Limit Switch choices?

[duwayne.]

Well I got most of my servo mounting parts made. I still have work to do on the ball nut mounts and the new linear rail head mounting slide plate. But anyways, I did get my servo's and drivers and power supply last week and I'm working on my electronics enclosure. One thing I still have to decide on is limits switches. I am familiar with what comes on the IH turnkeys. I've seen williamsmoto's machine. I have been looking at industrial grade lever action limit switches as one possible option. Then I was looking at these inductive limit switches that CNC4PC sells. I was wondering if anyone has experience with those. They also sell the optical ones as well. I have the C32 rev3.11 BOB, so it's pretty much sky's the limit as far as what I want to hook up. I was also curious as to why some of these limit switch kits they sell have pickups for each end of travel for an axis, such as the inductive kit from CNC4PC which has 5 pickups? Is it just a mounting option as far as how the pickups get activated? IE 2 switches 1 activator. I would think that one switch per axis would suffice with a trip block or activator at each end of the travel would give the same results. This is an area that I need to be educated a little on.

Thanks guys

[jetflatline]

After far too much fun with optical sensors I'm switching to mechanical switches on my lasersaur. I thought the optical ones were better but they're much harder to trigger reliably with acrylic triggers, even once they're layered in electrical tape.

The mechanical ones also are easier to diagnose as working or not, you can hear them click and read them with a simple multimeter.

[wildwestpat]

Hi Duwanye

There are several switch technologies each has its own list of pros and cons. The choice is to a large extent down to what is likely to contaminate the switch. For example a reed switch and magnet system could have problems with steel swarf collecting arround the magnet - a light beam break switch might have problems with dust if MDF were your bag. I found inductive sensors on eBay from China at £6 each and these are proof against coolant and since they do not use magnets are ok with steel chips.

Hope this helps - Regards - Pat

[RustedOut]

I built my own using Optek slotted optical switches. Had lots of issues early on until I found that the voltage separation between high and low wasn't enough to satisfy the break out board.

Installed pull down resistors on each switch at the board end and they've worked perfectly since. They repeat very well, less than .001". You have to keep the light out of the switch housing and I cast my own gaskets for the covers to keep everything out. Use flat black painted aluminum flags.

I even designed my own circuit boards to mount the components on and etched them. First time I'd ever done anything like that. Housings were one of the first parts I cut on the mill after converting it to cnc.

I use one per axis with two way switch rods. You have got to use twisted pair wires all the way from the breakout board to the switch. Keep them twisted all the way. Otherwise the induced noise will kill you. Biggest thing though was the voltage differential. Took a long time to figure that out even after I knew the high and low voltages. Didn't make the connection. I'm no electronics guy, so I work from a level of stupidity there. Every thing comes hard in that science.

They are actually a fairly easy project and a fun one if you can control the level of frustration when they don't work out of the gate. Glad I did them as I learned a bunch.

Bob

[duwayne.]

So I take it that whether you use 5, or 6 inductive, or optical's, or mechanical's with one actuator for each axis. Or if you use just 3 switches and 2 actuators per axis the result is the same. According to the diagrams I've been looking at they all get wired in series anyhow. I guess it's a mater of preference. The cost factor of the 3 different types that I've been looking at is about the same. Spent part of the day today working on my electronics cabinet. Here is a few pic's.

I ended the day working on the panel for all the external connectors. I am using the snap in panel where all the ports from the computer's mother board used to go out. I'm putting in 4 of the round 4 pin connectors for each axis motor and 6 RJ45 jacks for the limits, encoders, a probe and a future VFD control.

[jetflatline]

We're still developing our controller board but I find having unique switches at the + and - ends of each direction has been nice. If you hit +, you know you can back off in direction - and if you drop to -,- you know you're near zero then can edge a bit towards +,+ to get off the switch.

[RustedOut]

So I take it that whether you use 5, or 6 inductive, or optical's, or mechanical's with one actuator for each axis. Or if you use just 3 switches and 2 actuators per axis the result is the same.

I guess I'm not following all of that. One actuator rod per switch per axis is what I'm running. Here's a pic of the X axis switch assembly. There is a stop block along the front of the bed on each end to bump the rod and move the flag. The other two switches just have longer rods and a bracket that travels along the rods. Pushing or pulling the rods when contact is made. It's pretty much just like the IH system.

According to the diagrams I've been looking at they all get wired in series anyhow. I guess it's a mater of preference.

You've only got so many inputs on the breakout board for such things. I used 3, one for each switch. So, moving in a + direction or a - direction doesn't matter. I know what direction it was moving when it hit a switch and I know what direction to go to back off of it. Mostly anymore they're only acting as homing switches as I rarely hit a limit switch.

You can add more inputs by adding other boards, but I'm cheap and chose not to go that route. And like you say, you can wire them all into one input and they'll work the same. But the led's in Mach won't be right, if you care about such things. If and when I need more inputs I may go that route. Here's the little pcb I built for them. The twisted pair cables aren't mounted in that pic yet. These are just single sided pcb's. 2 diodes, a resistor and 2 switches.
I went with two optical switches per board for a reason I don't recall now. But I think it had to do with movement of the flag, or maybe not.

Bob

[duwayne.]

