Woohoo! Another series coming from Marcel! And I think you are right. NO contacts should be in series while NC contacts in parallel. As far as I know, NC is the standard in the industry and you've explained it right why: even a cable break would trigger the safety function. Which in this case is NO because close proximity is the normal condition. For example on a KUKA robot the safety features triggering the E-stop funcion are all daisy-chained on one loop. (An ex business partner left me his failed attempt of building a cnc-plasma, so I decided to finish it and make it work -a couple of days ago, and then out of the blue comes your video... -I have to consider it as a sign.) Keep up the good work!
Thanks for the nice words! And I agree, what you want is an NC situation, but indeed in this case that is the case when you use NO contacts. Anyway, your sign might be right ! But it probably means you are here to help me :) (your machine is already up and running 💪🏻
@@RoboCNCnl I'm here to learn, to think together with you and to contribute if possible. All I have left is a large welded steel frame, some linear rails and a 100A plasma cutter. The project had failed previously because the guy had issues with the high frequency noise he couldn't have figured out how to resolve. So I'm about to follow your lead on the controller and such. This anti-collision device you've found is a huge leap alone.
@@h4z4rd42 make sure you use double insulated/ shielded 5. Core wire so the inside shielding mesh can be terminated to a earth bus bar.... electrical interference has to be collected by the shielding and driven to ground .... literally.... hypertherm have an article on it. Gantry earthing, driving rod into ground etc... and if you controller is usb controlled from computer, you will be having issues forever as usb has too much sensitivity and interference and isnt built to run robots, for which these machines really are... i went with avid centroid acorn controller ... ethernet connection!
Looking forward to following along with this build, I've been thinking of building a CNC plasma table myself, though probably smaller than you're planning. Just a note on the sensors, you cannot (or at least should-not) connect transistor style outputs in series - at least not directly. As they are, the only way to connect them in series is to have each sensors ground run to the previous sensors output, resulting in the down-stream sensors becoming unpowered when the upstream sensor breaks the circuit, as they lose their ground connection. Also note that the (low) output at the final stage, will be somewhat offset from the system ground as each transistor stage has a small offset associated with it. This could end up being at a level too high to be properly/reliably detected, especially in a noisy environment. (you also need to ensure that the sum of the load currents posed by each sensor, do not exceed the transistor output capability, though you're probably okay there with only the 3, so 2 are in the load) This is likely why the manual suggests connecting the NC outputs in parallel. In that case if any one triggers, the input will get pulled low. The downside to the parallel connection is that you can't detect cable break faults, as you could with a series connection. If you had relay/switch outputs, then series connection would be correct for the way you are thinking.
Nice to hear you'll follow along ! and please feel free to engage when you have ideas ! Thanks for sharing your thoughts on the sensors in series… I should do some calculations here, although I will not find out the transistor types. I do feel that since I’m not powering them with TTL level but a 24V and no load (or at least very small) that I suspect to only see the voltage drop of 3timea the PN transition. Anyway… I did not think of it, so it’s a good thing to test, thanks for sharing !
Nice! Your earlier videos were a big inspiration and information source for my build. Thanks for that, I have now many many super precise boat parts done! And I think you’re right with the sensors. It’s different of course because you’re “constantly sensing” the normal situation and if sensing stops, that an error. With normal crash detection it would be other way around of course. I haven’t been able to use sensors at all so far because of interference. I need to get the VFD out of the controller cabinet… And I would like to use NC sensors for that to be able to detect faults in the wires as well. Haven’t just found them in reasonable prices though…
Super cool to hear, and hope this series will be inspirational to.. Thanks for the confirmation on the sensors. And if I recall correctly those were about 9usd a piece so that’s quite affordable. My VFD is still inside the control box and had no interfere on sensors ever. Might be a bit lucky, or due to a lot of grounding and shielding.
@@RoboCNCnl I might have to take another look for new sensors. The current ones work with 5volts (if i remember right) and there's definitely interference spikes of that voltage inside the cabinet. I cannot use even the emergency stop buttons... I think it's specifically inside the cabinet because connecting or disconnecting the actual machine wires doesn't make any difference there. They trigger even nothing is connected to the outside connectors... But otherwise the machine has been working great!
Yeah TTL level is something from the past, HTTL or anything like 24V is much preferred if your controller is capable. Signal to noise ratio is something to take into account. Nice video subject 🤔
Supertof project om te volgen! Ik heb zelf een aantal jaar geleden een cnc plasmatafel gebouwd en ik heb ook voor NO in series gekozen :) Ik gebruik twee microswitches die detecteren of de torch goed op de Z as plaat zit, werkt ook goed. Ben benieuwd met welke oplossingen jij verder allemaal gaat komen :) Destijds was jouw cnc machine al een bron van inspiratie :)
Bedankt voor je leuke woorden ! En hopelijk leer ik ook weer van jou ervaring. Ik ga ook voor de ‘touch of’ een micros with gebruiken omdat die veel lichter in te drukken zijn. Maar waarom twee? 🤔
Correct, NO contacts in series, this way, if any switch is activated, or the cable breaks some how then the Estop will activate. NC in parallel , any switch will activated the Estop, but a cable break will not be noticed and in the event of a crash damage may occur..
