PM932 CNC Build

[TomS]

Hi Tom,

Glad to hear your project is coming along. The proximity switches are okay. At first they were very repeatable but over time have become less so. I have not spent the time to figure out why yet so the jury's still out.

Yes they are normally closed.

I'm currently working with a couple of friends building a CNC router for cutting aluminum sheet metal. We will be using hall effect sensors and I want to see how they perform before digging into the proximity switches.

Please post some pictures of your progress.

Jay

Thanks Jay. Couple more questions. If I were to leave about 1/8" to 1/4" of over-travel when setting the proximity switches would they be repeatable enough to prevent a crash? If you had to do it over again would you still use proximity switches or go with standard type limit switches? Just wanted to get your opinion because I'm at a point in my project where I need to start thinking about limit/home switch type and placement.

Tom S

 

[jbolt]

The proximity switches work fine as limit switches and will always trigger. The repeatability issue I am having is with homing to within 0.001". My Z homes to within 0.003", the Y within 0.004" and the X within 0.007".

I like the proximity switches for their ease of installation and setup. If the hall effect sensors we are going to use on the router work better I may change to those but they will require enclosures to keep ferrous metal chips away since they use magnets.

I need to do some research on the resolution of the proximity switches I have to see if they are capable of what I am expecting. They may not have the sensitivity I need or my triggers need refining.

I have never used mechanical switches but it is my understanding they don't have the repeatability I am looking for. They would work fine for just limit switches.

Jay

 

[JimDawson]

A flag type photo eye may be a better choice for home switches. Mechanical limit switches are very repeatable if you use the correct switch. A snap action switch is not the correct switch. What is needed is a slow acting switch, much more repeatable and no dead band. Snap action switches work OK for travel limits.

The other problem I have run into is the speed of the homing routine. It should be a 3 step process Move to the switch at moderate speed, very slowly move off of the switch, very slowly move back to the switch. This normally gives a pretty accurate position. The only reason you really need to set the home position is to set the software limits.

The real question is why would you need home a mill at all. Normally you would zero off of the work and use that as the reference. If I need to park my table after an operation, I pick a convenient location and tell the table to go there after the operation is complete. There are other machines that do need to have a fixed home position, like a dispensing machine where the stations are fixed to the table and never change.

 

[John Hasler]

The proximity switches work fine as limit switches and will always trigger. The repeatability issue I am having is with homing to within 0.001". My Z homes to within 0.003", the Y within 0.004" and the X within 0.007".

I like the proximity switches for their ease of installation and setup. If the hall effect sensors we are going to use on the router work better I may change to those but they will require enclosures to keep ferrous metal chips away since they use magnets.

I need to do some research on the resolution of the proximity switches I have to see if they are capable of what I am expecting. They may not have the sensitivity I need or my triggers need refining.

I have never used mechanical switches but it is my understanding they don't have the repeatability I am looking for. They would work fine for just limit switches.

Jay

This could be a software problem. If the processor is polling the switches rather than being interrupted by them repeatability will depend on how busy it is and how fast the table is moving. Even if the switches generate interrupts you could still have a repeatability problem if the interrupts are low priority. This is where real-time design becomes important.

 

[TomS]

Thanks Jay, John and Jim for your input. You've clarified my questions so I will order proximity switches for my conversion. My plan is to reference off of the Y and Z axis ball nut mounts so I can place the switches inside the base and column thus protecting them from coolant, swarf, etc. The X axis is a bit more trouble to mount under the table so I may mount it on the front of the table.

Tom S

 

[jbolt]

The real question is why would you need home a mill at all. Normally you would zero off of the work and use that as the reference. If I need to park my table after an operation, I pick a convenient location and tell the table to go there after the operation is complete. There are other machines that do need to have a fixed home position, like a dispensing machine where the stations are fixed to the table and never change.

I've run a few projects multiple fixtures and work coordinates setup at the same time. G54, G55, G56 etc. Havng a repeatable home position allows me to simply reference "all home" at the beginning of a session without having to set the offsets each time.

My vises and tooling plates are also indexed to the table with pins. The vise edges and tooling plate index pins are a known offset from home position so with a good repeatable home position it is simple to switch between the vises and tooling plates.

On complex parts that require many hours of machining I will reference the home position between parts just to verify there have been no missed steps.

Jay

 

[jbolt]

This could be a software problem. If the processor is polling the switches rather than being interrupted by them repeatability will depend on how busy it is and how fast the table is moving. Even if the switches generate interrupts you could still have a repeatability problem if the interrupts are low priority. This is where real-time design becomes important.

How would I determine if it is a software problem?

 

[John Hasler]

How would I determine if it is a software problem?

Do I understand correctly that the software that monitors the switches and controls the motors is running under Windows XP? If so make it do something challenging while it tries to home and see if the error increases.

To fix it the best you can do is try to minimize the number of processes running while the switches are being read. Uninstall or shut down everything you possibly can. XP was designed for the desktop, not for control.

 

[JimDawson]

How would I determine if it is a software problem?

If you can hang loose for a couple of days, I'm getting ready to fire up Mach3 on a 4 axis machine I just put together. It will be running on a clean install of XP Pro, SP3. I am also using snap action switches for the travel limits and they also double as the home switches. Normally I would not use snap action switches, but they were cheap and good enough for my application.

I'll post my findings and all of the setup information for both XP and Mach3 when I get it going. I don't need extreme accuracy for this machine, but I will check the repeatability of the limits and see what I can do to get them to at least 0.001.

 

[jbolt]

John - Yes it is on XP. It was a clean install of XP pro. The PC only runs Mach3 and is not connected to a network, no anti virus etc. I have disabled as many unneeded processes as possible but there could be some I missed. Not sure what I could do to challenge the software while homing.

Jim- no hurry, I'm too busy with other things to tackle this right now. Next spring I plan on doing a partial tear down to make sure everything is still tight and true. Plus the machine needs a good cleaning. One of the downside of flood coolant.