I guess I'm not following all of that. One actuator rod per switch per axis is what I'm running. Here's a pic of the X axis switch assembly. There is a stop block along the front of the bed on each end to bump the rod and move the flag. The other two switches just have longer rods and a bracket that travels along the rods. Pushing or pulling the rods when contact is made. It's pretty much just like the IH system.

You've only got so many inputs on the breakout board for such things. I used 3, one for each switch. So, moving in a + direction or a - direction doesn't matter. I know what direction it was moving when it hit a switch and I know what direction to go to back off of it. Mostly anymore they're only acting as homing switches as I rarely hit a limit switch.

You can add more inputs by adding other boards, but I'm cheap and chose not to go that route. And like you say, you can wire them all into one input and they'll work the same. But the led's in Mach won't be right, if you care about such things. If and when I need more inputs I may go that route. Here's the little pcb I built for them. The twisted pair cables aren't mounted in that pic yet. These are just single sided pcb's. 2 diodes, a resistor and 2 switches.
I went with two optical switches per board for a reason I don't recall now. But I think it had to do with movement of the flag, or maybe not.

Bob

I guess I was getting confused when looking at the offerings from CNC4PC CNC4PC and CNC4PC. I'm assuming they mean for a switch to be placed at each end of the movable travel of an axis and then for you to use 1 actuator at a fixed position. In other words the switches would travel with the table, saddle, or head and would get tripped by a single actuator for each axis mounted at fixed locations. Like what you said about your setup is what I understand of how it should be setup, no matter what type of switch you use. If I were to use mechanical switches I would use a single switch for each axis and 2 buttons, or blocks at each end to trip the switch lever. Also like you said it's mostly for homing the machine and you can see where the machine is, if you hit a limit and have to back off.

I think I understand it's more for Mach3 knowing and or showing which end of travel has been tripped when using 5 or 6 switches and like you said Bob, it would require more inputs to be wired. I thinks I will try and follow the KISS rule for now. If I feel a justifiable need for more than 3 switches I will address it then.

This is one of the mechanical switches I was considering. AEM2G42Z11-3 Products.

Thanks for the help guys, it's always appreciated.

[RustedOut]

I certainly can understand the need for multiple switches per axis on larger machines and the z travel of my mill is pushing it a bit with the long actuating rod, but it works real good.

I was real hesitant to go with a mechanical switch up front not knowing the direction my milling was going. I cut a lot of different materials, but don't run coolant per se. I do spray both steel and aluminum and an upcoming project will be a micro drop system. So now I can see a quality mechanical switch could have been a candidate if coolant was the only concern.

My one shot lube system does oil down my Y switch a bit, but the other two stay dry and clean. I run an enclosure around the table to contain chips, but the back is limiting and going to come off for good. It has done a good job of keeping chips away from the switches, but with that removal the Y and Z switches become exposed. And I cut a good deal of MDF, so dust would be a real problem for some types of switches. Especially the second ones you linked to from CNC4PC if they were exposed. I'm running a C11G board from CNC4PC, so I had seen those back when I ordered the board.

My first round with my switches I used remnants of cable from my encoder cables so that was a cost I didn't have to bear. But for some reason I kept messing up one of the switches as I installed it, burning out the switch. Remember electronic idiot here. I ended up buying more cable and that was the real cost of building mine. All the other components are dirt cheap. A couple diodes and a resistor. Here's the switch I'm running now, but these would work as well I think. Just a different voltage handling as I was intending to move to a higher voltage thinking that would solve the problem of intermittent switching. Never got there as that was about the same time I figured out the voltage differential issue.

There was some discussion way back of running the higher voltage through the switches to combat noise, but in my case noise wasn't an issue due to my twisted pair wiring. I pull 5v off of the board and feed the led it's required 1.7v through the resistor and feed 5v through the collector through the second of a pair of twisted pairs. Back to the board and hook up a pull down resistor to lower the signal real close to 0v. If I recall correctly, the max off state voltage for the board is .8v and my low signal originally was around 1.3 or so. The high signal was only a bit under 3v I think. That was the real issue with my switches all along. No noise, no spikes, no switch failures (other than my own incompetence), just a voltage deal. It was easy up front to blame other things, but good trouble shooting prevailed. Of course it took me months to get to the end of it. Did I mention I'm slow too.

Bob

[ninefinger]

Hi Dwayne,

You may want to consider inductive switches from DMM http://www.dmm-tech.com/Files/DMMTec...st_Rev.1.1.pdf ( HW-04-NPS) - $10 each and very good quality, fully sealed so you don't have to worry about coolant / oil etc. I bought 3 but still have to install them - been too tied up with other things They do have an LED built in on them to help with setting up so you can see when they are triggered.

Mike

[duwayne.]

Hi Dwayne,

You may want to consider inductive switches from DMM http://www.dmm-tech.com/Files/DMMTec...st_Rev.1.1.pdf ( HW-04-NPS) - $10 each and very good quality, fully sealed so you don't have to worry about coolant / oil etc. I bought 3 but still have to install them - been too tied up with other things They do have an LED built in on them to help with setting up so you can see when they are triggered.

Mike

I was looking at a set from CNC4PC. They have a 5 pickup kit with the circut board all plug and play for the C32 BOB that I have. All I would have to do is mount them and plug it in. Have not decided yet on what type of switches to go with yet. Kinda leaning towards mechanical switches.