Just found your channel mate, well done. Im here to learn and help if i can. I need to gear down my motors and was looking at those gearboxes but i like your rack setup too on other cnc... you doing the same gearing for this?...
Thanks for finding RoboCNC :) hope you subscribed. No on this machine I will not use the same system. It will be a more conventional method using some beautiful planetary gearboxes. Which I will cover for in one of the next videos.
normally open in series or normally closed in parallel would both work but I would agree that normally open in series is more robust since if the cable goes it will "crash" the machine. If it were normally closed in parallel, the switches would all be open when everything is okay meaning high would be fine and low would mean crash, if the cable goes it would just stay high and couldn't detect a crash which is a problem. Unless I'm also missing something haha
Well since the comment section mostly disagrees with the uploader, i will count this double ;) And as always there are more ways to Rome and indeed i'm quite sure I'm taking a robust way. But of course it will all be tested with the camera rolling :)
I think you're wrong, the 12x12x1mm dont means 1mm gap, it means the standard target tester to have 2mm is 12x12x1mm (1mm is thickness) The flat surface compared to the other to the right is the distance detection, and fore sur the right one detect more in the side of sensor. The right sensor can't be inside of metal (the head) but the left one yes, in french we say "noyable" when you can fix into metal (flat from the holder) You choose NO, but the sensors always detect the target, ti means they are always closed excepted when a crash is detected, it comes open without target... no ? I don't know how you will fix 3 sensors in series ??? because output are not planned to be in serie, you can't put negative sensor with output of the other sensor (i worked 10years to Contrinex who built sensors like this)
Thanks for sharing… I don’t really understand the gap vs thickness explaining… sorry.. Eventhough, in this case we of course need the one that is able to be placed inside metal (flush mounted). And for the sensing distance I have no real way to be sure since it is even aluminum instead of metal. But a short distance is preferred.
Woohoo! Another series coming from Marcel!
And I think you are right. NO contacts should be in series while NC contacts in parallel.
As far as I know, NC is the standard in the industry and you've explained it right why: even a cable break would trigger the safety function. Which in this case is NO because close proximity is the normal condition.
For example on a KUKA robot the safety features triggering the E-stop funcion are all daisy-chained on one loop.
(An ex business partner left me his failed attempt of building a cnc-plasma, so I decided to finish it and make it work -a couple of days ago, and then out of the blue comes your video... -I have to consider it as a sign.)
Keep up the good work!
Thanks for the nice words! And I agree, what you want is an NC situation, but indeed in this case that is the case when you use NO contacts. Anyway, your sign might be right ! But it probably means you are here to help me :) (your machine is already up and running 💪🏻
@@RoboCNCnl I'm here to learn, to think together with you and to contribute if possible.
All I have left is a large welded steel frame, some linear rails and a 100A plasma cutter.
The project had failed previously because the guy had issues with the high frequency noise he couldn't have figured out how to resolve.
So I'm about to follow your lead on the controller and such. This anti-collision device you've found is a huge leap alone.
@@h4z4rd42 make sure you use double insulated/ shielded 5. Core wire so the inside shielding mesh can be terminated to a earth bus bar.... electrical interference has to be collected by the shielding and driven to ground .... literally.... hypertherm have an article on it. Gantry earthing, driving rod into ground etc... and if you controller is usb controlled from computer, you will be having issues forever as usb has too much sensitivity and interference and isnt built to run robots, for which these machines really are... i went with avid centroid acorn controller ... ethernet connection!
Thanks for sharing your knowledge!
Looking forward to following along with this build, I've been thinking of building a CNC plasma table myself, though probably smaller than you're planning.
Just a note on the sensors, you cannot (or at least should-not) connect transistor style outputs in series - at least not directly. As they are, the only way to connect them in series is to have each sensors ground run to the previous sensors output, resulting in the down-stream sensors becoming unpowered when the upstream sensor breaks the circuit, as they lose their ground connection. Also note that the (low) output at the final stage, will be somewhat offset from the system ground as each transistor stage has a small offset associated with it. This could end up being at a level too high to be properly/reliably detected, especially in a noisy environment. (you also need to ensure that the sum of the load currents posed by each sensor, do not exceed the transistor output capability, though you're probably okay there with only the 3, so 2 are in the load) This is likely why the manual suggests connecting the NC outputs in parallel. In that case if any one triggers, the input will get pulled low. The downside to the parallel connection is that you can't detect cable break faults, as you could with a series connection. If you had relay/switch outputs, then series connection would be correct for the way you are thinking.
Nice to hear you'll follow along ! and please feel free to engage when you have ideas !
