OT - smooth motor rotation

katratzi

Member
Join Date
Mar 2005
Posts
210
I'm interested in opinions on which type of motor and drive would
perform the best if your requirements are extremely smooth
rotation at low rpm (< 10). And, which feedback device could best
measure this "rotational smoothness" with accuracy. I have heard
of using frequency analysis to analyze velocity disruptions, but
I'm not sure how this could work. Right now this is just a "what
if" type of thing, so the motor does not even have to exist - yet.
Since the measurement of velocity is always going to be time-based,
does that introduce error into devices such as encoders, etc.?appreciate the input.
 
I can't tell just what you define as "extremely smooth" rotation but, a modern flux vector drive with 1024ppr encoder will rotate very smoothly at 10rpm even with significant torque disturbances.

If you limit maximum speed, you could probably even go to a 2048ppr or 4096ppr encoder and improve that.

I have used 1024ppr encoders on ABB ACS600 and 800 drives to develop full torque at zero speed and I mean zero speed. The rotor position doesn't vary more than the equivalent of one pulse in the rotation. I'd say that was pretty, maybe even "extremely" smooth.
 
Confused again

Is there enough information provided? Is travel in one direction? Will travel change direction?

Speed is based on time
Velocity is a vector quantity with direction (rate of change)

Wouldn't the type of motor etc be based on application requirements for torque, speed, and velocity and capabilities of any accessory devices needed...ie gearboxes etc.
 
Dash- yeah, gearboxes are great for speed reduction but even
the so-called "zero backlash" planetaries have some degree
of backlash that affects motor smoothness

DickDV- "EXTREMELY SMOOTH" aye, theres the rub. this is hard to
define, much less produce. I'm going to define it as a
measurement of motor shaft position relative to elapsed
time. if the motor shaft position is incremented the
exact same distance over exact same time intervals then
motor speed should be uniform. I have heard of sine
encoders with an incredible amount of ppr feedback
that could be used to hold a motors speed close to
set speed, and I've read some about using a high
end tachometer connected to a voltage to frequency
coverter and feeding its output to a spectrum
analyzer so you can see velocity ripple in the time
domain.
ABB drives have fantastic torque control
capability. But one pulse variance in 1024
at the motor shaft isn't that much, but lets say
you had a large diameter cylinder coupled to it.
the one pulse error then gets magnified at the
cylinders surface speed. not trying to nitpick,
its just that I don't think that a "regular"
feedback device will give optimum speed
regulation, but I don't know what a really
good velocity control system is made up of.
Servo motors, perhaps?
 
rsdoran-

the motor will travel one direction only,
max speed requirements 50 rpm, no gearbox
desired, and as long as torque is sufficient,
the only concern is speed regulation
 
katratzi,

What's the application? What type of application requires absolute ZERO backlash?

I can tell you from experiance that DickVD is right on target. With the ABB drives and 1024 ppr encoders, our problem is that we have to be carefull not to make the ride so smooth that the riders dont feel the motion and think their trapped. How much better can you get then that?

Anyway there's not enough info on the application to make anything near a specific reccomendation or comment (other then full torque a < 10 rpm). One direction sounds like a conveyer of some sort, but that's not much to go on...In short your creating a technical debate without really letting us know what it's all about.
 
hey elevmike,

not trying to be secretive, its just that this
process, while not patented, could be called a
"trade secret". anyway, I promised I wouldn't
reveal the nuts and bolts, but I can tell you
that there will be a motor connected to a large
metal cylinder and that speed deviations from
setpoint by this cylinder will cause imperfections
in the end product.

So, the better the velocity regulation, the
better the product. and even minute changes in
velocity have a measurable effect. From research
so far, it seems that controlling velocity at
extremely low rpm is not a simple task. Motor
cogging, torque ripple, etc. I'm just trying
to match a motor and drive package to this
application. I'll get back with the process
guy and see if I can divulge more information.


Mike, do you guys use ac or dc drives in the
elevator industry now? Are the Magnetek
DSD 412s still used?
 
When I was building/rebuilding cnc machines we had a lot of problems with cogging. For years we only used DC brush type motors with a tach going back to the drive for velosity and min 1024 back to the controller. We could get great smooth control with this setup. We were able to go to a AC bushless servo motor with a resolver feedback. We then got a simulated encoder feedback out of the drive that fed the controller with 4096 feedback. We were using Baldor motors and drives. We also did some with Indramat (Bosch/Rexroth) ECO drives. They were even better than the Baldor but were more $$'s.
 
katratzi said:
Mike, do you guys use ac or dc drives in the
elevator industry now? Are the Magnetek
DSD 412s still used?

We still use the 412 DC drives on large gearless machines when doing modernizations on older equipment. However, nobody's doing new installs with DC motors that I know of. It just dosnt make sense anymore. Geard machine mods also almost aways get a new AC motor & drive.

What's nice about Magnetek is that their specalized in the elevator field so there's less engineering & setup issues on our end of it. We have a controller (entire system package) supplier who seems to like ABB though. We've had great expireance with them also.

Magnetek, ABB & Baldor seem to have it togeather as far as elevator applications go.
 
BTW, if the application requires a constant speed, (no dramatic accell & decell ramps), consider a flywheel to remove small ripples.
 
Why does this sound a lot like a film-casting application?
I've used DC Pancake motors (large, many many poles) in the past and had satisfactory results, though todays AC drives coupled to a good zero-backlash gearbox would probably be superior.

As far as measuring the 'smoothness', well, you could put on a second high pulse-count encoder, and measure the time for 'x degrees' of pulses. That is, measure the time it takes for every 15 or 30 degrees of rotation knowing that there are say 200 pulses per degree. Still, you might find locating encoders like that difficult.
 
It sounds like a good aplication for AC Slow Sym (Synchronous) motors to me. However I have only used fractional horse power slow sym motors I do not know ifthey make them for large horse power.
 
Last edited:
With a large drum and the need for smooth surface speed, it would seem desireable to drive this mechanism from the outer surface rather than from the center. Much smoother rotation and higher motor speed would result
 

Similar Topics

I am using an analog laser level detector to detect the rubber level in a 2 cylinder mill. The rubber is introduced on one side in softball sized...
Replies
18
Views
5,904
i try to make a car that move smoothly in HMI, i use propertie Vertical Position of car ( CarVp): CarVp=CarVp + 1 ( or + other number 2,3,... )...
Replies
6
Views
1,979
hy sirs, i need to smooth the speed setpoint on a simodrive (6SN1118–0NH00–0AA2). which parameter i should change? (preferably if there is one at...
Replies
0
Views
2,562
I found this link on another site http://www.holoborodko.com/pavel/?page_id=245 In the past I have posted examples of how to use the...
Replies
0
Views
2,633
At first, my projects run on Simotion C (C230) with SimoDrive 611U and there was no problem all motions was dynamic. But i had to switch Simotion...
Replies
0
Views
2,440
Back
Top Bottom