or 2 degree deadband when doing so from my testing. So far, no one has posted a solution for this problem yet to my knowledge. I believe I deduced a solution for this recently though. If it is possible to find a way of connecting two 360 degree hall effect pots to the same axis, and they are set 180 degrees out of phase from each other, then before a hall pot gets anywhere close to experiencing a flip over, you will start using to other pot. So, both pots would actually be outputting 90 degrees to 270. But since they are 180 degrees out of phase from one another, they will cover the entire rotation. If you want to try this I can provide more information for you on how to do it. It is not without some drawbacks though because it would require more resources from your microcontroller.


Ps I was wondering why bsft said, “no choice but 240v” myself, but I assumed he only meant because that, or more, is what it sounded like you were going to use regardless.

@BlazinH , I am probably wrong and large DC motors, 12, 24, 48v would do the job as well. In Australia , out friend here @dontcarefilmer has easy access to 240v, so that may be of option to him.
He also is looking at a flight sim, so maybe this frame http://www.acesim.com/main.html is a cheap option to start in flight and 12v motors would move it, granted its not 360 deg roll though.
Oh and good info there you have added blazinh.

if you are after flight , then slower motors would be fine.
Generally a 2DOF balanced frame doesnt need huge power to move it, but when you go to 3DOF lift and or 360 deg roll 2DOF, then you need to look at more powerful motors and expense.
It does come down to sim design to suggest the best motors to use.
Up to you of course which path you take.

Just so you have some idea of what I purchased for reference, this is the kind of motor I bought for pitch on my continuous rotation but in a 24volt version. http://www.ebay.com/itm/3-4-hp-1800...680?pt=LH_DefaultDomain_0&hash=item232718e698 4 or 5 years ago after the economy turned, there was a ton of surplus new/old stock motors and gearboxes available here in the US (but not so much now). They are pricey now too! I bought my motors for less than $200 dollars apiece.

This is an example of the gearboxes I bought but not the same ratio of this one. http://www.ebay.com/itm/New-Boston-...629?pt=LH_DefaultDomain_0&hash=item2eb41b3945 I got both of mine, again, new/old stock, for around $100 dollars each including shipping. The gearboxes themselves are stainless steel models and retail for over $1000 each so I got a real deal on them. The gearboxes have a keyed shaft for attaching sprockets etc. They can be quite heavy though. One of them I bought was $50 but cost $60 to ship it! It was the pitch gearbox that was heavy due to a 60:1 reduction and high torque output making it beefy. But, since that one is stationary, it causes no problem. Since the roll motor, gearbox, etc., will need to move with the frame, weight is more of a consideration with it. My 25:1 gearbox is much lighter (and smaller).

thats a NUKE POWER motor for sure!

Hi, it is Newton.meter
10 Newton is the force you need to lift 1 kg
( i.e 1 kg x 1 g = 1 kg x 9.81 m/s² = 10 N)

If your motor has 12 N.m torque, you can lift a 1.2 kg with a 1 meter lever
or 12 kg with a 10cm lever.
Quite easy as well!

@BlazinH : I understand now. I thought you spoke about your simulator in the video... but you were talking about 360° simulator!

I love how you calculated your roll and pitch rpm @BlazinH, simple but clever. I laughed out loud with appreciation when I read this. I will also try this method myself as a starting point.

Thanks for your detailed explanation of torque related calculations. Your summary was well written and easy to follow. This gives me something to play with at home regarding my own designs. I like learning .

With regards to "flip-over" data from the software, I anticipated that was going to be a problem. To overcome this I wrote a sketch for the arduino which takes positional data from the serial line (at the moment it's just hand typed into the serial monitor) and calculates the shortest route. So if I am at position Roll 350 degrees and want to move to Roll 010 degrees, the prototype drives 20 degrees clockwise without a hiccup. I am only new to programming, so I'm quite proud of this code . I am not sure exactly what kind of data Simtools sends to the controller, but if it is just positional I think this should work OK.

As for feedback, my experience is again limited. Until now I have been driving stepper motors for my small prototype, and so all my tracking is done via software. Not great I know, but it works for a demo . I like your suggestion with 2 hall effect pots, quite novel. I had considered an incremental rotary encoder for this task, but it takes up valuable interrupt pins on the arduino. I had also toyed with the idea of using an accelerometer / gyroscope package to roughly track position in tandem with incremental method. Hopefully this would eliminate tracking drift in the software (if there is such a problem). See link: http://www.ebay.com.au/itm/200949159205?ssPageName=STRK:MEWAX:IT&_trksid=p3984.m1423.l2649
If anyone else has overcome this problem I would like learn about it too.