[Fastest1]

You might get lucky and never have a case of over travel. I and many others werent.

Or when setting up home and limits, you turn them off for a moment and jog the wrong way. Your switch and housing are destroyed.

Though I like the attention to detail you 2 are performing, in this case the switch either needs to be activated by a ramp that can pass by while simultaneously pressing the switch. Or use an opto switch with a gate and blade arrangement. Or even a proximity switch. Those will allow for over travel.

[atomarc]

You might get lucky and never have a case of over travel. I and many others werent.

Or when setting up home and limits, you turn them off for a moment and jog the wrong way. Your switch and housing are destroyed.

Though I like the attention to detail you 2 are performing, in this case the switch either needs to be activated by a ramp that can pass by while simultaneously pressing the switch. Or use an opto switch with a gate and blade arrangement. Or even a proximity switch. Those will allow for over travel.

It's interesting you mention the over-travel thing. I have installed many limit switches over the years and have always used a ramp or toggle arm..never straight on as we have done. Nick was adamant about mounting them in this fashion, making the point that the machine rarely looses counts and can reverse in tenths of an inch so "not to worry" Dad! They are at the very extreme ends of the axis travels so I can't imagine jogging into them and am not sure why a person would turn them off as you would crash the machine into itself...but I haven't done this before and don't know.

Can you give me an example of how this could happen. Why would we turn off the home switches? If this truly was a error we have plenty of time to change the way these are tripped..thank you for the heads up, I appreciate it.

Stuart

[109jb]

If you hit a limit, then you go into the controller and override the limits to jog off the switch. Jog the wrong way and....

[atomarc]

Why would I ever want to override the limits...that is what I don't understand, they are at the end of travel. If a person did this and jogged back into the limit switch, when the switch was actuated would it reverse the movement the same way it does when homing the machine. I'm not arguing, just attempting to understand how the software works that you might drive through the switch. I guess if it's possible to blast the switch under the worst case scenario then it's a moot point that it could happen..right! Can the software be configured so that it can't happen?

Stuart

[109jb]

If you inadvertently hit a limit you have to override then to get back off of the limit. It happens. Imagine setting up a part that uses the extreme travel of the machine. You will be very close to the limit switch and a simple bump of the job will send the machine to the limit, tripping the switch. Now the machine can't move because it is on the switch. You have to override the limits to jog away from the switch to resume normal operation.

Sent from my Z958 using Tapatalk

[Fastest1]

What he said ;-)


Sent from my iPhone using Tapatalk Pro

[Fastest1]

In general it will protect the switches.
It is in the example 109 gave that will most likely occur.
I am not near as accomplished as you 2 appear to be. I learn things the hard way.


Sent from my iPhone using Tapatalk Pro

[atomarc]

Just to be clear on this, the switches pictured in the post are home switches and they are at the extreme limit of the travels. There are no mechanical stops on these travels so if a inadveertant jog was able to be initiated, and it was in the wrong direction, a mashed home switch might be the least of my worries. I watched a video by "Daz" on the Artsoft forum dealing with setting home and soft limits. The machine he uses for the demo has roller limits that are mounted like ours are...dead on, in harms way.

While his demo doesn't show it, I'm hoping that it's possible to have the machine go to home, zero, then return in the opposite direction some programmed distance. This gets it off the home switch and allows room for a screw up if one were to jog in the wrong direction. I'm also hoping/assuming that any jog command cannot go further after hitting a home switch. Maybe the problem here is the jog speed versus the approach speed when homing.

What am I not understanding? We want to proceed with the install on the home switches, but if it's more sensible to change the method or switch style, now is the time to do it....HELP!

Stuart

[DJEkim33]

You are on the right track. But what everyone is trying to say is that sometimes, especially with a new machine you can have an issue where you over run the limit switch. Whether you lost steps or had something come loose it "could happen". You can home to the far extreme. Which I would establish the switch location about 1/4" off the far mechanical extreme so you have some wiggle room. After it homes you can set it up to return to "0" which you set at any distance away from the switch. With a soft limit at the opposite end of the table. What I would also do is set a soft limit near the home switch of say .1" or so. It would look like this assuming 10" of total mechanical travel.
Far - mechanical travel is -5" (physical limit)
Home switch is at -4.75" (set in programming)
Soft limit is -4.65" (set in programming)
Zero is 0 (set in programming)
Soft limit at 4.75" (set in programming)
Far + mechanical travel is 5" (physical limit)

The numbers can all change with in the envelop of your said travel like my 8" y axis (following the above order) -5.25", -5", -4.9", 0, 2.65", 2.75". Now atleast with linix I can move my table within the -4.9"to 2.65" range and never trip a limit switch. When I hit home it will go past the soft limit to hit the home switch then back off and rapid to 0. I use proximity switches with a 15mm range so the wont be hit if they fail. I just crash the axis. Just a reminder linuxcnc is free and to get the equivelant of smooth steppers (mesa cards) and mach 3 you are about half the cost.