Thanks for sharing your thoughts on the sensors in series… I should do some calculations here, although I will not find out the transistor types. I do feel that since I’m not powering them with TTL level but a 24V and no load (or at least very small) that I suspect to only see the voltage drop of 3timea the PN transition.
Anyway… I did not think of it, so it’s a good thing to test, thanks for sharing !
Nice! Your earlier videos were a big inspiration and information source for my build. Thanks for that, I have now many many super precise boat parts done!
And I think you’re right with the sensors. It’s different of course because you’re “constantly sensing” the normal situation and if sensing stops, that an error.
With normal crash detection it would be other way around of course.
I haven’t been able to use sensors at all so far because of interference. I need to get the VFD out of the controller cabinet… And I would like to use NC sensors for that to be able to detect faults in the wires as well. Haven’t just found them in reasonable prices though…
Super cool to hear, and hope this series will be inspirational to..
Thanks for the confirmation on the sensors. And if I recall correctly those were about 9usd a piece so that’s quite affordable.
My VFD is still inside the control box and had no interfere on sensors ever. Might be a bit lucky, or due to a lot of grounding and shielding.
@@RoboCNCnl I might have to take another look for new sensors.
The current ones work with 5volts (if i remember right) and there's definitely interference spikes of that voltage inside the cabinet. I cannot use even the emergency stop buttons...
I think it's specifically inside the cabinet because connecting or disconnecting the actual machine wires doesn't make any difference there. They trigger even nothing is connected to the outside connectors...
But otherwise the machine has been working great!
Yeah TTL level is something from the past, HTTL or anything like 24V is much preferred if your controller is capable. Signal to noise ratio is something to take into account. Nice video subject 🤔
Keep it up Robo.... Looking Foreward!
Thanks mate !! When its ready lets make some sparks together !
Now....wouldnt that be cool!
Oh man hes backk
Thanks for waiting :)
Supertof project om te volgen! Ik heb zelf een aantal jaar geleden een cnc plasmatafel gebouwd en ik heb ook voor NO in series gekozen :) Ik gebruik twee microswitches die detecteren of de torch goed op de Z as plaat zit, werkt ook goed.
Ben benieuwd met welke oplossingen jij verder allemaal gaat komen :) Destijds was jouw cnc machine al een bron van inspiratie :)
Bedankt voor je leuke woorden ! En hopelijk leer ik ook weer van jou ervaring. Ik ga ook voor de ‘touch of’ een micros with gebruiken omdat die veel lichter in te drukken zijn. Maar waarom twee? 🤔
Correct, NO contacts in series, this way, if any switch is activated, or the cable breaks some how then the Estop will activate. NC in parallel , any switch will activated the Estop, but a cable break will not be noticed and in the event of a crash damage may occur..
Thanks for your reply… always nice to use the ‘collective brain’… 🙏🏻
Just found your channel mate, well done. Im here to learn and help if i can. I need to gear down my motors and was looking at those gearboxes but i like your rack setup too on other cnc... you doing the same gearing for this?...
Thanks for finding RoboCNC :) hope you subscribed.
No on this machine I will not use the same system. It will be a more conventional method using some beautiful planetary gearboxes. Which I will cover for in one of the next videos.
normally open in series or normally closed in parallel would both work but I would agree that normally open in series is more robust since if the cable goes it will "crash" the machine. If it were normally closed in parallel, the switches would all be open when everything is okay meaning high would be fine and low would mean crash, if the cable goes it would just stay high and couldn't detect a crash which is a problem. Unless I'm also missing something haha
Well since the comment section mostly disagrees with the uploader, i will count this double ;)
And as always there are more ways to Rome and indeed i'm quite sure I'm taking a robust way. But of course it will all be tested with the camera rolling :)
Yooo Marcel long time has passed;)
Although it was not much... you might have missed some, since the last vid was last month
@@RoboCNCnl i Watch all of ur content no mather the length 😉
Much appreciated!! 🙏🏻
I think you're wrong, the 12x12x1mm dont means 1mm gap, it means the standard target tester to have 2mm is 12x12x1mm (1mm is thickness)
The flat surface compared to the other to the right is the distance detection, and fore sur the right one detect more in the side of sensor. The right sensor can't be inside of metal (the head) but the left one yes, in french we say "noyable" when you can fix into metal (flat from the holder)
You choose NO, but the sensors always detect the target, ti means they are always closed excepted when a crash is detected, it comes open without target... no ?
I don't know how you will fix 3 sensors in series ??? because output are not planned to be in serie, you can't put negative sensor with output of the other sensor
(i worked 10years to Contrinex who built sensors like this)
Thanks for sharing… I don’t really understand the gap vs thickness explaining… sorry..
Eventhough, in this case we of course need the one that is able to be placed inside metal (flush mounted). And for the sensing distance I have no real way to be sure since it is even aluminum instead of metal. But a short distance is preferred.