Follow RealPars on Instagram instagram.com/realparsdotcom/ for daily practical tips and tricks on Industrial automation and electrical engineering topics.
Hi Nell, Thanks a lot for your kind compliment! We are always extremely happy to hear such positive feedback! If you ever have any questions, feel free to reach out to us. Happy learning!
Modern Ethernet switches have port to port latency of single digit nanoseconds. Arista has one with less than 4ns of latency. What advantage does EtherCAT have over that?
Thank you for your question! EtherCAT processes signal frames through nodes, so no Ethernet switches are needed. It operates in real-time to minimize delays, and you can adjust the bandwidth if necessary. For more detailed benefits, I recommend checking out the EtherCAT website. Happy learning!
@@realpars I mean it seems as if EtherCAT was created because Ethernet switches were considered to be too slow for real-time process control but that is no longer true. Modern Ethernet switches are insanely fast. I suppose EtherCAT does save the cost of the switch but it adds cost to every device. Latency also grows with the number of devices on the bus while latency is constant with a switch. It may be easier to wire devices to each other than to a central switch.
Having seen this video, I think a video comparing “EtherCat vs ControlNet Network” advantage and comparison should be aded ti the RealPars Team queue for future videos. This arise too from a internal conversation at my OEM Engineering Company as to why we still design brand new plants in 2019 with ControlNet vs EthernetIP. On my side of company we do all plant upgrades as that what we do in all EthernetIP. Main side of Company on greenfield new projects still do ControlNet Networks. Reason was attributed to slower speed and not missing any communication data as compared to EthernetIP, but it made no sense to me logically. So I’d love a video on that. Or also include EthernetIP in the 3 way comparison.
We will take your suggestion into consideration. ControlNet is deterministic. It was created using standard 10Mb Ethernet and using 5Mb for data transfer and 5Mb for insuring that data packets sent to a node are actually received by a node. It adds the ACK and NACK features that standard Ethernet is lacking. Standard Ethernet is like the post office. You send a letter (Ethernet packet) and you do not know if it has been received until you "ask". ControlNet adds a function that is used to guarantee that the packet is received, making it deterministic. Ethernet/IP is not deterministic, although it is very robust. A couple of layers have been removed from the OSI stack to make it is as reliable as possible and as deterministic as possible, without imposing the overhead and limitations of ControlNet.
Scott Sommer 🤔... That is such a valuable and well articulated reply that I actually get that in your explanation. Though I’ve had some very smart engineers at work try to explain things to me before with the ControlNet issue. None ever gave me a simpler explanation than that where it’s meaningfully understood. I’ll surely screenshot this response for future reference because I can’t win with so call smarter thinking engineers. Still, i think a video would be great in comparing all three and explaining the issues or what you so well explained and more. Either way, glad it’s being considered. But I’m most thankful for such an easy well explained explanation that didn’t leave me lost at all still wondering. Well Explained! Reply so well appreciated.
Not a fan of either of them. I would go with EtherCAT. Low cost, phenomenal performance and easy to implement. People use EIP because Rockwell has been around for a long time. Many suppliers chose EIP due to its market share not because of its performance. EtherCAT is open technology with more members joining every year.
Thank you for the video it's very informative but I' am little bit confused about the timestamps. So are there two timestamps, one when the node received the message and one when the frame leaves the node?
there are none. Literally the easiest coms protocol. To the best bit is that it's now available wirelessly. No need for connecting anything to your PC.
Well, one disadvantage is if you need to make an unplanned topology change on a live specimen. That is quite straightforward with ethernet IO, but with ethercat it's a very theoretical possibility, practically you just can't, you need to take it offline. Which is a problem in some cases. Maybe some time in the future there will be better software support for it and it will be somewhat doable in limited fashion, but due to system architecture I don't ever see it being as simple as with ethernet IO. Also, if your controller is not realtime and you can't put out the packet on time, that will cause many ethercat slaves to misbehave and just plain not work properly. Which is fine if you actually have realtime controller anyway, but that is not the case in every sort of system. So if you need to make your control system realtime just so you could use ethercat, then that's a lot of extra headache and cost for something you just don't need. Ethercat does need a very sizeable software and hardware stack, which can be an issue in embedded systems, serial comms in comparison are trivially simple and much cheaper. For that reason I don't think you are ever going to find ethercat in your coffee machine. Ethercat is absolutely awesome, but to say it's a one size fits all solution, is perhaps an excessively narrow view. There absolutely are cases where ethercat is not the best option.
