[catahoula]

Hello,

I'm working on this build: https://www.cnczone.com/forums/cnc-d...2-forum-6.html

Per advice on the forum I picked up an Antek 63v 24v 5v combination toroidal power supply. I'm also aware that the common advice is to add a contactor to directly cut power from both drivers and spindle when the E-stop is triggered. Also, the EdingCNC controller has hardware-level E-stop funtionality in addition to software E-stop which is advertised as being "extra safe"

So I have a few options:
1. go without contactor E-stop circuit because the controller has hardware-level e-stop. This means I can use the toroidal power supply directly for all voltages. There will be circuit breakers and fuses of course. Is this a horrible idea or reasonable?
2. use separate contactors, one for the AC spindle power wires after the main breaker and one for the 63VDC circuit to the stepper drivers after the power supply. In this case I have to buy an extra contactor, not the end of the world
3. use separate small 24v power supply mounted on a DIN rail which are upstream of the contactor, so the controller stays powered regardless of e-stop condition. This is fairly simple, but I'm wondering if I lose the benefits of the smooth toroidal power if the controller is powered by a small switching power supply? I think the toroid thing was more about smooth power to the stepper drives but I just can't remember.

I'm guessing #2 is the best course of action but would appreciate any input!

-Brian

[joeavaerage]

Hi,
I've been using Mach for years, first Mach3 and a parallel port and for the last 8.5 years Mach4 and an Ethernet SmoothStepper.
I've never had a contactor to cut the AC power supply to any components.

I rely instead that when I hit the <Estop> button or a Limit or servo fault or whatever that Mach will stop. In thirteen years CNCing it never failed to stop.
Do you actually need a contactor to cut the power?

You might be required by Electrical or Industrial regulation to do so.....if you were selling this machine......but for your own machine there is no legal requirement to do so....and who would enforce
it anyway?

It occurs to me that if you have a contactor to depower the machine it suggests that you do not have confidence that your CNC software solution and/or CNC motion control system will not reliably stop
when commanded eg <Estop>. That suggests that you should get a better CNC software and/or motion control.

Craig

[catahoula]

Yeah that seems reasonable to me. The PrintNC diagrams as well as the Centroid Acorn manual both recommend the contactor circuit, and I understand the idea that the E-stop is not a normal use function but a last resort so having redundancy built in may be appropriate. But the VFD is already going to be controlled by a relay/contactor so it does seem a bit overkill.

[Al_The_Man]

As mentioned, it depends on whether you want to do it "By the Book"?
If you want the industry required way, there should be a copy of NFPA79 out there. "Electrical Control for Industrial Machinery. which is required by NEC & CEC.
There are basically three forms of E-stop that can be implemented, 0,1, & 2. .
All end with the total removal of all motive power, the controller itself can stay 'live' , just that it should be advised an E-stop has taken place.
E-stop is typically not used for regular machine work stoppage.

[joeavaerage]

Hi,
'National Fire Protection Association'???? Why should anyone take any notice.....the US doesn't adhere to its most basic of laws, the Constitution, is there any reason anyone else should
be bothered with them?

The current and active international standard is IEC 60204, last revised in 2021. It defines the stop procedures in a similar manner to the US standard Al refers to.

https://machinerysafety101.com/2010/...op-categories/

IEC 61508 is the overarching standard and IEC 62061 being the machinery specific interpretation of 61508.

If a manufacturer wishes to make and sell machinery then he will need to get a type safety certificate, and for the majority of the world, excluding then US, then the requirements of 62061
must be met, and that in turn requires one of the three stop procedures of 60204.

Craig

[phomann]

Yeah that seems reasonable to me. The PrintNC diagrams as well as the Centroid Acorn manual both recommend the contactor circuit, and I understand the idea that the E-stop is not a normal use function but a last resort so having redundancy built in may be appropriate. But the VFD is already going to be controlled by a relay/contactor so it does seem a bit overkill.