[atomarc]

DJEkim33,

Those are encouraging words, thanks. If the chance of overrunning the home switch and mashing it is just a case of some mechanical component failure I'm not too worried. The basic question I have is..will the software, when properly configured, NEVER allow you to move beyond home switch '0'.
Also, it sound like you are saying that after a home command you can program the machine to return to a selected position which would of course be in the opposite direction that it homed...is that correct?

If a operator had a piece to machine that was near the capacity of the machine and he was jogging, say the 'Z' axis upward to clear the piece, and he inadvertently came to the home switch, would that stop the jog function? If that's true, would the jog speed be too fast for the home switch and machine to respond and stop before mashing/crashing?

I hope these questions make sense!

Stuart

[DJEkim33]

Yes properly configured you can jog anywhere without hitting a mechanical limit or limit switch. Assuming your first action is to home the machine which it should be. Now you could over run a soft limit if you lost steps. But that is obviously something causing the issue. Along the same thread with the setup I mentioned you can jog anywhere and never actually hit a limit switch. You will be bound by soft limits on both sides. When you reach the soft limits the machine will stop jogging that direction and you can jog the opposite way without having to override a limit switch.

Yes you can program the machine to go to wherever after a homing operation. After going the axis will go to "0" so you can make that spot where ever by programming. You will be entering the location of the limit switch as it relates to this "0" position. So on a 20" travel you could say limit switch was at -19 and soft limit is +1. It would home all the way to the negative, then rapid 19" away to the far side to the machine absolute origin point or "0". Your can set a work offset and that is different that absolute position so don't get those confused.

The last question is a yes it will stop at the limit, and will stop before the limit if you have a soft limit set just a hair off the switch(this helps to avoid the need to override the limits). You will not be going so fast as to smash the switch. You should not rely on the limit switch to stop the machine though, as it needs to be overridden, and should have a soft limit set so the machine just stops and allows you to go back the opposite way without an override. So the limit switch is actually the second layer of protection after the soft limit and you are really only using your limit switch as a home switch primarily.

As for a work piece, at least with Linux, you can place the piece on the table and locate it. If you run the program and the program would cause the machine to go past it's travel limits a warning will pop up saying so and ask for an override, and if you accept and run the program, you will crash a soft limit on one side or hit a limit switch on the other side. Not sure if machine does the same, if it does I would never see a reason to run a program that caused this warning without adjusting the work to fit inside the travels.

[109jb]

A couple things.

If your intent is to rely only on soft limits to prevent running into the ends of the machine travel, in theory this will work. However, theory and real life are many times different. If your machine is to be run with open loop steppers and you think you will never miss steps, I wish you luck.

If you set those switches up to be ONLY homing switches then running into them will not disable jogging. Most people when using a single switch like that enable it to do dual duty as both a homing switch and also as a limit switch for that end of the travel. The switch is there, so why not ? If used as a limit switch, then tripping the switch will disable the machine requiring temporarily disabling limits in order to jog off of the switch..

When a limit is tripped, the drives are disabled, so the stopping of the axis is by friction, motor drag, etc. How far that takes is machine dependent. Quite a few years ago, I tested my machine to see how far the overrun from switch activation was when jogging as full speed. I don't remember the exact numbers, but on my G0704, it was about 0.040" worst case overrun. I don't recall which axis that was on. I use inductive proximity switches and the gap at activation is about 0.100", so using these as limit switches as well means that even a full speed trip the target would be 0.060" or more from hitting the face of the prox switch. I have my homing routine configured to set the machine zero position 0.050" from where my switches activate but this is just to prevent false triggers at the soft limit boundary.

One caveat is that I run LinuxCNC, but I think Mach 3/4 has a similar logic when it comes to homing and limits.

[atomarc]

So..here is what I get from all the great responses, and what I think I understand!

I'm running Mach4 with a ESS. I have 3 home switches at the extent of the 3 'axes'. They are positioned in a plunger orientation so any gross over travel(+ .040) will mash the switch.

I can configure the software to home the machine to the switches, then return to a position of my programming. I can configure the software to home at a speed of my choosing, possibly homing at a fast rate, then slowing to a crawl as it nears the home switch position.

I can configure the software to have soft limits a bit shy of the home switches so during normal operation the home switches will never be contacted.

A jog in any direction will only go as far as a soft limit....and maybe be programmable to slow to a creep as a predetermined distance from that soft limit.

If all these things are true, then I feel that I can finish installing the home switches in the manner I presented in a earlier post and be fairly safe that they won't be destroyed unless there is a catastrophic failure of some component or a software glitch.

In regards to the steppers missing steps, this is my first rodeo, so to speak. I will say that my son has a small 3D printer that is stepper driven, he has printed hundreds of items, many very intricate, and to my knowledge that machine has never lost a step or skipped a beat so I've got my fingers crossed that "missing a step" my be way overblown.

