Estimate stiffnes of rotational system

aand74

Member
Join Date
Dec 2005
Location
Deinze
Posts
131
We have a system where a servomotor will rotate a big steel disc.
To obtain an allowable speed of the motor for the speed of the disc we want, there will be at least a gearbox after the motor, and maybe also a chain or belt with reduction factor (chainwheels with different diameter).
We know the accelerations we want from the disc, so we can calculate if the motor torque and speed is allowable, and can select the gearbox that is best fit, and if neccesarry also the
speed ratio of the chain or belt.
The disc will have quite some inertia. Now I am worried about the stiffnes of the chain or belt (if we use one).
I suppose choosing the ratio of the gearbox so that no chain or belt is necessary anymore will be the stiffest construction.
But the fact is that we already have a gearbox that we might use.
In that case a chain or belt will be necessary for further speed reduction.
But how can I know in design phase if the drivetrain will be stiff enough? Especially when stopping very fast, I want no oscillations. What tools or formula's can help me in that determination?
The servodrive and the gearbox is Siemens.
 
The shaft will be like a torsion bar if it twists due to applied torque. This is kind of like a spring. You should be able to get specs on the shaft for the amount of twist per torque.
you must calculate the natural frequency and the damping factor. If th system is critically or over damped then the system will not over shoot with open loop control. If the system is under damped then the derivative gain and second derivative gain may be required.


I have NEVER seen anybody do this calculation. I usually rely on the auto tuning to make sure the closed loop poles are on the negative real axis in the s-plane so there is no overshoot.
 
i will admit that i am no expert, any device of good size in motion must have a slow down curve before it stops.
you have a large steel disk, lets say 500 pounds rotating at 200 rpm.
you cannot simply tell it to stop turning, it must be coasted or braked to a stop.
that is like telling a trailer truck pulling a full load going at 65 mph to suddenly slam on the brakes and stop in 10 ft. that's not going to happen. the energy must go somewhere. the belt may slip, the chain may break, the gearbox may tear apart, the mounting bolts break.
please do your calculations on everything forwards and backwards to see what your system can do. i say this to help alert you to the possibilities. say safe.
james
 
Please give some data for your 'big steel disc' application - your big and my big could be worlds apart, also the speed/accel requirements.

Have you considered a Siemens Torque motor - I've used these in the past to eliminate zero backlash gearboxes.


I wouldn't consider using a chain in a servo motor application as the chain will wear and introduce more and more backlash over time.


I've lost count of the number of times I've comes across inadequate elastic bands (sorry drive belts) because they have been chosen based on their power requirements and no consideration is made for the stiffness of the belt. Ask your mechanical designer what the stiffness is and see how long it takes to get an answer:whistle:
 


Excellent, but I have never seen anybody really do it for real let alone a controller for it.
The goal would be to control mass 1 without overshooting. I PID with feed forwards will not place all the poles one the negative real axis in the s-plane so there will be some overshoot.


If I am provided with values for those variables I can do the simulation but NO ONE HAS EVER PROVIDED THEM. I have always had to use an auto tuning program that does system identification that estimates these values.
 
You should fear no evil

LOL, yeah, the video seems to leave that, plus the DifEQ solver, as "an exercise for the user."


You can modify the SOPDT system identification program to solve for the two inertias, spring constant and the damping. This assumes you get some excitation data. This system won't be as weird or difficult as that heating system with the valve that didn't respond well.


The next trick is to make an open loop transfer function for the rotational position of the load that is being controlled. The do some symbolic math to compute the closed loop gains that place the closed loop poles on the negative real axis.


Easy! It will give you something to think about.
 
You should be able to simulate this system in the software tool IndraSize from Bosch Rexroth. It should be a free download (they may ask you to create a free account).

With the disk geometry and material it will calculate the inertia. You should have the gearbox info from the manufacturer.

You’ll get a speed torque curve with points showing the requirements of the moves you program in.

It will select a Servo motor and drive if you want as well.
 
Norm, how does that help in estimating the STIFFNESS, torsional spring constant, between the TWO inertias?

Repeating what I said above. I have never seen this calculated. With an RMC you would go through the auto tuning procedure and use the second derivative gain to place the extra pole.


Tuning a single pole system without torsional compliance is super easy.
 

Similar Topics

Hi Rockwell experts. I have a customer who is using two non-Rockwell EtherNet/IP scanners for control of car-body line robots and IOs. Both...
Replies
0
Views
1,292
For those of you on your own - how are you estimating hours on program design? I've heard of people basing it on the I/O count but that doesn't...
Replies
6
Views
2,992
How do you estimate how much time you need to draw an electrical circuit of a machine automated by means of a plc?
Replies
2
Views
2,206
Hi: We are in stalling a Ethernet/IP Network with 4 ControlLogix with two A-17 chasis each one, connected by Cnet between chasis. Each Clgx has a...
Replies
2
Views
1,702
Hi Automation Professionals, I found this web page is very helpful finding many questions and answers.But I also expect some more about my career...
Replies
2
Views
2,803
Back
Top Bottom