Hi,
I would suggest that you follow the lead of the CNC manufacturers and have your Estop remove AC power from the drives and spindle. Leave the motion controller powered off Its 24vdc or other dc power.

Removing AC power from the drives and VFD is safe and easy to do. There really isn’t a reason not to do it.

As to your relay controlling the VFD, many breakout boards/ controllers use FETs to turn the relay On/Off. when a FET fails, the majority of time it fails in the on state, meaning that the VFD will not turn off if you press a software Estop.

This issue is also the failure mode for solid state relays. It’s why mechanical relays are used in contactors.

If you’re CNC system can’t harm anyone then maybe you don’t even need an Estop, but once you VFDs and spindles then it should be controlled by a proper EStop circuit.

Cheers

Peter


Sent from my iPhone using Tapatalk Pro

[joeavaerage]

Hi,
if a VFD MOSFET or IGBT fails short circuit it will short the DC bus and blow the fuse/breaker/ contactor quick smart. No need to turn it off, it will do that itself in milli-seconds!

I would agree however about solid sate relays. I tend to use them on pumps, light circuits etc. None of which I would call safety critical.

Much of IEC 62061 and IEC 61508 is about assessing risk. If a machine fails and it highly likely to kill multiple people....then a very strict safety regime is warranted.
If however no injury is likely to occur then a more relaxed attitude is acceptable. The IEC standards address and characterise such risks and specify the minimum safety regimes
required.

Just as a matter of interest, both Al and Peter have often commented favourably about the industry standard idea of an Estop depowering a machine. Do either of you gents have
door interlocks to prevent the machine from operating with the doors open? The worst, or at least potentially the worst accident I've had is where a smallish piece of sheet metal
came adrift and started spinning with the tool and eventually flung itself outward at blinding speed. It hit me dead centre of the chest. Had it hit me in the face I may well have been
injured rather more than the contusion I actually got. I had the doors open....poor practice. Door interlocks would have prevented it....if I didn't deliberately override them!!!

I think the bottom line is that it is the human operator behaving cautiously and consistently cautiously that prevents accidents.Good design is required, things like an electromagnetic
brake to prevent the Z axis from slumping in the event of servo power loss,for example. In designing my control system I have it such that it fails 'safe'. Should the motion control fail
or the PC go on the blink, the machine will stop naturally. I've never had a failure where either the PC, the motion control or any of the controlled servos/spindle etc have failed "ON".

Craig

[phomann]

Hi Craig,

I wasn’t talking about the VFD Mosfets, rather the small mosfets used on breakout boards used to drive relay coils that control the spindle on/off fwd/rev, etc.

Regardless of that I’m surprised that you are instructing beginners or people asking how to connect Estop circuits to just go with a software Estop button.

And yes, I don’t have door safety switches, but then I don’t have doors.
It’s my choice, but if a new comer asks, if they need safety switches on their doors, I’d recommend that they do if they could be injured by the machine. Whether they do is up to them, it’s their choice.
If they are asking for advice, one can assume that they don’t know what is the correct way to go. I don’t know how dangerous their machine is or how skilled they are. I give them advice on how to do it correctly. Whether they follow that advice is up to them.
The last thing I want is someone who gets injured says “Peter told me a software reset was sufficient “

A lot of people listen to your advice as it’s usually very good, but on this issue I strongly disagree when you say that hobby machines don’t need a proper Estop circuits that removes AC power from the drives and spindles.

On this we will just have to disagree as neither one of us is going to convince the other.

Cheers

Peter.


Sent from my iPhone using Tapatalk Pro

[joeavaerage]

Hi,
so you think it dangerous to not have a machine depower, but its perfectly OK to work without any physical containment of the part or tools or fragments thereof?
Both are legally mandated are they not?

Craig

[phomann]

Hi Craig,
Review what I said.
I said that there should be switches on the doors.
When people ask how it should be done, I provide the safe way it should be done.