Any comments or thoughts on what I think I know.

Thanks,

Stuart

[DJEkim33]

If the throw of the switch is greater than .040" then it should not get smashed. 3d printers have significantly less force on the steppers than a mill that is pushing through a cut. Missed steps are real and will happen. I am currently dealing with .018 to .025" worth of missed steps over the course of an 8 minute program in my Y axis only. Fun part is to track down how that is happening. Missed steps will happen, especially early on in your build. But if you build a buffer into your travels you shouldn't see too many issues.

[SCzEngrgGroup]

Missed steps are real and will happen.

Horsepuckey! Steppers are perfectly reliable, if the entire drive system is competently designed and built. Steppers will "lose position" ONLY when asked to perform a move that is
beyond its capability. The same is true of ANY drive system. Lost steps is a result of poor design, or improper operation. Period. I have run many hundreds of hours on stepper-based CNC machines and NEVER once lost a step other than when I crashed the machine. Sadly, far too many people here "design" based on recommendations from what are, effectively, strangers who, themselves, based their "design" on recommendations from other strangers. That will always be a hit-or-miss proposition. Proper design REQUIRES knowing the detailed motor characteristics, the DETAILED machine requirements, and a very good understanding of HOW steppers work More is NOT better. Bigger motors, more voltage, more current, etc. can be every bit as bad, sometimes even worse, than going the other direction. Steppers are used in literally millions of machines worldwide, in applications where a single lost step would be catastrophic, and they operate flawlessly for years and years. Again, steppers "lose steps" ONLY due to poor design, improper operation, or hardware or software problems that require correction.

As for the Home switches, your switches will be just fine, provided:

1) Your hardware never screws up
2) Your software never screws up
3) You never make any mistakes

#1 I'd not be too concerned about, except for your limit switch wiring. A fault there, and you WILL break something. #2, if you're using Mach3 or Mach4, I would not bet five cents on that one... #3... Well, it #3 is true, which I sincerely doubt, you're going to save a bloody fortune on metal stock and tooling compared to the rest of us!

Regards,
Ray L.

[DJEkim33]

Respectfully you can thump your chest all you want but in the context of our conversation, given the people we are (you know the ones who just copy poor ideas) steps will be lost at some point. It could be from a number of things. That bolt that comes loose, something we did not know about in the design or just plain poor design. But we are a bunch of guys learning the craft in our garages in our spare time. So we have to start somewhere and following someone else's trials is as good a place as any. We will learn as we go, the way we learn best, through our errors and imperfections, one at a time.

You too were once a doe eyed student of the craft who I am sure had to learn some lessons along the way. You may have been able to learn from someone far more knowledgeable then we have available to us and by the sounds of it far more time to devote to the endeavors. 6 months ago 50% of this made no sense to me. Today about 60% of it does but I am all ears to hear good ideas and give them a try. The hobby is a stress reliever from my regular job and as such it does a great job of relieving stress and learning some fun stuff I admired as a child. That is the interesting part of the hobby/bench top crowd. It is quite the melody of people from different backgrounds, mechanical, electrical, programming and some people with no relevant background to CNC what so ever. But we are all here to learn.

And missed steps is part of the learning process...

Nick and Stuart have come along way in the last few months as have I but none of us claim to be an expert, we are just here to share the experience.

Mike

[atomarc]

https://youtu.be/HwFD3Re_Ccg





https://youtu.be/7Us8S1g6aoI



.....

[louieatienza]

If the throw of the switch is greater than .040" then it should not get smashed. 3d printers have significantly less force on the steppers than a mill that is pushing through a cut. Missed steps are real and will happen. I am currently dealing with .018 to .025" worth of missed steps over the course of an 8 minute program in my Y axis only. Fun part is to track down how that is happening. Missed steps will happen, especially early on in your build. But if you build a buffer into your travels you shouldn't see too many issues.

If you size the steppers correctly, and don't try to push them to the extremes of their performance, you won't miss steps, period.

That is, the gibs are adjusted correctly without any binding at any point, and lubed, to prevent any stiction.

[gcofieldd]

I know some people have already said this, but putting the limit switch inline with the movement of the machine is problematic. When I am jogging my machine around manually, I run past my limits all the time. It is usually when I am zeroing on a part or working on the machine. Your limit switch will become a mechanical limit. It would not take much to move your limit switches out of the path of travel. I do not think you will regret it.

[atomarc]

Graham,

The switches we are installing are 'home switches' so they are at the extreme limits of all travel for all axis. Per some members suggestions, soft limits will be programmed in front of the home switches so I'm hoping that between those soft limits and the fact that the home switches are so far out of any normal work envelope that we will be protected. But I do agree with you, a better design would have been a roller type of switch with a ramp or a non-contact capacitive switch. Our switches are mounted, so we're going with this design for now...if we have crashes and destroy stuff we'll drop back and regroup.

Thanks for the comments, we appreciate all the help we can get.

Stuart