Skip to main content
12 events
when toggle format what by license comment
Feb 9 at 14:59 comment added MrGerber On the conclusion, we can agree.
Feb 9 at 7:43 comment added Lundin @MrGerber Around year 2007 I was involved in a project developing a Profibus DP product. Back then, Profibus DP had maybe 40-50% of this whole pie chart and Modbus something like 10% perhaps. The Profibus DP product sold nicely when it was new, but nowadays we sell a trinkle of a few units per year. These are rapidly dying technologies and good riddance. The only reason it stayed alive well past the best-before date is becaues Siemens were always pushing for Profibus. (Which in turn is the only reason for Profinet at all existing as well.) We only need two buses: CAN based and Ethernet based.
Feb 8 at 18:13 comment added MrGerber 2/* So yes, Industrial Ethernet is for sure taking over, and that will most likely keep gaining ground as APL gets more widespread, finally you can do proper, modern network over existing field cables. But Modbus will not be obosolete for a long while yet.
Feb 8 at 18:10 comment added MrGerber Well. In industrial equipment modbus is for sure quite widespread. It doesn't matter if an engineer deems something to be outdated if that is what the industry wants. And the industry definitely still wants Modbus. As they want HART over 4-20mA - which is even older. According to Automationworld.com - modbus RTU was the next largest fieldbus, after Profibus DP. Modbus TCP is the fourth largest Industrial Ethernet standard. EDIT: link: automationworld.com/communication/article/33009016/…
Feb 8 at 7:21 comment added Lundin @MrGerber UART based technologies in general are outdated. Yeah there are car enthusiasts buying cars from the 1960s, there are retrocomputing fans who would love to get their hands on a 1960s computer and many enjoy music from the 1960s. But that doesn't mean that you as an engineer should specify in 1960s technology to be used for your next supposedly cutting edge project in the year 2024.
Feb 7 at 18:08 comment added MrGerber I would have to say that the statement about Modbus (EDIT: I considered "original" to be RTU, but you might have referred to Modbus ASCII?) being phased out might be dependent on the field you work in. New equipment keep coming out with modbus (ed. RTU) support "every day". Not that it changes the fact that it's unlikely to be the correct choice of protocol for motion, though!
Feb 2 at 10:38 vote accept 19172281
Feb 1 at 14:45 comment added Lundin @19172281 In that case you'll most definitely want to use a sync mechanism as mentioned in this answer, to ensure that each actuator is moving at the same time as the rest of them.
Feb 1 at 13:09 comment added 19172281 I'm controlling 6 servo linear actuators that drive a Stewart platform. Smooth motion (high setpoint generation frequency) is essential which is why is why I wanted to get an understanding of the limitations of Modbus, CAN, etc. I imagine 50hz would be sufficient.
Feb 1 at 12:56 comment added Lundin @19172281 Timing considerations mostly depends on how many nodes that send on the bus simultaneously. Once per 10ms or once per 100ms is standard. You could go faster still, but then you have to do the homework and calculate how long it takes to send the frame itself. If sending at 1m periods then the frame transmission time in itself needs to be taken in account. As for what PLC modules there are, I have no idea since I only work with microcontrollers. PLC in themselves tend to be sluggish. What kind of object are you controlling?
Feb 1 at 12:51 comment added 19172281 I suspected that may be the case. How would CAN fare if we were to send the position, say, 1000/sec? Are there any PLC motion modules that implement motion control in this way (i.e. generate the setpoint on the PLC side, and transmit to drive)?
Feb 1 at 12:36 history answered Lundin CC BY-SA 4.0