This is now just getting silly.

Peter.


Sent from my iPhone using Tapatalk Pro

[joeavaerage]

Hi Peter,
other than some BS rule you have yet to give one cogent reason why you should depower the machine on Estop.

I have repeatedly said that in the years I've been CNCing never once have I had the machine not stop when I've commanded it bit by <Estop> or Limit
or servo fault or whatever. I have asked and no one has ever reported such a situation where the machine has ignored an Estop. So what safety situation are
you trying to solve by insisting the machine be depowered? What evidence do you have that its actually needed?

I do have evidence, indeed personal experience, that suggests mechanical containment of the parts and tool is worthwhile. I have had only one incident
that caused a minor injury, but I have had several ocassions where either the tool or bits of material have flown around the workshop. For my safety
I regard containment more beneficial than depowering the machine.

Craig

[phomann]

Hi Craig,

Sure I did. I said that if a breakout board output get fails in the on state, the spindle will not be able to be turned off with a software reset.

Cheers
Peter.


Sent from my iPhone using Tapatalk Pro

[phomann]

many breakout boards/ controllers use FETs to turn the relay On/Off. when a FET fails, the majority of time it fails in the on state, meaning that the VFD will not turn off if you press a software Estop.

For reference.


Sent from my iPhone using Tapatalk Pro

[joeavaerage]

Hi,
well kool, and this has actually happened to you? Then you have evidence that a depower type Estop is worthwhile.

I, on the other hand have never had such a situation. For 11 of the 13.5 years I've been CNCing I was using two of your own (MB2) breakout boards, and not once did they ever let me down
with respect to any Estop event. If I recall the outputs of your breakout boards are CMOS IC's....and yet never a problem. I did manage to 'blow' a couple of inputs over the years, my own fault
on each occasion, but I never had Mach fail to signal a stop event for which the breakout board ignored.

I've built my own breakout board for my new mill, and its been in service now for 2.5 years, and neither has it ever failed to stop the axes or stop the spindle when so commanded.
The ESS signals the breakout board, a CMOS buffer IC and that in turn switches a 2N2007 MOSFET to turn ON the spindle relay. It not impossible that something could fail that would
preclude the relay from turning off, or the relay contacts could get 'sticky' and remain ON despite the armature of the relay going to OFF state but its not done so yet.

Craig.

[phomann]

Hi Craig,

No not personally that I can recall but, I have corresponded with multiple customer who have experienced the problem. For example, below is the most current email I have from a customer.

Hopefully that is enough evidence for you to be convinced

From: XXXXXXXX
Email: xxxxxxxx

------------------------------------------------------

Hello, I am currently having an issue with my Gecko G540. Both outputs are on all the time, as soon as the g540 is turned on the outputs and relays are active. This happens even when the PC is not connected so I'm pretty sure that it is not software related. Have you encountered this before and think the g540 might be damaged? I was hoping to get an expert opinion before replacing the whole thing or the motherboard. Looking forward to hearing your thoughts. Thank you,

------------------------------------------------------

Office Use Only:
From: XXXXX
Email: XXXXX
Login Name: Not logged in
Login Email: Not logged in
IP Address: XXXXX
Host Address: XXXXX
Date and Time: Fri Dec 8 2023 1:25:10 AEDT



This email address was given to us by you or by one of our customers. If you feel that you have received this email in error, please send an email to contact at homanndesigns.com Change at to @

This email is sent in accordance with the US CAN-SPAM Law in effect 01/01/2004. Removal requests can be sent to this address and will be honoured and respected.

Cheers,

Peter

[joeavaerage]

Hi Peter,
that could end very badly indeed. Do you think it a fault with the G540 or some other factor?

