Mitsubishi Servo - Controlled Stop

[Plastic]

We have a Mitsubishi servo (with an electromechanically brake) that rotates a table with a few hundred pounds of mass.

Through set-up and tuning, the servo makes smooth, accurate, and deterministic movements of the table and stops in repeatable positions for the shot pins to engage, and work performed.

The problem we have is when there happens during one of three events.
Light Curtain Interrupt during movement
E-Stop during movement
Power loss during movement

The mass of the table causes a little slippage in the shaft coupler between the table and the servo gear box. This introduces a mechanical offset, and the shot pins won't engage.

Other than allowing the slippage, and re-home on each of the three abnormal occurrences, what options may there be.

I am in controls, I believe the root cause is mechanical deficiency, but as usual, they are asking about controlled stop, and "power-off delay".
Thoughts?

 

[jkerekes]

I'm not really a motion guy, but the three conditions you mentioned are in my definition, emergency stops. I worked with old AB IMC 123 servos. When someone tripped the safety mat, it didn't do a smooth controlled stop, but instead did a STOP RIGHT NOW! stop. After awhile of these kind of stops, a shaft on a reducer failed. We change the action of these emergency stops to a controlled one.

 

[Steve Etter]

As jkerekes says, you have pretty much defined the emergency stop conditions. So based on what you have laid out, I agree you do indeed have a mechanical issue, not a controls issue. Furthermore, I see four possible solutions.

First (and probably least expensive initially) would be to allow the slippage and to create a homing sequence that runs each time the system recovers from an estop condition. If you don't already have them, this would require a "home" switch and flag. The down-side of this method is that you are allowing the slippage to remain and this can potentially cause physical damage to the shafts and coupling. This IMHO is not the right choice.

Second would be to find another coupling technique that prevents the slippage. This is clearly a completely mechanical solution and, in my experience with similar efforts, will probably not be successful.

Third would be to redefine how estop conditions are handled and allow the table to coast to a stop when the an estop is initiated. If this is even remotely acceptable, it will require you to go through a complete risk assessment to make sure no unacceptable dangers are created. Once again, I doubt this is the right choice.

Fourth would be to install a physical brake on the table itself that engages anytime the estop condition is initiated. This, obviously, is a mechanical fix but feels like the right one to me.

Hope this helps.

Steve

 

[muttleydoo]

Maybe the most robust solution would to do a homing routine after all the interruptions. I guess that the interruptions would only happen infrequently so the delay would be insignificant in relation to the stoppage caused by the interruption.

A homing sequence using the shot pins could be the least costly solution as there would be no extra outlay, you could use a data set method using the shot pins as the home position. Using a sensor as a dog switch would not cost the earth, and using this method with a home offset would be my preferred option.

Dropping the power instantly is the only real option under e-stop or safety gate interruption so homing would be the only realistic solution in my opinion.



This is definitely a post that could stir some interesting debate.

 

[iant]

"The problem we have is when there happens during one of three events.
Light Curtain Interrupt during movement
E-Stop during movement
Power loss during movement"

The powerloss during movement will need to be addressed seperately as per the above suggestions

The E/Stop - Light curtain - is caused by the power to the drive being cut out - (Check safety Cct wiring)
Depending on the servo drive age
Most current servos require the safety relay contacts be directly wired to the servo safety inputs.
- NO POWER ISOLATION VIA SAFETY CONTACTS ARE NEEDED
so the supply can be kept on

Older servo's motion need to be stopped by the servo before the power is dropped off. - Use a safety relay with an integrated timed output.
i.e. stop the servo via the safety signal then after a time interval. drop out the power to the drive

Very Old servo's can't have the power dropped off you need to use the the cycle hold only.


Hope this helps

 

[muttleydoo]

I know with modern Mitsubishi servos like the J3 the only way to control the safety is to remove the power they do not have the safety gate input facility as far as I'm aware.

How you can control the outcomes depends on the servo controller type and its safety function. As Iant suggested you may have to go for a timed shutdown under these conditions, but it would have to be subject to a robust risk assessment with other control measured such as pressure mats or secondary barriers.

James

 

[LadderLogic]

1. Fix the coupling issue, pure and simple. There is no such thing as a positioning application with slipping couplings.

2. It is only power loss that should ever require re-homing. As long as the motion controller and the drive control circuitry are alive, there should be no need to waste time to re-home. I do assume that e-stop or light curtain kill only the drive power lines while control voltage stays alive.

3. If acceptable, invest into a safety relay with 0.5 sec off-timer. That way, you may be able to tell the drive to execute "immediate stop" command before dropping the power contactor; the drive, if sized properly, should be able to dynamically brake the motor using its internal circuitry.

4. If movement after the complete power loss is an issue, or if it is not possible to come to complete stop within 0.5 seconds, you would have to install braking resistors following the drive manufacturers instructions.

 

[robb676]

I would check the coupling, there are special couplings for encoders that limit slip or compensate for it.

 

[iant]

You are correct with the lack of safety inputs - bugger.
The J3 manual shows the wiring of an E/stop in series with the electromechanical brake cct.
this would be a secondary stop method and would be acceptable as the next step is the primary safety stop.
It is nesecary to institute a controlled stop command before the drive power is de-energised
Wire the dual safety relay to first stop this (secondary safety). then after the time out period drop out the power contactors (primary safety).

A free wheel power stop is more dangerous and in many other machines, destructive - this is what you are seeing now.

the risk analisys will still need to be done as a matter of safety requirements

 

[leitmotif]

I do not see how you are going to get good repeatable results if your mechanical drive can slip. I am kinda surpresed you dont have a key and keyway on output shaft of reducer. I would do one of two things
1. Cut keyway and insert keys
2. Cross drill coupling and install shear pins

In both cases I would use brass to avoid tearing up shafting and coupling especially so during setup and testing.

Dan Bentler

 

[iant]

Hi Dan,
Like you i have no idea exactly what this machine is.
But I have experiance with the PowerLoc style coupling
They won't slip and do not need keyways.
The slip type coupling reffered to - i would guess - is a torque coulping
But I would say not suitable at all for this application.
It sounds like the turret for a turret punch or similar.
I am sure the servo motor is going to 'Coast' because it does not know where it is - the power is off

 

[leitmotif]

Hi Dan,
Like you i have no idea exactly what this machine is.
But I have experiance with the PowerLoc style coupling
They won't slip and do not need keyways.
The slip type coupling reffered to - i would guess - is a torque coulping
But I would say not suitable at all for this application.
It sounds like the turret for a turret punch or similar.
I am sure the servo motor is going to 'Coast' because it does not know where it is - the power is off

Is a Power loc coupling also known as taper lock? If so they are good for the assembly guys who get adoration from mangagement because it is quick. Us maintenance guys also appreciate them for the job security when we gotta get em apart - especially when they do not install those with jacking bolt holes. I try to avoid them when possible.

If he is only coasting to stop on power loss I dont think this should be causing slippage in couplings

Uhhh is he using worm gear drive I wonder - in that case I can see the couplings slipping especially with high reduction worm gear driving a heavy mass load.

So many pickyuny details behind simple questions.

Dan Bentler

 

[iant]

This is a powerlock coupling - it is not a taperlock.
It is similar but is paralell - the image shows two devices not two halves

Also the actual drive for the table is only a guess.
If during normal operation it stops normally.
than I would suggest that the dropping of power is preventing the controlled stop.
It would normally not be a wormwheel drive with a servo.
More than likely - planetary geared or direct driven.
The wormwheel type is a high ratio (20:1) and would be both inaccurate
And a huge loss of control. the gear clearance is too much for a servo.