Suggestions for D-Box Type Home Theater Simulator

Here's the screw and as can be clearly seen there's not much meat at the place where the bearing goes (just 1mm on either side and just in one place where the thread ends)
[img600px]http://www.netikka.net/aros-stuff/DSC05509.JPG[/img600px]

I'm planning on doing something like this, but with a bit thinner(shorter...) sleeve. It gets welded on to the screw. I'll drill some 3 holes on the sleeve and weld it in place from those holes.
[img600px]http://www.netikka.net/aros-stuff/DSC05512.JPG[/img600px]

The bearing surface is pretty bad as I grinded it on a mill... This one's just a prototype, I'll take the real screws to a shop and they'll put them on a lathe and make threads for the locking nut as well.

I grinded the flange off from the ball nut, it really looked like this when it arrived http://www.zappautomation.co.uk/product_info.php?cPath=1_63_64&products_id=153

I think I'm going with the screw design, but I'll take a look if those planetary gears could bear the load that I calculated for the belt drive system, but thanks anyway!

I bought the screws from a member of Finnish CNC forum and they are good quality, they were originally from his workplace, but were discarded. Nut is some Chinese brand and a bit noisy, it also misses some 16(!) balls, I'll have to order them and reball the nut. Quality wise it's not the best, but was pretty cheap...

This forum is so awesome! adgun, you have been especially helpful. Thank you. Thank you. Thank you. You obviously know a lot about this kind of thing. What do you do for a living?

jyrki.j.koivisto, thank you for sharing. You did a good job grinding that flange off. I didn't even notice at first. What is your project? Do you have another thread? I will be very interested to see how it progresses.

I am really enjoying this. This is going to be a really fun project. With this information, I will re-work my design which I'm sure will create more questions.

Mike

At this point, it is clear that my system will easily exceed D-Box's performance in pretty much every way. It will handle more weight, travel farther (I still want to limit the travel to something fairly small) and be faster. I want to make it as good as possible within reason. I know it will be over-engineered in some aspects and that is OK. I think I can calculate the weight stuff pretty easily but I'm kind of scratching my head with speed or travel.

How fast should I be shooting for? I want it to be responsive but I want to be realistic about the speed I can achieve. Speed is based on RPM (and we are limited to 3,000 with ballscrews) and the screw lead right? I don't want it to lull people to sleep. Any ideas?

D-Box travel is 2 or less. I was thinking of possibly taking this to as much as 6 but wasn't sure if that is too much or too little. I'd rather not have to install seat belts...then again maybe I should plan a seat-belt mode for the adrenaline junkies. Basically, in normal use, this shouldn't really be thought of as a ride as much as mostly just adding subtle motion cues that will not be too distracting from the film itself. I welcome any input.

Mike

I suggest you put your design to a Excel spreadsheet. That way it's easy to spot what happens when you change one thing to another, the math is pretty simple.

For example:

Axis velocity (maximum speed target) 250mm/s results in (250mm/s)/(Screw pitch of 5mm)=50rev/s on the screw or (50 1/s*60s)=3000rpm

With that you can calculate the needed torque on the screw (remember that the screw isn't 100% efficient, you can use a value of 80% for that) as soon as you've calculated the forces needed to move up and down (will be different), before those forces can be calculated you need to decide on the acceleration you want. 1G upwards would be the same as free falling down, so let's take that and decide the maximum weight bearing down the axis that we move, lets say 100kg (pretty much, but we want to know the maximum values the system should cope with)

100kg times the acceleration of gravity => 100kg*9,81m/s2 =980,67N (Force of gravity on the axis bearing it down)

Now with 1G acceleration upwards we need to add the force of gravity to the force needed to accelerate axis upwards (1G is the same as 9,81m/s2, 2G would be double of that)

Force needed to accelerate 100kg upwards with acceleration of 1G => 980,67N+(100kg*9,81m/s2)=1961,33N =FAxisUp

Now the downward force would be 0N as the gravity pulls the axis already down (at the same 1G acceleration), but remember that the screw isn't 100% efficient so we need some force pulling the axis down if 1G acceleration down is desired, but the neede force will not be anywhere close to that needed on upward motion, so we can ignore it for now, but remember that the motor HAS TO HAVE stall torque large enough to make the axis stay at the same place and not giving up!

Now we can calculate the torque on the screw (motor actually) (remember to multiply this value with the efficiency!)

(FAxisUP*Screw pitch (5mm))/(2000*pi*Efficiency(80%)) => (1961,33N/5mm)/(2000*3,14*0,8)=1,95Nm

2000 comes from the fact that our screw pitch is given in millimetres not in metres so we multiply 2*pi by 1000 (1m=1000mm)

Now we can calculate the power we need the motor to produce

2*pi*1,95Nm*50 1/s (=3000rpm) = 205W

Now we know that in order to accelerate mass of 100kg with acceleration of 1G with maximum speed of 250mm/s we need a motor that can rotate 3000rpm and has at that speed torque of 1,95Nm and can therefore produce 205W of power on the output shaft.

There may be errors in the calculations as I'm a bit sleepy at the moment

Wow! Great info. I even understood most of it!

This may be simple for people smarter than I am. I know I will wrap my head around it but it will take some time to digest. Someone must have created a calculator/spreadsheet to do all this already right? I have some stuff to think about. I'll be back...

Mike

The calculations are based on physics I merely took what I needed. Those calculations are just simplifications of the problem to be solved, but they will get you started. Dynamic motion is bit harder to calculate.

