Need help with 4DOF (FD301/401) like simulator

Hello everybody,
I just wanted to say this site is freaking amazing!! :yippiee: And I'm impressed with all the resourcefulness and ingenuity of the members of this community. :thbup:

I stumbled onto this site about 3 weeks ago from a video I saw on youtube. Needless to say my interest was peeked when I realized there was open source software available to use in my own DIY simulator project. It took me a few days to decide what kind of simulator I wanted to build, and after searching this forum and reading post after post for 3 weeks straight, I think I have a fair amount of knowledge to start my own project. I decided to go with a simulator similar to Force Dynamics 301/401 just because of the sheer awesomeness that this simulator can dish out! I really like the over exaggerated movements that this simulator provides over the SCN5 type simulators. I do realize however, that this type of simulator will cost more than a standard 2DOF, and so I plan on dragging this project out over the length of a couple years to be able to source out cheap parts and find good deals on the required hardware components.

I am somewhat disappointed by the lack of support for 4DOF type simulators like the FD 301/401. Being a newbie here, its hard to find references or some sort of guide to go by to build my own. I have only found 2 projects that are similar in nature to the 4DOF, but they all seem to stop short of completion and don't really go into detail as to the exact components that are used and compatible with X-sim.

I have a few questions that I would like help with, because I can't seem to find a definite answer in the forums.

1. I'll be using 3-phase AC motors with 25mm diameter, 10mm pitch ball screws. 25mm linear shaft will be attached to the ball nut much like the FD 301/401 design. Would it be better to use a brake motor or a regular motor? Would the ball screw back drive at a 10mm pitch?

2. Other than the Omron J7, what other VFD's are compatible with X-sim ? What features should I look for when purchasing a VFD for X-sim?

3. Which motor controller will give me the fastest, most accurate results? I've read that the AMC1.5 is probably the best around right now, but will it be closer to the performance of the FD 301/401 versus using the VM140?

4. I can't find any good information on which encoders work with X-sim. I see some people using HI-RES encoders like the MAB40, but how does it interface with the AMC1.5? Do I need an analogue or a digital signal from the encoder to make it work?

5. I'm also having trouble trying to wrap my head around the basic flow of how motion simulators work. As I understand it:

(Game PC data)----->(AMC1.5)----->(VFD)----->(Motor)----->(Encoder feedback)----->(AMC1.5)??

Please verify or correct my thinking on this.

6. In other posts, I've also read similar explanations about how the flow is supposed to go, but everyone uses different terms to describe it, which makes it confusing. What does the Profiler do and how does it work with X-sim and the AMC1.5?

7. I'm also assuming that X-sim isn't the only piece of software that I'll need to download in order to make this simulator work. What other programs do I need?

Sorry for the long post, but I've done so much reading over the last 3 weeks, I'm bursting with questions! :?

I appreciate any help you guys can give and I look forward to your advice.

Awesome website!!! :thbup:

Hello and welcome.
4DOF units are probably out of the price range on even us hobby guys to develop and build as well. Still, here a good place to start my-3dof-project-is-born-t3096.html?hilit=aldoz , go through the thread. He uses 12volt winch motors as the 3 DOF. I guess its just a matter of putting a rotating base underneath such a design. And he uses the Pololu 12v12 JRK boards.
The x-sim program should be enough to need.

I do realize that these types of sims can get a little pricey, but the way I see it is, if you're going to do it, do it right. Go big or go home! And I just wouldn't be satisfied with a 12v dc motor setup. Cost really isn't a factor here. Now I'm not saying that I'm made of money, but if you look hard enough, you can find what you're looking for for very reasonable prices. That being said, I've already acquired a 25mm x 700mm ball screw and nut with a 10mm pitch, 2 heavy duty end supports for the ball screw, a 1000mm x 25mm linear shaft and two 25mm linear bearing blocks for the shaft, all for around $130. I am also going to acquire a 2HP(1.5KW) 3 phase AC motor for $50 next week. All I will need now are a few pulleys, some scrap metal, an encoder, and a VFD. All which should cost around another $230. Aside from building an AMC 1.5, I'll be ready to start testing the actuator. Considering the cost of the parts so far, I think its very reasonable to build a FD301/401 type simulator that compares to the commercial models. Granted, not everyone can find good deals like these for their DIY actuators, but $410 per actuator is pretty awesome! :thbup:

So I guess what I want to know is, if money didn't matter, or if you had more money to spend....how would you go about building one of these 4DOF sims? What type of actuator would you build - Toothed belt driven or Ball Screw? Which VFD to buy? What type of AC motor - brake or no brake???

I've seen where Jyrki was attempting to do the same thing I am, but I haven't seen anymore progress from him lately.

I know some of you guys in here have knowledge about these types of things and the answers I'm looking for. So if you were going to build a 4DOF like mine, how would you do it?

I mean, how awesome would it be for someone in this forum to finally build a working 4DOF like the commercial models for thousands of $$$ less and also using Xsim!! 8)

hi
A safety brake will always be better
Dont forget : You need continuous stalltorque to hold the system
with a 10mm pitch you need around 2NM stalltorque to hold 100kg
Seak for a fast controller (updates second counts)with encoder input.
ps. reading at youre parts ,i cant understand how youre gone make your actuator?
regards Ad

I don't quite understand how you calculate the continuous stall torque of a motor. I realize that the stall torque is the torque required to stall the motor so that the rpm's are at zero. But given the calculation for torque of a motor, with a known HP or KW value:

Torque = (5250 x HP) / RPM =
Torque = (5250 x 2) / 1 ----(can't do zero, because it would equal zero) =
Torque = 10500/1 =
Torque = 10500 ft-lbs (14236 Nm)

It just doesn't make sense to me. That sure is a lot of torque!

Adgun, you said I need about 2Nm of stall torque for 100Kg(220 lbs), which is basically the same load that my actuators have to support anyway. 2Nm is equal to 1.475 ft-lbs, which is roughly the same amount of required torque that I calculated to move each actuator with a load of 200 lbs(90Kg). From this information, it looks like the torque to move the load and the stall torque are the same thing. :?

I did calculate the required torque to lift my cockpit (per actuator). Roughly my cockpit with a 250 lb(113Kg) rider will be at max 450 lbs(204Kg). I know you can't just divide the total weight by the number of actuators because it won't give you the true amount of weight that each actuator should be able to handle, so I divided by 2.25 and to give room for error and came up with approximately 200 lbs(90Kg) per actuator. In order to lift 200 lbs(90Kg) with the ball screws I have, the motor needs to be able to supply approximately 1.3 - 1.4 ft-lbs(1.76 - 1.89 Nm) of torque. By this information, all I would need would be a 3/4 HP(.56Kw) motor turning at 3000 rpm's. The slower the motor turns, the more torque it puts out. The 2 HP motor I'll be getting should easily be able to do that.

So from all of this information, is my motor's stall torque sufficient enough for this application or am I missing something?

I've also attached a rough drawing of the actuator I plan to build so you guys can understand what I'm trying to do a little better.

(Edit: Don't know how or it won't let me attach an image file. I guess I'll figure it out later.)

Are you asking for help to design and build a simulator for yourself?
Or are you planning to use this information to have commercial units built for re-sale?
Your questions are suspicious.

No, no... this is just for me. I have no intention of re-selling these. I'm always looking for some sort of new and challenging project to start up. Its a hobby of mine. Always building things....My last project was a full size arcade game equipped with light guns, trackball, racing wheel and pedals, and over 7,000 games to play. Its a beast of a machine!

I've always wanted to have one of these simulators since I was a kid. And I have the means and capabilities to finally build one myself. I've got the design all worked out in my head for the most part. I'm just having trouble with a few calculations of certain things, and exactly what hardware components are compatible with Xsim.

