limit and home - best practice - conceptual

# Thread: limit and home - best practice - conceptual

1. ## limit and home - best practice - conceptual

Howdy,

I'm not looking for a technical discussions on details, but a statement/report from Users about concepts:

we all know what home, soft and limit switches can/should do, and we most likely all at least one variety of them on our machines.

my understanding is that

LIMIT switches should SOMEHOW interrupt a motion when one axis goes beyond a certain PHYSICAL limit.
SOFT limits should do the same, but the interruption is sent by some control software in charge of the whole job.
HOME switches are useful at start up in order to define a reference frame for a machine that does not have "absolute" encoders.

The question that arises is the following:

Assume a (30 years +) system that has almost no friction (air bearings) for XY, and a spindle like system on Z.
The gantry traveling on air bearings is about 500 kg, can travel 2 m by 3 m on a polished stone surface, slow motion, and is "protected" with limit switches.
If this assembly hits one of the limit switches, the power to the servos is removed, but the inertia can suffice to have the assembly pass (by destroying it) the switch and fall on the floor. Dangerous.

What scheme would you suggest to improve this situation ?

regards, Philippe

2. ## Re: limit and home - best practice - conceptual

I wish I had a system that size, floating on granite!

If this assembly hits one of the limit switches, the power to the servos is removed, but the inertia can suffice to have the assembly pass (by destroying it) the switch and fall on the floor. Dangerous.
What scheme would you suggest to improve this situation ?

Easily solved.

Change the switches from direct action mechanical switches (microswitches?) to either magnetic sensing or optical, both from the side. That way the table can go straight past the sensor without ever touching it. (I am assuming it was the switch which fell on the floor here!)

Second, I would NOT remove the power from the servos under these circumstances. The idea that you have to do so 'for safety' is very old and thoroughly out of date. Instead use a driver with an INHIBIT input. Get some electronics help and set this up to be latching, so that when you have got the table out of the Limit zone it does not spring back into life immediately. Instead you will need to use a Reset button to re-enable the drives.

I would also implement the Soft Limit system in parallel. It could be used to tell the controller to bring the table to a fast stop, before hitting the hard Limit switch trigger. This would preserve axis calibration, and be a huge saving in time.

Cheers
Roger

3. ## Re: limit and home - best practice - conceptual

First of all, moving parts of the machine should never be able to just fall on the floor, no matter what. There must be mechanical stops (bumpers) preventing that.

Activating a limit switch should not just turn the power off and allow the machine to coast. The runaway axis must try to stop as quick as possible using servos, brakes, or whatever it has available.

And if the axis still hits the mechanical stop despite all efforts, the damage to the machine should be minimized.

4. ## Re: limit and home - best practice - conceptual

"I am assuming it was the switch which fell on the floor here!"

much worse, it the "gantry" that could potentially get off the table ... at this point the damaged switch is of no concern ...
it never fell in the life time of the machine, but there are obvious design weaknesses that I need to to fix.

"Instead use a driver with an INHIBIT input" and as CitizenDreams puts it "Activating a limit switch should not just turn the power off and allow the machine to coast. The runaway axis must try to stop as quick as possible using servos, brakes, or whatever it has available."

I agree with you both of you.

"I would also implement the Soft Limit system in parallel. It could be used to tell the controller to bring the table to a fast stop, before hitting the hard Limit switch trigger."

I agree, this is why more recent systems rely on SOFT limits. This change may require a complete re-write of the controls SW, or a change of paradigm.

"First of all, moving parts of the machine should never be able to just fall on the floor, no matter what. There must be mechanical stops (bumpers) preventing that."

Correct. I have no idea why they are missing, but they are. Blueprints are not showing them either, so that the they were not forgotten when the thing was assembled.

Since the "gantry" is on air bearings we can imagine to couple the hard LIMITs (not the SOFT limits) to trigger a reduction of the pressure in the main line. This is a brake for sure. One has to make sure there will be no resulting damage. What I'm looking for is a scheme that brings a runaway axis to stop as fast as possible WITHOUT the need of an action from the SW controller which could have crashed ... Also, what worries me is inertia. The machine is operated slowly, and used to reach precise positions. Encoders send their signal to a FB loop. At each target location, the main axis overshoots, changes directions, two or three times before it comes to halt. Again, all the process is slow, and not meant for manufacturing but for measurement processes.

regards, Philippe

5. ## Re: limit and home - best practice - conceptual

Ah - a CMS rather than a CNC. OK.

At each target location, the main axis overshoots, changes directions, two or three times before it comes to halt.
This is bad. It is well-known in the CNC world as ringing in the PID controller, and is NOT tolerable there. It definitely should be possible to retune the CMS controller so there is no ringing. That often means tweaking the P down a bit and possibly the I up a bit.

I would also put hard buffers at the ends. Well - padded PU buffers maybe. I doubt you would do any harm to the granite table if you put a couple of small anchors in at the sides to hold them. This is often done to granite tables as the rock is fully stress-relieved.

Now, bringing the runaway axis to a halt without using the controller. That is now standard with modern servo drives. They detect what is called a 'follower error' (the encoder signals no longer match the drive signals), and when this becomes too large they halt and raise an error flag. Have a look at the specs for the Gecko G320 drive for this.

Cheers
Roger

#### Posting Permissions

• You may not post new threads
• You may not post replies
• You may not post attachments
• You may not edit your posts
•

We are the largest and most active discussion forum for manufacturing industry. The site is 100% free to join and use, so join today!