Sorry to be such a noob but I am having a bit of trouble with the torque calculation. First, I believe you meant to multiply FAxisUP by the screw pitch as you originally said. I have highlighted the error in red. Assuming that, now comes my confusion. Doesn't torque = force * distance? Force is 1961.33 and isn't distance 250mm? I don't see the correlation in your formula. I am not saying you are wrong. I am just trying to understand the principals.

Also, I read that the efficiency of a ballscrew is 90% forward/up and 80% backward/down. Can't I just compensate for the 10% loss of forward efficiency by adding 10% to the weight? Am I thinking of this wrong?

Also, is 1G sufficient for what I am trying to do?

Mike

Torque is similar to linear force, but acts on a radius or a circle, the length of the axis has nothing to do with this. Pitch of the screw tells how far the mass and axis moves per rotation. Multification of efficiency is done with the 0,8 (=80% screw efficiency) and final answer is the torque needed for the motor.

You're correct, I erronesly typed division when it should have been a multification (FAxisUp*Screw pitch)/(2000*pi*eff) is the correct term to use.

Efficiencys of ball screws are mostly estimates, if you use lower value for the efficiency then at least you won't be disappointed if the screw and nut combo you acquire don't meet your specifications. If they exceed the initial estimate then all just works better.

1G is just the acceleration. Nothing in this world moves from zero to something in an instance, it takes time to accelerate anything to any speed, faster you accelerate it the faster it will come to the desired speed, but then you have to be able to control the mass as not to exceed the desired position or else you might shoot out the roof (just kidding), here's the start of the dynamic part

You could of course calculate the needed acceleration for the screw to meet the desired speed. There is not always going to be enough screw travel left to get to the desired speed. For example if the axis is in the middle and you order the axis to move all the way up with max speed, there might be a situation when the acceleration is not just enough to ever get the screw going at the max speed.

250 mm/s is the speed in millimiters per second, which in this example is the maximum speed ever attained. It is taken in when calculating the power of the motor, 50 revolutions per second is 3000 revolutions per minute.

In the end it's just a big pile of compromise...

Here's some more reading
http://en.wikipedia.org/wiki/Classical_mechanics

hi
D box actuators are stiff secured on the frame.This means when actuators are moving the distance between floor points chance.D box uses big cone formed floor points ,and little travel to avoid damage to the floor.This also create stresses on the actuator
Making a 6 inch actuator and use 2 inch create room for error,this is good.
You dont need to thank me ,there are more people that are willing to help as you see.
Take one step at the time and take the time you need! There is a lot more.
Work every thing out,and decide before you spend youre money
I,am a mechanic and build industrial machines
Ad

adgun,

I appreciate your willingness to share your expertise.

I noticed the D-Box floor points and I realize something will need to move somewhere. I haven't quite figured that out yet.

Are you saying it would be good to use 2 inches of the 6 inch actuator as a margin of error?

Mike

hi
sorry i did,nt made it all clear!
with a 6 inch actuator you have about 5 inch freedom for movement
and 1 inch for error
if you use les it,s allways fine
when the sofa stands up the frame you decide how far you want to go.
regards Ad

Yeah, so maybe I will actually make the actuators have 8 of travel so I have plenty of room to decide how intense I want to get.

Like you said, I will have to elevate it a bit so I can get my movement correct.

Mike

Consider the limitations of brushed motors before going ahead.

Using Jurki calculations you will end up with reasonably high static torque (almost 50% of peak) that motors have to create on a permanent basis while sitting most of the time in stalled position (no movement.) Brushed motors don't like this. They naturally have arcing but under high loads in stall electric arc eats into commutator creating pits because heat is not spreading around the whole of the commutator and electric contact is not re-circulating since rotor does not move.

DC motor under 50% peak load rotating at 3000RPM might survive for years while the same motor at 50% in stall might not be a happy camper.

I am not saying it is going to be a problem but do some research and consider this if you want to run the platform for many years on the same set of motors.

Leo

Thanks for the advice. This application is a torture test for motors. No doubt about that.

Do brushless motors do better in this regard?

Mike

They do ,but you need a Xsim compatible bldc controller for it

ps.
no one told you to use 50% from MOTOR peak for continuous stall
50% is difference between 1 and 2G of acceleration!

Stall torque on 100kg load against gravity would be (FGravity*Screw pitch (5mm))/(2000*pi*Efficiency(80%)) => (981N*5mm)/(2000*3,14*0,8)=0,98Nm or 50% of the max upward force but thats just because of the chosen acceleration (1G) which is just as much as the acceleration of gravity and is 50% of the total upward force (1G+1G)*mass

Companys give continuous stall torque not for nothing,they want to keep there name.

Here's some more reading regarding torque on dc-motors http://lancet.mit.edu/motors/motors3.html and here http://www.me.umn.edu/courses/me2011/ar ... index.html

Some forced air cooling might be a good idea to keep the motors cool as they produce more heat when stalled or when running in low speed (most of the time this is the case) Motors own fans don't make any (or not enough) airflow at those conditions, some pc fans could be adapted.

Sorry, but I really don't think so.
You'll have good static and low frequency performance but the power of dbox is its wide band response (near 100Hz), you cannot have such dynamic with the motors that you are quoted.

jyrki.j.koivisto did some calculation for linear acceleration, now if you just take a 20Hz sinusoidal motion, it's not the same story.

Dbox actuators are a masterpiece of hardware, don't expect to have the same performance with very low cost materials.
Whereas SNC5-like actuator (iwht integrated control system) is really easier to DIY.

Interresting reading: http://eetimes.com/design/audio-design/ ... technology

thanks for this link, very interesting!
Bruno Paillard a été formé comme un technicien à l'IUT-Le Creusot (France) en 1982
Dbox is French !! ;D