Inertia Mismatch - Please Explain This...!!!

I am currently working on a servo application in which I made a big screwup specifying the correct motor. It apparently comes down to inertia mismatch which is something I can't quite grasp. This is what I was hoping to get some more information on.

Here are the details...

The application required a small servo motor to spin a dial. The dial is used to flip over some parts that are coming in upside down. The reason for the servo motor is we need acurate positioing and speed control. The initial design was to direct drive the dial with this motor.

Based on the inertia of the load (dial) and the speed requirements, I figured this was not a problem at all. I figured out the torque requirements for the motor based on the total system inertia and the acceleration requirements. This allowed me to pick out a small Y series motor from AB. The rest of the system was already on a AB Sercos network so to integrate this cheap motor and corresponding drive was a sure thing. Or so I thought...

It appeared as though the motor had plenty of torque to move the load at the required rate. At least that is what I thought (basec on the my calculations.) Then I started working on the application and realized that the motor just can't do it. So naturally I asked myself why...

So I started doing some research and got my local AB motion control guy to help. He explained to me about the inertia mismatch thing but really did not explain it all. Just kinda of said it was a rule of thumb and reconfirmed that my current configuration was not going to work. That was not quite the answer I was looking for.

So I am currently in the process of adding a planetary gear box to reduce the reflected inertia to the motor. Not because the motor cannot handle it but because the load-to-motor inertia ratio is way too high.

I can't quite get the theory of this down. If the motor has the torque to do the job, why does the mismatch override this.

Can someone please explain the theory and importance behind inertia mismatch when dealing with servo applications?

I am throwing a layman's thought out here. IF there is motion involved with the part then the motor has to be able to override that motion to develop its motion.

If there is no existing motion involved I have no idea what they are talking about, inertia implies motion.

Google gave me this:
It has some math ( fancy formulas ), trends and stuff. ( Peter should love it.)
I don't have a clue what their talking about, ( I'm just a bubba from california) but maybe you engineering types will.

http://kmtg.kollmorgen.com/welcome_center/articles/Inertia.PDF

Mickey beat me to it

I was reading the same article while Mickey was posting. I was trying to figure out how I would explain this.

OOOOh, more calculus. Yes Mickey it is good stuff. I copied it into my pdf library. This is the same kind of research I do with hydraulics only I don't worry about impedance matching. Hydraulic cylinders don't have gear boxes. Hydraulic motors don't usually need them either as they have a lot of torque. I must use the techniques mentioned at the bottom of the article and compensate for compliances using modeling and pole placement.

Anyway, look at the proof. However, the alpha_m and alpha_l are missing the alpha. This is too bad because it ruins an otherwise good article.

Ron, inertia is resistance to change in motion. Whether the load is moving or not makes no difference.

How does one make Greek letters again?

Peter

I am not sure how to do it nowadays
BUT
in the ole days you looked up a table of ASCII codes for various characters then you looked up a command in the software to allow insertion of that ASCII code and voila it printed out the greek letter.

Don't know how it is done these days but have had no demand for it.
Hopefully it is like riding a bicycle.

Incidentally the one book I have kept from college is physics.

Dan Bentler

I got slapped down, so nothing more to contribute.

how does one make Greek letters again?

Click to expand...

Not sure if you are taking the P or not, but I only found this out recently, character map under system tools, has all sorts of useful symbols etc.

Greek Symbols in Word

Peter,

I will assume you using word to create your documents.
Click on Insert then Symbols and then select basic Greek from the upper right drop down box.

Hope this helps,
Bob O.

abwiz said:

It appeared as though the motor had plenty of torque to move the load at the required rate. At least that is what I thought (basec on the my calculations.) Then I started working on the application and realized that the motor just can't do it. So naturally I asked myself why...

Click to expand...

What exactly was it that made you realize that the motor just can't do it?

LD,

The motor application was setup in position mode configuration. If I give the motor a jog command with a reasonable accel rate and speed, the load tries to move and has a hell of a time doing it. It actually starts twitching, then it jerks a little bit and then starts to move, but then it gets a position error fault.

I know it sounds like the motor torque is insufficient, but my calculations on inertia and acceleration torque say otherwise. The only possibility is that the inertia I was given (by another designer) was way off.

I am currently using a program from AB called Motion Analyzer. This program tells me the exact same thing that I calculated (on paper) except it does throw up a red flag about inertia mismatch. But the torque of the motor is well adequate...Once again, this is using the numbers for inertia that I was given.

I am still going to sit down with the original designer and see about the inertia numbers he gave me.

I suppose it is either that inertia mismatch REALLY does matter or my inertia calcs are way off.

I'm not discounting Inertia mismatch, but what about friction ?

Back in my previous life as OEM I use to work with Indramat servos mostly and remember really well that the only thing salesman calculater was inertia mismatch. Remember one small table that can't be rotated fast because inertia ratio was 700:1. Adding gearbox took care or the problem.

Same thing said engineer from Fanuc - add gearbox or axis will not accelerate - why old was OK? - because old men decided to increase lead screw diameter a little...

Too bad I did not touch this stuff since 2002....

The obvious thing is for abwiz to show us his or the AB motion calculator's calculations.

That article Mickey posted is pretty good. I remember seeing it before and I think I saved it then.

People tend to throw around rules of thumb, especially old ones, as if they are laws of physics. The whole inertia mismatch thing is one of those items. As the article states, and Peter alludes to, talking about inertia mismatch without talking about compliance doesn't make too much sense. For example, the Indramat guys in my area are applying frameless motors as direct drive elements to print cylinders on printing presses. These assemblies have inertia mismatches in the thousands to one and they don't have much trouble. The trick is that the motor rotor is attached directly to the roll shaft, they have extremely high count feedback devices and they have drives that can handle very high gain settings. I'm pretty sure Kollmorgan can do the same thing with their DDR series of motor.

As Peter said we need to see the information that abwiz is working off of. But my guess is that there is a 'spring' in this system that isn't being accounted for.

Keith

Here is another article that may be of interest:

http://www.ormec.com/csedocs/coupling.htm