I don't understand how it works in a tree/star network if it's supposed to go around in a fixed time. Wish there was a way to get to grips/tinker with PLCs outside of a work environment
Hello Kestas Kuliukas. Thank you for your question. I can understand it could be confusing. In our short video, we just talked about the ring topology in a simple understandable format, and we did not discuss in detail all the other possibilities. But to help with your question and learning I am including a link to the EtherCAT org group where you can learn more about this network. Happy learning! www.ethercat.org/en/technology.html
No. EtherCat enabled devices will have two RJ45 connections; one for the incoming network (from the previous node) and one for the outgoing (to the next node). Each EtherCAT connection act as a switch and can pass the data packets on to the next node. You can't just add EtherCAT to your existing network; you need to install EtherCat-enabled processors. You can add an Ethernet switch port to an EtherCat processor in order to interface to Ethernet peripherals, such as printers.
about Distributed clock : what did you mean for " Each node adds the timestamp as the message is received" for timestamp you mean that the slave adds to the frame its internal current time respect the GMT?
Normally 1st slave device is your master clock, when EtherCAT frame reaches each subsequent slave node the time stamp is saved. EtherCAT master uses timestamps to calculate delays between each node. This will help you synchronize nodes in the same sync group. This has nothing to do with GMT.
Thank you for your question! No, EtherCAT does not use IPv4 addresses. EtherCAT uses a "processing-on-the-fly" method, eliminating the need for IP routing. Hope this helps :)
EtherCAT was designed for implicit messaging in a ring topology, giving the user the fastest speed and reliability. Three-p[ort EtherCAT devices allow a "spoke" or "drop" to be attached to the device. In order to employ this configuration, explicit messaging on that port is required. The same goes for the star configuration. Essentially, EtherCAT is "demoted" to CAN over Ethernet for these configurations, which is not as fast or efficient.
@@realpars Star topologies are possible where the EtherCAT frame follows down the "spoke" and returns back along the same physical connection before continuing on, effectively making a star network appear like a line/ring network if you were to flatten it out. See for example CU1123 or CU1128 from Beckhoff. Internally, every EtherCAT slave uses a special ASIC such as the ET1100, some of which support 3 or 4 physical ports. There is no performance penalty to this architecture beyond the delay introduced by the additional nodes in the branch.
How many motors can 1 PLC control? I have 5 (2 axis drilling machines each has 3 motors) which is 15 motors total, and I need to know how many PLCs do I need to control them at the same time.
Hi Bassel, As far as I know, the PLCs are not used in controlling the CNC machines. I am not sure if you are aware of Siemens SINUMERIK CNC automation systems or not but you can find out more about them here: sie.ag/2u87gPL There are other manufacturers of CNC automation systems which you can find them with a simple search. However, you can use PLCs to control your machines, but I'm not sure that PLCs can function as good as the CNC control systems and in addition, the code would become complicated. Probably a Siemens 1200 CPU would be enough if you insist on PLCs. Thanks
Bassel, this could easily be done with 1 Beckhoff PLC. TwinCAT NCI supports 31 channels, each channel can have 3 path axis + 5 auxiliary axis, so 248 axis in total is the software limit. You would need a controller with enough calculation power, but 15 motors is like a basic application for TwinCAT. Something like a CX5140-0155 would do that.
Both EtherCAT and Profinet are used for the same purpose of industrial Ethernet. EtherCAT is developed by Beckhoff and Profinet is developed by Siemens.
Hello Sir I have a question, my machine gives an ethercat error showing different cards every time and goes away for some time after powering on and off, can you guide what could be the problem, I have checked all the connections and their tightness, I have also checked the topology but it seems that doesn't matter in this case because it points out a different module every time
Use TwinCAT’s EtherCAT diagnostic feature and check again. Are the slaves getting enough power? Also check your Ebus power consumption. It should be fairly easy to diagnose.