I just want to be able to build the best possible simulator I can build....you know what I mean... Sorry if I sounded suspicious, but rest assured I have no interest what so-ever in making money off of these. I just want something to do.

Maybe build a low cost 2DOF to start with, so you can get the idea how it works, and this would allow you to get used to the software and its running. Just design the build so you can expand to a 4 DOF later.
Start small and reduce overall cost.

Here is a rough drawing of the actuator I'm trying to build. Hope this gives you guys a better understanding of what I'm trying to do.

Build it and give it a run...thats the way to see if it works.....practical, not just theory.

Here's some of my ramblings about the subject design-of-a-timing-belt-actuator-t3232.html

Hello Jyrki,
I was hoping I was going to hear from you sooner or later. I've read all of your posts a few times already, and I'm still trying to figure out how you calculate stall torque. Very interesting ideas by the way. I kinda understand, but I was wondering if you could explain it too me a little better.

For example:

My specifications and payload are very similar to the calculations you came up with in this post - suggestions-for-d-box-type-home-theater-simulator-t3519-40.html

But the thing that confuses me is that the stall torque you come up with is almost the exact amount of torque that I calculated I would have to have from the motor to lift each actuator. So from this, it would seem that stall torque and lift torque are the same, which I know they're not... :? :?

Also, how do you know how much stall torque a motor has? Have you done any more work on your project? What kind of motor were you planning on using?

Sorry for all the questions, I just want to make sure I get the right part before I spend the money.

Thanks again for all your advice.

The needed stall torque is simply the amount of rotational force needed to keep the axis in one place and is exactly the same force as gravitation excerts to the arm e.g. F=m*a (plus some more as some of the force is lost in the system, remember efficiency)

Torque is quite linear in relation to speed and as the rpm increases torque decreases and vice versa, mostly it can be said that stall torque is higher than the running torque.



If you get stall torque to be the same as force needed to lift the arm then you might have forgotten to add the force pulling the sim down on your equations, e.g. force needed to lift sim up is F=m*acceleration of gravity + m*wanted acceleration up. On my examples I calculated with wanted acceleration to be 1G or exactly the same amount as acceleration of gravity.

Summa summarium:

stall torque > F=mass on one axis * 9,81m/s/s
maximum force needed to lift F=mass on one axis * 9,81m/s/s PLUS mass on one axis * sims acceleration

I'm planning on using some chinese ac-servo motors and drivers, I once looked some up and found out that with 1500€ I could get 3 servos (some 750w ones) and drivers. If my memory serves me right they were some 2,4Nm motors (peak torque is higher as always, but motors aren't supposed to be driven with peak power all the time)

http://www.yaskawa.com/site/dmservo.nsf/link2/NKOE-7TKLD2/$file/PR.DN.01.pdf

Thanks Jyrki
I think I've got it figured out now. I did some calculations last night and here's what I came up with.

Force on each axis = 90Kg(~200lbs)
Force of Gravity = 9.81m/s^2
Lead Size of Screw = 10mm(~0.4in)

Force on Axis at rest = 90Kg*9.81m/s^2 = 882.9N

Force on Axis up while accelerating at 1G = 882.9N + (90Kg*9.81m/s^2) = 1765.8N

Torque to lift axis up at acceleration of 1G = (1765.8N*10mm)/(6280*0.8eff) = 3.51Nm(~2.6Ft-lbs)

Stall torque = (882.9N*10mm)/(6280*0.8eff) = 1.75Nm(~1.3Ft-lbs)

Preferred Axial Velocity = 400mm/s (~15.7in/s)

Required RPM = 400mm*60/10mm = 2400 RPM


Required motor HP = (2400rpm*2.6Ft-lbs)/5250 = ~1.2HP (895W)

So a 1.2HP(895W) motor at 2400rpm's should give me 2.6Ft-lbs(3.51Nm) of torque at an axial velocity of 400mm/s. In all reality though, I would go with a bigger motor of course because I wouldn't want to run the motor at max capacity for fear of shortening the lifespan. If I went with a 2HP motor, the true HP would be:

This is based off a motor I'm looking into purchasing:

Volts*Amps*Efficiency% = 208v*6.15amps*0.8eff = (1.37 True HP)

So I might even need to go with a 2.5HP so there won't be as much stress on the motor. What do you guys think?
Also, based on my calculations (if they're correct), would you think the motor in question would have enough stall torque for my application? I still don't know how to find the stall torque of a given motor other than looking at the data sheets for the motor. And I'm very skeptical of used motors having data sheets still intact.

So after doing some reading, I found out that Inverter Duty rated motors are more efficient to run with VFD's than standard general purpose motors. They are capable of withstanding higher voltage spikes without their insulation failing and can deliver 100% continuous torque at low speeds without over-heating. It's like these motors were designed to be used for simulators!

Anyway, I found a motor here: http://www.globalindustrial.com/p/motor ... e-1180-rpm

I talked to a sales rep already just to make sure that the price wasn't a typo, and its not. This has got to be the cheapest Inverter Duty rated 3HP motor I've found so far. The data specs are here: http://www.globalindustrial.com/site/im ... 10-1VT.pdf and the model number you need to look for when looking at this sheet is U3V3BC. It doesn't say much about it, but this motor is 90% efficient at full load! And the true horse power for this motor is 2.4HP(1.8Kw)!

The only thing I don't like about it, is that it weighs 110lbs(50Kg). Other than that, do you guys think this would be a good motor for my application?

I'm kind of drunk at the moment so I haven't gone through your calculations thoroughly but they seem to be in line what I got with my 5mm screw pitch (your's is around x2 the torque)

I would recommend against normal squirrel gace ac-motor as I've noticed that they tend to give poor performance while running at low rpm (5hz seems to be the cutout speed on some that I've run on inverters) and that motor is rated only at 1180 rpm, you would have to push the inverter with higher freqyency to make it go fast enough. Because of the poor performance at low speeds it might result in your sim not keeping in one place when stationary. AC squirrel gage motor needs to be lacked behind the rotation of magnetic field to produce any torque and at low speeds it jumps or jerks and doesn't give constant torque but falls to zero and then again it tries to drive the shaft.

You are better of using a real servo motor as VFD with position control are quite expensive and even then it would be like pushing something with a rubber band then again there are thesis on making them act like real servos, but they aren't meant for that.

It's in my opinnion a bit marginal issue if motor is VFD rated or not, they all have more than enough of insulation on their windings but the bearings might suffer because of the induced higher freqyencies (or that what I believe is the real issue).

One more thing against normal ac-motor is the inertia of the rotor, on real servos it is designed to be very low but squirrel gage motors have a lump of aluminium and copper cast that have high inertia. High inertia means you would have to push the motor with greater power to gain the same acceleration. On real servo motors the rotor is hollow and has low inertia.

So what you're saying is at 5Hz, the motor just stops producing torque and whatever load you're holding up will begin to fall? Is it also safe to assume that any 3 phase induction motor uses a squirrel cage rotor, regardless if its an Inverter Duty rated motor or not?

I'm a bit confused on this subject as well:
I realize that speed of a motor is dependent upon voltage and frequency, so would you say a motor of 3000 rpm base speed at 5Hz is faster than a motor of 1180 rpm base speed at 5 Hz?

The specs for the motor I referenced say:
10:1 (6-60 HZ) Speed Range Variable Torque
5:1 (12-60 Hz) Speed Range Constant Torque

How slow will the motors need to turn during operation of the simulator anyway? Lower than 6Hz?

This motor could be fitted with a brake to eliminate loss of torque at low frequencies while trying to hold its position. But if I'm going to do that, I might as well just buy a brake motor instead that already has the brake built-in.

