servo homing question

shawnhimself

Member
Join Date
Oct 2006
Location
LBC
Posts
84
I'm not extremely familiar with servo motors however I know there must be a better way...

In an application of ours, a servo compression unit is 'homed' by smacking against two lock collars. I believe the theory is that when the motor senses an overcurrent, it establishes 'home' position.

Feel free to school me on this. I just feel there is a better way. Using an encoder position, sensor, etc.

Thanks.
 
The positive stop homing method is often the most accurate. Usually the velocity and torque limits are reduced so that instead of "smack" it is more like a "gentle push".

The controller may be using the position error during this routine to recognize home position, not necessarily the current limit. At least that has been my limited experience with it.

It can probably be changed to work with a limit switch but the repeatability and reliability will be that of the switch used. Any switches I've ever used will vary more and fail more often than a fixed positive stop.
Paul
 
What you are refering to is homing to a hard stop. The advantage to this method is that if you need to move the home position, you can do so by moving the hard stop. The disadvantage is that if you have the homing current set too high, you can break the weakest link in your system. Also, accuracy may be affected by any contamination between the pieces that come in contact during the home or wearing away of material.

One of the most common ways of homing is to a switch. There are several ways this occurs and many controllers have built in routines to handle this. You can just move in a direction until the switch is made and call that home. This is the crudest method. Moving up from there, many routines will find the switch at some moderate speed, then reverse direction at a slower speed until the switch signal goes away and call that home. The concept is that the move at slower speed will provide a more repeatable home position. Again, if you need to adjust the homing position, you can move the switch or the feature that actuates it.

Next up you can run until the switch is sensed and then reverse until the control "sees" a marker pulse on the encoder or resolver. This would be more repeatable than the previous metohds as it doesn't rely on a switch actuation for the position. However, if you need to move the home position, you can only do so in incremental revolutions of the motor or by adding homing offsets with a computer.
 
The positive stop homing method is often the most accurate. Usually the velocity and torque limits are reduced so that instead of "smack" it is more like a "gentle push".
This method is not the most accurate as it relies on the machine being stressed. The machine could deform by 0.01 inches or more. It is very important that contact does not involve high torque. The lower the better. There is an offset proportional to how much the machine is stretched or compressed.

Better is to make physical contact and then move out to the z pulse where the machine is no longer stressed.

Normally one backs up to a prox and then extends to the z pulse. Note, a good motion controller should allow the z pulse to be anywhere. I know the last project where I did this we see the prox so the limit switch was behind the 0 point. The z pulse was at about -0.25 inches. This should be a parameter for the homing routine. The axis would retract to the prox and then extend out to the z pulse and stop at 0. It was very simple and fast.
 
What type of control? You should have options but you need to make sure you have what you need to make the changes. Do you have the programming software etc to do the job. Give some more details and I am sure someone here can get you started.
 
I forgot one other option which is to use an absolute multiturn encoder or linear MDT type absolute encoder. Home once and then never again until you take something apart.
 
Peter Nachtwey said:
This method is not the most accurate as it relies on the machine being stressed. The machine could deform by 0.01 inches or more.

That depends on the machine. Back in the day before we used absoulute encoders, it was the best method for homing slides used in machining transmission housings. We had to hold in the neighborhood of .0001" You just need a rigid enough machine.

The Indramat system had home to positive stop as a built in function.
 
I tend to agree with Rick.

When I worked for an automotive parts manufacturer we had a rotary transfer machine with 59 axes! This machine manufactured one 'wheel cylinder' every 6 seconds for car drum brakes. Accuracy was paramount and the main axes (Z's) were hydraulic servo axes with linear scales and were all homed by sending them to a fixed stop. Never caused us a problem?

Rich
 
Last edited:
Shawn,

I recently completed a hydraulic press control system that was comprised of two good sized opposing rams. The cylinders were designed with LVDT which provides absolute position feedback. The position feedback is used to slow the flowrate by ramping the servo valve as the ram reaches the end of travel to soften the transitions. This has been a huge hit with the operators and the maintenance people since the direction reversal is almost silent and it is much easier on the equipment. As a secondary benefit, an anti-collision feature was provided that prevents the two rams from banging together.

Maybe something similar is possible in your case.

Best Regards,

Bob A.
 
I recently completed a hydraulic press control system that was comprised of two good sized opposing rams. The cylinders were designed with LVDT which provides absolute position feedback. The position feedback is used to slow the flowrate by ramping the servo valve as the ram reaches the end of travel to soften the transitions.
What was this press used for and what controls did you use?
 
We are using a micro 1500, ultra 3000 drive, and an mp-series ab servo motor.

This is just sort of a side project of mine. It just seems like the customers don't really like the crude nature of it. While I do realize it works and it's fine. It's more of a cosmetic fix to me.

I also do not like the fact that if the lock collars were to be moved that the servo would continue on its path until it finds resistance.
 
We have a stress column design for our frames. You have a bottom plate and a top plate. Four tubes seperate the plates and 4 bolts sandwich the hole thing together. By applying more force with the bolts you can stretch the columns to eliminate stretch. We see about 0.025 before stretching the frame. After we still see 0.003-0.005

Each of the bolts are 2-1/2 inch.

Here is what I think. If you can get a better home position by doing it this way you must not have very good resolution. When I was doing machine tools we used a prox sensor then went to the Z channel (marker) on the encoder. If the Z channel is not good enought then your not going to hold a 0.0001 tol.
 

Similar Topics

Hey all, I have a couple older style Yaskawa servo drives (not sure on exact model at this moment, but i can get it) We had a power failure...
Replies
5
Views
2,127
Hi, I am trying to get my head around as to what the absolute homing does. Below is a statement that I read in the logix 5000 motion instruction...
Replies
4
Views
2,135
I have a machine I'm trying to refurbish. Currently, it won't home without intervention. It's using a 1746-L532 processor with two 1746-HSRV SLC...
Replies
0
Views
1,178
Hello boys n' lads, I'm trying to make a step into my sequence which main function is "homing" the servo motor. Here is a screenshot of my...
Replies
5
Views
3,932
Back
Top Bottom