Question about CNC Router limit switc (and/or zero point switch)

[kagliostro]

My CNC Router has no limit switch and I would like to install it

there are many different ways to connect (electrically) the switch

Which is your preferred arrangement ?

Someone has councils and/or a schematic for me ?

Thanks

Franco

[kagliostro]

OK, may be a pair of pictures will help to understand what I'm asking for

as told there are many ways to connect (electrically) the Home & the Limit Switch

in this example each Home swithc is connected separated from the other and the Limit switch are connected in series with the Stop button



while in this example the Home switch of each axis is connected in series with the Limit switch of the same axis, the Stop button is separated


Those of you that have a CNC which arrangement adopt ?

Thanks

Franco

[Heinz]

Hello Franco,

I may be wrong but I think the serial arrangement of the limit switches and the panic button in the first picture does not make sense. You should connect them in parallel so that each switch triggers a stop. With the serial connection the stop will only be triggered if ALL limit switches are pressed AND you hit the panic stop. I assume that is not what you want.

The second circuit seems a bit more logical. The panic stop can be triggered individually which is good. However, I don't think that the home and limit switches are in serial. If I understand the basics correctly home and limit switch should be at opposite ends of each axis. This would make it impossible to press both switches at the same time so with the series connection you'll never get a signal. Having home and limit switch of each axis in parallel, doesn't make sense, either, since you'll never know whether you're homing (desired) or limiting (undesired, stop). Even having multiple limit switches per axis in serial is very unlikely as all must be pressed at the same time in order produce a signal. Having multiple parallel limit switches per axis makes more sense from an electrical point of view although I can't think of a mechanical setup that would require them.

The first example with parallel switches looks like an efficient and economical solution to me (only one input pin required) unless you want to be able to distinguish the reason for the stop, e.g. for an error display.

[kagliostro]

Ciao Heinz

Thanks for the answer, the switch I've under hand has both

but I think that the connection on the switch that is used is the NC (Normally Closed) not the NO (Normally Open)

This changes things or not ?

It will be interesting to know how Doug has the switch arranged on his CNC

Franco

[Heinz]

You're right. If the switches are NC then it will work. I had not thought of that.

[shooter]

Fwiw, all the medical systems I worked on, primarily used || design with redundant circuits, typically opto switch, backed up by mechanical switch.  The E-stop always over-ruled everything.  G.E. did have a series string for all *door* interlocks, around 30 in series, over 3 rooms, with NO indication, so about 90% of the *non-critical* switches were by-passed.

[kagliostro]

Thanks for the info Shooter

Franco

[Heinz]

I think the main plus of the series connection is the fact that any failure - including a break in the wire - will trigger the stop.

[kagliostro]

I agree


Franco

[PRR]

Normally closed (kill on open) is widely used in alarm systems.

Window-tape (metal foil around the glass) is naturally normally closed (opens when glass cracks through the foil), no real choice.

Normally closed also reduces the building wiring. You have ONE wire going around the building and back instead of two.

And normally closed is "fail safe" in the sense that if a wire breaks, the alarm goes-off as soon as you set it, or the machine won't run.

Some other sensors are naturally normally open, and an alarm system will have additional terminals to sense that.

None of this is important to a small machine. Mechanical switches are often both NC and NO. Wire cost is small.

I would have thought they would run individual wire-pairs to each switch. Then the main display could say "X limit hit", "Red button hit" instead of just "fault". But I guess on a small machine you can quickly see which motion has gone too far. And individual pairs greatly increases the number of screw terminals. If they give you a single "kill" circuit, you can series-up all your panic and limit switches so if any one goes OPEN the machine stops.

[kagliostro]

Thanks PRR

Franco

[G._Hoffman]

Mine are all normally closed, normally high, and all in one loop, so it only takes one input.  It works great!


Gabriel

[kagliostro]

Thanks Gabriel

Franco

[G._Hoffman]

Thanks Gabriel

Franco

The idea is to make sure there is current flowing through the whole loop, so the chip is sure to notice a change.