No reason other than the redundancy capability, flexible network topology, speed, ease of use, distributed clocks functionality, advanced diagnostics of each slave device, no feasible limit on devices, and the ability to use any other fieldbus in a master or slave configuration via one single module over EtherCAT. Yeah, other than all of those there’s barely any benefits
![KAYTRANADA - Witchy (feat. Childish Gambino) [Official Video]](http://i.ytimg.com/vi/jGK3YVmGZ3Y/mqdefault.jpg)
Follow RealPars on Instagram instagram.com/realparsdotcom/ for daily practical tips and tricks on Industrial automation and electrical engineering topics.
Is RealPars also in LinkedIn? I would like to follow...
The way every device timestamps the frame when it enters and leaves is brilliant.
Visualization is the edge of this channel. Each topic is a nugget of knowledge. Engaging explanations in a very short time videos.
Hi Nell,
Thanks a lot for your kind compliment! We are always extremely happy to hear such positive feedback!
If you ever have any questions, feel free to reach out to us.
Happy learning!
using RealPars videos and content on my current INDUSTRIAL COMMUNICATION PROTOCOLS.
Thanks RealPars
You're very welcome! Happy learning.
Thanks for the sharing and neutral review on EtherCAT , totally love it !
Great to hear that, Daniel! Thanks for sharing.
Great video again! Please share more Beckhoff tutorial with us!
Thanks for the suggestion! I will definitely pass that on to our creator team. Happy learning!
What do you want to learn about Beckhoff?
Another piece of automation knowledge brought to me by www.realpars.com .No CATs were hurt in this commercial.
Renegade Flower not funny.... at all
Thank you! :)
Thanks, RealPars, very informative!
Thanks a lot, Bob! :)
Please upload more videos about EtherCAT
Thanks, Shashikant! I will let our creator team know. Happy learning!
Incredible explanation, thanks!!!
Glad it helped! Happy learning
Modern Ethernet switches have port to port latency of single digit nanoseconds. Arista has one with less than 4ns of latency. What advantage does EtherCAT have over that?
Thank you for your question! EtherCAT processes signal frames through nodes, so no Ethernet switches are needed. It operates in real-time to minimize delays, and you can adjust the bandwidth if necessary.
For more detailed benefits, I recommend checking out the EtherCAT website. Happy learning!
@@realpars I mean it seems as if EtherCAT was created because Ethernet switches were considered to be too slow for real-time process control but that is no longer true. Modern Ethernet switches are insanely fast. I suppose EtherCAT does save the cost of the switch but it adds cost to every device. Latency also grows with the number of devices on the bus while latency is constant with a switch. It may be easier to wire devices to each other than to a central switch.
Having seen this video, I think a video comparing “EtherCat vs ControlNet Network” advantage and comparison should be aded ti the RealPars Team queue for future videos.
This arise too from a internal conversation at my OEM Engineering Company as to why we still design brand new plants in 2019 with ControlNet vs EthernetIP. On my side of company we do all plant upgrades as that what we do in all EthernetIP. Main side of Company on greenfield new projects still do ControlNet Networks. Reason was attributed to slower speed and not missing any communication data as compared to EthernetIP, but it made no sense to me logically. So I’d love a video on that. Or also include EthernetIP in the 3 way comparison.
We will take your suggestion into consideration. ControlNet is deterministic. It was created using standard 10Mb Ethernet and using 5Mb for data transfer and 5Mb for insuring that data packets sent to a node are actually received by a node. It adds the ACK and NACK features that standard Ethernet is lacking. Standard Ethernet is like the post office. You send a letter (Ethernet packet) and you do not know if it has been received until you "ask". ControlNet adds a function that is used to guarantee that the packet is received, making it deterministic. Ethernet/IP is not deterministic, although it is very robust. A couple of layers have been removed from the OSI stack to make it is as reliable as possible and as deterministic as possible, without imposing the overhead and limitations of ControlNet.
Scott Sommer 🤔... That is such a valuable and well articulated reply that I actually get that in your explanation. Though I’ve had some very smart engineers at work try to explain things to me before with the ControlNet issue. None ever gave me a simpler explanation than that where it’s meaningfully understood. I’ll surely screenshot this response for future reference because I can’t win with so call smarter thinking engineers. Still, i think a video would be great in comparing all three and explaining the issues or what you so well explained and more. Either way, glad it’s being considered. But I’m most thankful for such an easy well explained explanation that didn’t leave me lost at all still wondering. Well Explained! Reply so well appreciated.
