Emergency Stop Circuit

This code will allow you to have an external loop with your power to your drives. I guess that's is what you want. Then there is the Linuxcnc estop circuit which takes an input from the external latch. If linuxcnc throws an estop via whatever, it will drop out your external loop via the output. If your external loop drops out it will fire linuxcnc estop.

The catcher is you need a button that will start your external latch, which could be your cabinet on button, that's what I do.

Just some suggestions,

I am not an expert on this so now that is said.
The VFD has an option to do DC braking on the spindle. I want to use that.
"Real" vfd's for spindles are safety-aware and can go to a safe stop with STO (please google).
My vfd can't, the best it can do is to brake. I'll let it do that, then switch off the power.
So the safety relay will trigger the DC break of the spindle for a settable <3sec period
then switch off the power to the vfd. There is a relay KM6? that forcefully triggers the stop signal
to the vfd. And if that doesn't work for some reason, the power will be shut to the vfd after 3 secs.

There are dual contactors in series to break the power, which I have been advised to use.
And dual E-stop circuits as well.
I think the scenario is that "nothing works" and I just want it to stop. It's the delay that
pushed me to use the safety relay.

The software is QElectroTech, can be downloaded and used without cost.
As always, there is a bit of treshold with new software. It took me maybe somewhere 2-3 hours
to sketch this up

I understand why you would want to remove power from the drives. I do not do that, but that is a preference thing.

Why you want to remove power from vfd and not let Linuxcnc handle the estop I don't agree. e-stop will shut off your vfd and let it brake. And Linuxcnc will be at a known state. If you dont trust Linuxcnc to handle an estop and shut the spindle off, then whats the point? You will be resetting errors in Linuxcnc gui.
The safety switch is in alot of machines I work on but the controller handles the estop also. The safety switch does the same thing you can do with relays and inputs.

Just my 2 cents.

I think I may not be clear or maybe you are responding to some of the other posters here but I do not want to have two seperate loops by any means. In fact the whole original purpose of this post was trying to get some ideas to integrate LCNC control to my Estop circuit.

The idea would be that at power up the control would have power boot completely and once that was done would provide a ready signal to the estop circuit. At that point it would require pressing a machine ready button then latch the main contactor providing power to the drives.

In the case of a software related alarm LCNC would generate an "estop signal" (or w/e I want to call it) that would then open the circuit and kill the mains through the contactor to the drives.

In the case of a physical E Stop the NC switch would open, breaking the circuit dropping the contactor and main power. this would also generate an alarm on the contol. That alarm would prevent the machine from being able to power back on without clearing the alarm and pressing the machine ready.

Drives should have power when estop is reset not when the machine on button is pressed. (machine-is-on)

When machine is on button is pressed, the drives should then be enabled, either by mesa ena or by having a relay follow machine-is-on and that relay enabling and disabling the drives.
Estop will take card of turning machine off and disabling the drives and in the case of removing power from the drives the drives will drop out. But the drives should have power before machine-is-on.

My 2 cents,

I think that is really the question. Do you trust linuxcnc to be operational and thrusthworty of the e-stop?
Even when the display hangs for some reason and the program is running, it doesn't
react to the "program stop" button and the spindle is at top rev.
After some beers I concluded I don't trust linuxcnc for such a vital task. And my design follows from that.

It is just my take on it. Hopefully you get more confident in your way of acting after giving this a thought,
I don't intend to change your mind. After all I get to understand that real safety includes risk assessment,
risk mitigation and and a number of other things. And i haven't been there. I have just drawn up a solution
that suits my immediate thinking. Hmm, but I think I am happy with it. I just wanted to share and raise some thoughts.

My current standing is a combination of both. Allow the control to trigger it's software estop and then allow that to trigger the physical estop circuit or trigger the circuit directly via mechanical switches.

I won't soley rely on software because there are to many risks. If the software had some kind of glitch causing the machine to run away you will be thankfully to have an Estop circuit that drops drive power.

I think the biggest question now is do I allow for a sub second delay to allow the control to try and stop motion prior to dropping drive power?

And how do you implement that?

What I have learned is that a safety relay is the way to go.
You should get one with a delayed output, just like the allen-bradley msr138 dp (many variants, the 440R-M23143 is the one I went for). It will break two contactor circuits and also allows for 0-3 sec delayed contactor breakage, .

There are no problems to let linuxcnc act as a NC switch in the estop chain. There are actually two, duplicate estop chains. And double contactors, in series. You decide where the limit is.

There are of course many safety relays out there www.google.com/search?q=safety+relay+del...q=safety+realy+delay

When it comes to what you actually can do, you are limited by the outputs.
There is one Hi->Lo at 0sec and one Hi->Lo at say 2.0 seconds,

If you want to stop your slides in any way, you get the first signal at 0sec: Hi->Lo
you need to design the servo drives' control or whatever, to start braking your slides at this point.

The second, tough point, is at 2.0 sec where the power to the slides/axis is shut.
Well, if if done right the power will be switched off now no matter what.