When I designed my breakout board I specifically designed it so that should the PC turn off or otherwise lose control of the ESS, the ESS lose power, or even the breakout
board lose power the servo Enable output and the spindle ON/OFF output would go low and therefore the servos would all stop as would the spindle. This does rely
on the servos disabling and falling idle, and the VFD coasting to a stop, if the outputs from the breakout board go low. To date they have never failed to stop
when commanded to do so.

Indeed, you may recall I emailed you a while back about the pull-up resistors on your MB2 outputs 1,14,16 and 17. I was troubled that the machine into which I was fitting the MB2 would cause the VFD
to operate should either the ESS and/or the PC lose power. In particular I removed the pull-up resistors off the breakout board such that they are naturally low when uncontrolled by
the ESS and Mach which suited my purpose. I do not want a 5hp three phase motor starting up without a specific command to do so.

Do you suppose the chappy in the email has such a clash? ie a critical output pulling high because of some downstream control not operating. I can well imagine even if you had an Estop
with a contactor such that the machine depowered on an Estop, it would not necessarily prevent this fault.

Craig

[Al_The_Man]

I don't agree with the BS inference, the fact is that NFPA79 and the EU have similar regulations that are there to protect both operator and machine.
When I build an electrical enclosure for a customer, I have to follow these rules, and it is no big deal, Yes, if the machine has operator doors, specialty E-stop switches designed for this can be fitted.
The same with specialty start Palm-Buttons, wherein an operator is required to place both hands on them in order to automatically operate a manually loaded machine process.
I always leave power to the controller and/or PLC, they are advised of the E-stop, just the outputs are disabled,

[joeavaerage]

Hi Al,
have you ever issued an Estop, be it a manual pressing of the button, or a Limit event or other alarm and had the motion control fail to stop?.
It seems to me that you are advocating that an Estop depower the machine against the probability that the motion control will fail to respond.
Have you personally ever had that happen? I have not. Peter has not had a personal experience of that happening, although the email he attached certainly
suggests that it could happen.

In IEC 61508 there is a matrix with the probability of a failure event on one axis and the likelihood of injury/fatality on the other. The matrix results is a class
of risk ranging from I, II,III etc to about XII. The class of the risk determines the mandated safety protocol.

It occurs to me that the probability that a motion control does not respond to a stop command is very low indeed, and therefore the class of risk ie something like IX or X.
You and Peter seem to believe that the probability of a motion control failing to stop when commanded is much higher and therefore the class of risk is say IV or V, for which
a depower strategy is indicated.

It comes to the basic question 'What is the probability of a given motion control failing to respond to a stop command?'. My estimate is vanishingly small....what is yours?
You can see I have given this a great deal of thought, and I have come to the conclusion that the class of risk is much lower than you guys seem to think, and that means that
someone is not evaluating the class of risk according to IEC 61508 correctly. It seems probable this is a case of institutional inertia.

Craig

[metalmayhem]

For 11 of the 13.5 years I've been CNCing .

in those 13 years you've been machining what machines and controls have you used that do not cut power when estop is hit .

[joeavaerage]

Hi,
I used first Mach3 and a parallel port and then for the last 8.5 years Mach4 and an Ethernet SmoothStepper. Over that period at no time whatever
have I pressed <Estop> , or had a Limit induced Estop, or servo alarm induced Estop where the machine failed to stop.

About 13 years ago there was a bad earthquake in Christchurch where I live, over 200 people died. I can say that the chance of a fatal earthquake is higher
than the chance of my motion control not stopping when commanded to do so. Statisticians are likely to jump up and down....but consider this there has been a fatal
earthquake in the same time where I have never had my CNC motion control fail to stop when so commanded.

Al is a long time user of Galil motion controllers. For many years they have been the standard against which other motion controllers are measured....and thus I'm very interested
to know in all Al's long experience whether he's ever had an occasion that a Galil motion control failed to stop when so commanded. My guess is that he has not for otherwise
I feel sure he would say so. If that is indeed the case then is a depowering strategy necessary....aside from an outdated rule?

Craig