Gabriel

[EL34]

I use probotix boards

All the NC switches have their own connections on the board
They are not in series

In Mach 3, whatever switch gets opened shows on the display
If switches were in series, you could not tell which switch was triggered

[kagliostro]

Thanks also to you Doug

Franco

[EL34]

I have two NC switches for each of the X, Y  and Z axis

The two X Limit switches are in series
The two Y Limit switches are in series
The two Z Limit switches are in series

All the switches go to their own pin on the breakout board
The other side of the switches goes to the common ground pin on the breakout board

That way, if either one of the two limit switches for each axis gets opened,
Mach 3 shows it on the display for the axis that has reached it's limit

I use lever switches that can be NC or NC
You can set up an adjustable screw that touches the lever
That way you can set the switch open point very precisely with the screw

[G._Hoffman]

In Mach 3, whatever switch gets opened shows on the display
If switches were in series, you could not tell which switch was triggered

I've never had a problem figuring out which one was just moving, and having them all in series means I can use the other inputs for other stuff.


Gabriel

[EL34]

I have had problems before with an axis triggering a limit switch

It's nice to actually know which one is causing the problem

[shooter]

It's nice to actually know

+1
the more complex you get, the faster you can find n fix it.  Dougs *redundant* switching per axis is good practice, switches are pennies, motors and gears and possibly limbs are expensive

[kagliostro]

Thanks to all

Franco

[PRR]

> switches are pennies

Yes (many pennies for good ones); but controller INput screws are another cost and size issue.

It's not a big deal. A DIY CNC operator can spot a problem without clues from the controller. In a mass production operation with low-pay workers, more feedback and more log data might improve production, but still would not run itself and clear all jams.

[EL34]

Make sure you use precision switches
The ones with the lever action in my pic above work great
The contacts make and break very cleanly with no bouncing

I used some plunger type switches when I first built my machine
They triggered themselves from vibration once in a while
They also had less than accurate make and break contact points
It really sucks to have a limit switch get triggered in the middle of a job and nothing was even close to the switch to make it trigger

Here's a shot of Mach3 as I pressed on one of the X axis limit switches
Notice the yellow leds that are lit up

[kagliostro]

Thanks for the info and the photo

--

I got some switch at a Ham Fest, they are brand new and acting them with fingers seem of good qualitythey

are not too small, has lever with a small wheel on the top (now I haven't it under hand but I'll make a photo and post it)

--

On the CNC, for now, I arranged only the ass Z switch attachment and I've installed two screws as to be able to have a fine regulation of the switch intervention


EDIT:

Here the photo of the Matsushita switch and how I've arranged the screw for fine regulation











Franco

[EL34]

Looks good franco
Do you have extra nuts to lock the screws in place?

[PRR]

> photo of the Matsushita switch

A fine old design, still the go-to switch for many jobs.

Invented around 1932, MicroSwitch company formed in 1937. VERY common in 1950s-1960s, now owned by Honeywell, and never went out of style. Yours is the Japanese equivalent. I would trust a Matsushita to be good-as the originals. (eBay has 10/$1 look-alikes which might not be as good.)

https://en.wikipedia.org/wiki/Miniature_snap-action_switch

[kagliostro]

The hole in the aluminium is threaded and the nut is to lock rthe screw at the right point

---

I got the switch at an electronic fair, there were also other that were cheaper but I preferred this and paied ..... ??? .... 1 or 1.5 € each

Franco

[G._Hoffman]

It tends to be best not to run the switch directly into a stop - it can easily put too much force on the switch and break the actuator, and also makes the issue of switch bounce a bigger problem.  It is much better to actuate the switch with a ramped plane, which is gentler on the switch, and will induce less switch bounce (making it more repeatable).  Like so:


I wrecked too many switches before I finally listened to the folks telling me that one!!!



Gabriel

[shooter]

best not to run the switch directly into a stop

+1
all the "roller" type switches I serviced, and there were LOTS!, used Gabriel's ramp type stop, usually 2 switches and 2 slightly off ramps, beyond that was a mechanical stop, you didn't want Gramma shooting out the end off an MRI

[kagliostro]

I can agree about the use of ramp

my CNC controller is provided of the inputs necessary to the switch

but the CNC isn't mechanically prepared to install the switch

so I must use the best compromise I can

Franco

[EL34]

I have 6 limit switches and all of them are set up with screw stops

Never had a problem with this system
The lever switches I use are very dependable with snap action levers

They have been in use for many years on my machine and my machine gets used a bunch

Running the screw heads into the tips of the levers, there's no way it can damage the switch by pushing too far
The levers on my switches can bend if needed
The switch either works or it does not

Most likely, it's how you set up your switches if you are going to have problems or not
And the type of switch would matter also

[EL34]

I also use the same snap action lever switches on my 3D printer, but without screws

The moving Axis just contact the switches like in the pic below

[kagliostro]

Thanks for sharing this info Doug

Franco