Motion Control and Motor Drives

cancel
Showing results for 
Search instead for 
Did you mean: 

Which Compact RIO Chassis and Controlleur for 3 axis control

Solved!
Go to solution
I already ordered and receive a 9012 (128MB, 64DRAM) controller and a 9102 (1M).

A little logic of digital I/O, 4 high speed encoders read, 1 slow encoders read, 3 analogs outputs.

I'm a little bit afraid I will miss space? What do you think.

Thanks,
Patrick
0 Kudos
Message 1 of 32
(5,617 Views)
Solution
Accepted by topic author paubine
Hi Patrick,

with a 3M gate backplane it is possible to implement 6 full featured axes, each running with it's own PID controller. If you use the same PID controller in a loop (this requires some memory managment to keep the internal states for each individual axis), it's possible to run 8 axes with a cRIO system.
In this scenario there is a ratio of 2.66 axes per 1 M gate. So I doubt, that it is possible to implement three PID controlled axes with a 1 M gate FPGA.

The best solution for your application in terms of pricing and performance should be the NI cRIO-9074, which is a backplane with a 2 M gate FPGA and an integrated controller. Maybe it's a good idea to contact your local NI branch and check if there is a way to exchange your cRIO system. This should provide enough headroom for your application.

Kind regards,
Jochen Klier
National Instruments
0 Kudos
Message 2 of 32
(5,600 Views)
Are you using NI Soft motion?  If so, you will need the 3M back plane to do 3 axis of control (if using the trajectory generator).  However, if you are only talking about 3 pid, 4 encoders and a little digital logic, that should fit on a 1M without any problem. 
SteveA
CLD

-------------------------------------
FPGA/RT/PDA/TP/DSC
-------------------------------------
0 Kudos
Message 3 of 32
(5,583 Views)
I don't know what to do!
 
Should I invest in NI SoftMotion...?
It's my first project as a mechanical engineer junior ... lots of money to invest...I prefered NI to all other motion controller till the beginning because of the "custom" but now, if I have an other software to buy... I will have to do a part of my career in controling robots 🙂
NI Softmotion look simplier to control motion but... what will be the result of not using NI Softmotion? I'm doing gearing from a master encoder...(kind of contour move, no need of Softmotion, I think), other then that, straight line move...
 
gearing:
slave 1 = blablabla * cos(theta_master)
slave 2 = blablabla * cos(theta_master)
 
I return the 9102 and the 9012 and asked for the cRIO-9074 in case I have to use NI Softmotion. (it's 575$ more expensive, not too bad)
 
What do you think?
 
Thanks again,
Patrick
0 Kudos
Message 4 of 32
(5,581 Views)
The nice thing about NI softmotion is that it has the trajectory generator built in.  You could write this code yourself, but it's not trivial.  The softmotion module gives you coordinated axis moves, contour moves etc.  All the features you would expect from a motion control board.  The difference from a motion control board, is you have the ability to completely customize the implementation in LabVIEW.
SteveA
CLD

-------------------------------------
FPGA/RT/PDA/TP/DSC
-------------------------------------
Message 5 of 32
(5,563 Views)

What exactly mean "coordinated"? (setpoint of each axis has to be coordinated?)

Also, what would you do to implement these equations to NI Softmotion, you would give the setpoint from the equations to which fonction (straight line move?)  or where would you write the equations?

Thanks again

Patrick

0 Kudos
Message 6 of 32
(5,543 Views)
Patrick,

"Coordinated moves" means, that you can group axes into one vector space and provide move constraints for the vector space instead of the single axes. Here is an example:

On a X/Y stage you could define axis 1 as x-axis and axis 2 as y-axis. For a straight line vector move with NI-SoftMotion, you provide a target position as X/Y coordinates and a single value for acceleration/deceleration and velocity. NI-SoftMotion makes sure, that both axes start and stop at the same time and that the load moves along a straight line. Acceleration and velocity values are applied to the vector (= to the moving part) and you don't have to care about these values for the single axes.
The same thing is true for arc moves and 3D moves.

Does this answer your question? I didn't understand your last question exactly.

Thanks,

Jochen
0 Kudos
Message 7 of 32
(5,529 Views)

Jochen,

Yes, you answered my question for the "coordinated" move. thanks!

As for the gearing with NI Softmotion? My 2 slaves axes displacement depend on the counts of the master...

master counts in -> (black box of equations) -> position of the 2 slaves (slave count)

So each count of the master give a new position for the two slaves with acceleration and speed not constant for all the trajectory, which last 1 sec. (with accelaration and speed "ok" for the positionning tables and ratio of 3:1)

The master is a 2000 counts encoder, so labview will give 2000 positions targets to the slaves in 1 sec (master shaft is 1rps)... the master encoder resolution is enought? 2KHz ?

Everything should work with a simple PID, but I was wondering if NI Softmotion could help me.... and now... 🙂 if my encoder has enough counts/revolution

Thanks,

Patrick

0 Kudos
Message 8 of 32
(5,524 Views)
Ok,

obviously I didn't read the earlier posts to this thread thoroughly enough. As it looks like the trajectory is generated by an external axis, I don't see an urgent need for NI SoftMotion in this scenario.
The encoder frequency seems to be a little bit low, but it depends. Typically servo motors are controlled at frequencies between 2 and 20 kHz, but in your case I would limit the control loop rate to e. g. 500 Hz. You will have to experiment with this a little bit, but you should be aware of the fact, that changing the control loop rate significantly, requires also significant changes to the tuning parameters.

Jochen
0 Kudos
Message 9 of 32
(5,515 Views)
Jochen,
 
The higher frequency control loop is not necessarly the best for controlling? why do you suggest 500Hz instead of 2KHz?
Any theories on that subject?
 
Thank,
Pat
0 Kudos
Message 10 of 32
(5,512 Views)