Now, you recommend servo motors, and I did some looking to find out what kind of motors that Force Dynamics and Motion Sim use on their Sims. Motion Sim uses non-synchonous (or Asynchonous) motors and cogged belt gearing, which leads me to believe that they're 3 phase induction motors with worm gears and timing belts. Force Dynamics uses three-phase brushless permanent magnet motors with a spring actuated brake. They also use over sized ball screws. I'm assuming that the brushless permanent magnet motor is DC. So my question is, is X-sim able to control this type of motor with a digital servo drive? I know it can control VFD's using the AMC 1.5, but is the AMC 1.5 able to control other drives as well?

If it can, than I found a servo motor on ebay that would provide more than enough torque for my sim. http://www.ebay.com/itm/YASKAWA-AC-SERV ... 33686f199f

The only problem is, I haven't found a drive for it yet that has enough power to run the motor.

Jyrki, I know you originally started your project using the geared belt design. To me, it seems like this would be a cheaper route, and the AC motors would handle better at low speeds because of the gear reducer on the motor. I don't know, maybe I'm wrong. I may go with the belt gear design if I can't get the ball screw to work, but I've got to decide on a motor first! :?

Motors need to run at zero speed and produce torque at the same time. While driving along a race track at even speed the actuators would stay put...

All induction motors are squirrel gage motors and 3000rpm motor will run faster ofcourse than lower rpm motor when driven with the same frequency.

You will need to cook your own motion controller for this thing anyway there's no shortcuts. Here's some servos that I looked at http://www.zappautomation.co.uk/index.php?cPath=36_118 There are some cheaper ones too on the net if you dig deep enough.

Only reason I didn't take the belt route was that those gearings cost a lot of money and with the same price tag of 3 reduction gearboxes, 3 ac-motors and 3 VFD's one can get 3 ac-servos and 3 drives for them.

Thanks for all your help so far Jyrki. I hope to help others one day with all this new stuff I'm learning in this forum.

Now when you say cook my own controller, do you mean to build a motor drive (AC Drive) or are you talking about building something like Thanos AMC 1.5 board? Because I'm kind of excited about building the AMC 1.5. I've never etched a PCB before, so that will be a fun skill to learn.

Back to the belt design...I can get gear motors around here real cheap. For example: http://www.dougdeals.com/p-7004-sumitom ... r-101.aspx. I just never looked into them because of the ball screw design I originally intended to go with. But, if price wasn't a factor, would you go with the belt drive or ball screw? I'm kind of leaning more towards the belt design right now. They are quieter than the ball screws too.

I've already got the calculations figured up for pulley size, linear speed, reduction ratios, and rpm's. The only thing I can't figure out is how you got your original calculations for the torque needed on your belt driven system. I've looked all over the internet all day today, and couldn't really find the formulas I need to calculate it. I would assume the force on the up axis is the same for the ball screw, and that you have to figure in the pulley radius somehow... :? :? Now, I do know that pulleys are a little more efficient than ball screws, but not by much, and I'm also guessing that the torque required on the axis up will be more than it was for the ball screw design. If you could show me how you got your calculations, I would be very grateful. :thbup:

Thanks again.

I made a simple Excel sheet for my initial calculations, it's not a big secret here's the sheet http://www.netikka.net/aros-stuff/Simulaattori2.xlsx or one version of it. There may well be errors on it so don't take it too seriously. In this version there is some sort of calculations for gas spring aid, but the results are pretty odd and the calculations on that part are probably flawed.

The pulley for which there is pitch diameter in the sheet is 22 teeth HTD pulley which I planned on using. It's same same one os on my belt drive thread.

What I meant cook your own controller was that you need to somehow translate the data that x-sim export to what ever you are trying to control. For servo drives that would be to translate position data of axis into step/dir pulses for the servo drive. I'm planning on using a 32bit microcontroller on USB that listens to x-sim and controls the actuators (step/dir or whatever it will be)

Belt drive is far easier to make than a ballsrew one, I'd go with the belt drive if I could source those gearings.