Not a fan of either of them. I would go with EtherCAT. Low cost, phenomenal performance and easy to implement. People use EIP because Rockwell has been around for a long time. Many suppliers chose EIP due to its market share not because of its performance. EtherCAT is open technology with more members joining every year.
Thank you for the video it's very informative but I' am little bit confused about the timestamps. So are there two timestamps, one when the node received the message and one when the frame leaves the node?
Yes, this is required by the distributed clock system of EtherCAT.
Great explanation, thanks!
Thank you!
Interesting, but it would be great if you could also highlight some disadvantages.
Thanks for your feedback, will happily pass this on to our course developers!
Happy learning!
there are none. Literally the easiest coms protocol. To the best bit is that it's now available wirelessly. No need for connecting anything to your PC.
I can hardly think of any to be honest.
Well, one disadvantage is if you need to make an unplanned topology change on a live specimen. That is quite straightforward with ethernet IO, but with ethercat it's a very theoretical possibility, practically you just can't, you need to take it offline. Which is a problem in some cases. Maybe some time in the future there will be better software support for it and it will be somewhat doable in limited fashion, but due to system architecture I don't ever see it being as simple as with ethernet IO.
Also, if your controller is not realtime and you can't put out the packet on time, that will cause many ethercat slaves to misbehave and just plain not work properly. Which is fine if you actually have realtime controller anyway, but that is not the case in every sort of system. So if you need to make your control system realtime just so you could use ethercat, then that's a lot of extra headache and cost for something you just don't need.
Ethercat does need a very sizeable software and hardware stack, which can be an issue in embedded systems, serial comms in comparison are trivially simple and much cheaper. For that reason I don't think you are ever going to find ethercat in your coffee machine.
Ethercat is absolutely awesome, but to say it's a one size fits all solution, is perhaps an excessively narrow view. There absolutely are cases where ethercat is not the best option.
Thank you!!
Good explanation...Can you explain about I/O s
Thanks for the topic suggestion, I will definitely go ahead and forward this to our creator team. Happy learning!
Can you add the video to the Industrial-network playlist?
I don't understand how it works in a tree/star network if it's supposed to go around in a fixed time. Wish there was a way to get to grips/tinker with PLCs outside of a work environment
Hello Kestas Kuliukas.
Thank you for your question. I can understand it could be confusing. In our short video, we just talked about the ring topology in a simple understandable format, and we did not discuss in detail all the other possibilities. But to help with your question and learning I am including a link to the EtherCAT org group where you can learn more about this network.
Happy learning!
www.ethercat.org/en/technology.html
Fantastic video!
Thanks a lot, Oliver!
Dear sir please make video on indirect addressing in S7-300 plc..
Hi Vikram, thanks for the topic suggestion! I will pass it on to our creator team.
Thanks!
You're very welcome!
Thanks for the video. Quick question. Our PLC only has 1 ethernet port. If using EtherCat, can I use an ethernet switch in the loop? Thanks Again.
No. EtherCat enabled devices will have two RJ45 connections; one for the incoming network (from the previous node) and one for the outgoing (to the next node). Each EtherCAT connection act as a switch and can pass the data packets on to the next node. You can't just add EtherCAT to your existing network; you need to install EtherCat-enabled processors. You can add an Ethernet switch port to an EtherCat processor in order to interface to Ethernet peripherals, such as printers.
Interesting, never heard of this before.
That is a good sign that our course videos are helpful! :)
Happy learning!
Hey can u plz do a video on how braking modules works and how it stop the motor
Thanks for the topic suggestion, Richard! I will pass it on to our creator team. Happy learning!
@@realpars am learning alot from u guys thanks a Lot
about Distributed clock : what did you mean for " Each node adds the timestamp as the message is received" for timestamp you mean that the slave adds to the frame its internal current time respect the GMT?
Normally 1st slave device is your master clock, when EtherCAT frame reaches each subsequent slave node the time stamp is saved. EtherCAT master uses timestamps to calculate delays between each node. This will help you synchronize nodes in the same sync group. This has nothing to do with GMT.
Does ethercat have ipv4 address like profinet? Can it transmit over an ethernet network with routers and switches like profinet?
Thank you for your question! No, EtherCAT does not use IPv4 addresses. EtherCAT uses a "processing-on-the-fly" method, eliminating the need for IP routing.
Hope this helps :)
There exists 3-Port EtherCAT Slaves. I'm wondering how the frames moves in that topology? Better yet, how does the frame moves in a star topology?
EtherCAT was designed for implicit messaging in a ring topology, giving the user the fastest speed and reliability. Three-p[ort EtherCAT devices allow a "spoke" or "drop" to be attached to the device. In order to employ this configuration, explicit messaging on that port is required. The same goes for the star configuration. Essentially, EtherCAT is "demoted" to CAN over Ethernet for these configurations, which is not as fast or efficient.
@@realpars Star topologies are possible where the EtherCAT frame follows down the "spoke" and returns back along the same physical connection before continuing on, effectively making a star network appear like a line/ring network if you were to flatten it out. See for example CU1123 or CU1128 from Beckhoff. Internally, every EtherCAT slave uses a special ASIC such as the ET1100, some of which support 3 or 4 physical ports. There is no performance penalty to this architecture beyond the delay introduced by the additional nodes in the branch.
Amazing!
Thank you, Rajat!
How many motors can 1 PLC control?
I have 5 (2 axis drilling machines each has 3 motors) which is 15 motors total, and I need to know how many PLCs do I need to control them at the same time.
Hi Bassel,
As far as I know, the PLCs are not used in controlling the CNC machines.
I am not sure if you are aware of Siemens SINUMERIK CNC automation systems or not but you can find out more about them here: sie.ag/2u87gPL
There are other manufacturers of CNC automation systems which you can find them with a simple search.
However, you can use PLCs to control your machines, but I'm not sure that PLCs can function as good as the CNC control systems and in addition, the code would become complicated. Probably a Siemens 1200 CPU would be enough if you insist on PLCs.
Thanks
You can use 32 DI/16 DO PLC
@@realpars Thank you so much.
@@piyushviroja4648 Thanks a lot.
Bassel, this could easily be done with 1 Beckhoff PLC. TwinCAT NCI supports 31 channels, each channel can have 3 path axis + 5 auxiliary axis, so 248 axis in total is the software limit. You would need a controller with enough calculation power, but 15 motors is like a basic application for TwinCAT. Something like a CX5140-0155 would do that.
Can you do one for CAN/CANopen?
Sure, I will happily pass this on to our course developers. Thanks for sharing your topic suggestion!
is EtherCAT use for the same purposes as Profinet?? or what would be the difference?
Both EtherCAT and Profinet are used for the same purpose of industrial Ethernet. EtherCAT is developed by Beckhoff and Profinet is developed by Siemens.
Thanks
You are very welcome!
nice bro！
Nice bro. Pc links
Thank you, Asfand!
thanks.......
You are welcome!
zes,,,,,,,,,,,,,,,,,thaks ........
Hello Sir I have a question, my machine gives an ethercat error showing different cards every time and goes away for some time after powering on and off, can you guide what could be the problem, I have checked all the connections and their tightness, I have also checked the topology but it seems that doesn't matter in this case because it points out a different module every time
Use TwinCAT’s EtherCAT diagnostic feature and check again. Are the slaves getting enough power? Also check your Ebus power consumption. It should be fairly easy to diagnose.
Thanks ✌
You are very welcome!
Why would anyone choose EtherCAT over EtherNet/IP?
No reason other than the redundancy capability, flexible network topology, speed, ease of use, distributed clocks functionality, advanced diagnostics of each slave device, no feasible limit on devices, and the ability to use any other fieldbus in a master or slave configuration via one single module over EtherCAT. Yeah, other than all of those there’s barely any benefits
@@InvisibleSquirrelz Thanks for the comprehensive reply
Because EtherNet/IP sucks. It’s slow, lots of effort to configure, expensive (have to invest in routers switches).
By Christ you sold that well 😅
big lezion