Best network communication protocol for an extreme EMI environment

[Fred Derf]

I realize this might be like asking what's better, Ford, Chevy, or Dodge, and the opinions are likely as varied. I ask here specifically because I am designing a new multi-axis welding system  for my current employer. We build automated welding equipment for environments that are extremely prone to EMI noise interference. Besides relay logic and proprietary circuit boards, I've used the Samba and V700 to control our equipment. Those products have worked extremely well for us. Control has been done via analog out for speed and relay logic for Enable/Direction to simple SCR drives powering 0-90VCD gearmotors. I'm old enough to have witnessed the transition from tubes to transistors and have spent a lifetime mitigating EMI in our control designs. Part of that design mentality was to never design around a network centric control system. A good portion of our business has been replacing such controls that just could not reliably operate in such a noisy environment. Now I must design a multi-axis servo controlled network system. The welding systems are often portable and have to operate from regular North American 120VAC outlets. That unfortunately prevents me from trying the Unitronics line of servos and their Unican. I don't need to have fast communication with the drives, I need reliable communication. Axis travel speeds are measured in Inches/Minute. So considerable latency will be acceptable. The V700 seems very versatile and can use several protocols. Ethernet TCP/IP, RS485/RS232, MODBUS TCP, SNMP*, CANopen, CANlayer2, UniCAN, BACnet, KNX and M-Bus via gateway.

Initial servo drives are likely to be from Applied Motion Products. Unless someone has another suggestion. Their drives are available for many different protocols.

Given the current design parameters, what is the best network communication protocol for this type of an application?

[Flex727]

I think this is a great question and I hope you get an answer that is just as great from some of our members with the right experience. <following>

[kratmel]

I deal with ABB, Fanuс, Panasonic welding robots with different MIG, TIG process and Migatronic welding equipment.

Everything that really works and moves in these systems is basically galvanically isolated CAN (CANopen) networks.

In my opinion, CANopen -  the most common and standard network for motion control supported by the Vision PLC. By using additional isolated power supply for the CAN transmitter and proper shielding of the networks, it is possible to achieve transmission stability.

The main thing is to ensure proper grounding of all components of the welding system. However, I think for the author of the topic this should not be a problem, thanks to his many years of practical experience.

The only problem that the Vision PLC probably can't solve is the simultaneous movement of the axles like a CNC machine. This requires a third-party motion controller.

Therefore, the question to the author of the topic - will the movements of the axes be consecutive along the main axes of coordinates or should it be interpolation?

 

[Fred Derf]

Thank you Kratmel for the input. The ubiquitous CAN bus has survived for quite some time for a reason. With my clean slate approach I thought I’d ask those who have experience with the Unitronics products. 

 

As for grounding, I’m often in a situation where there is no real Earth or line ground. These welding applications are Hard Surfacing or Weld Rebuild. Portable trucks with engine driven welding power supplies that do not offer any practical ground. Undesirable energy must be dissipated internally. 

 

It is not a fabrication application there is no need for high accuracy or interpolated movements. Its mostly bang bang servo control with some dynamic speed variations. The goal of these welding applications is to rebuild with molten metal as fast as possible. Sometimes with six torches simultaneously welding in excess of 150kW total. 

 

The V700 is a great inexpensive controller for these simple tasks. I do it now via analog and relay signals. I just need to do the same via network communication. 

 

[kratmel]

4 hours ago, Fred Derf said:

Portable trucks with engine driven welding power supplies that do not offer any practical ground.

In this case, the main thing is that the welding currents do not pass through the shielding surfaces of the signal wires. The Faraday cage (control cabinet), extended by shielded cables to the motors and encoders, must work. You say no Unitronics servo possible. The use of a transformer to increase the voltage from 120 to 220V further isolates your system from parasitic interference. Accordingly, you will have the advantage - the availability of ready-made ladder  for motion control through the CAN.

The only thing you should pay attention to is the resistance of cables to UV radiation. Proper cables have been working on welding robots servo for over 10 years. Those closest to the UV source degrade quickly.

In your case, I do not know whether UV rays will fall on the conductors to the servomotors and how it will react to this.

[Ausman]

Kratmel has done a great job here...good stuff and well done.  Something you might like to also consider, but may not be very practical in the situation you're describing, could be optical interfacing.  I've never used such stuff, but it is available.  One quick link found:

https://www.prosoft-technology.com/Products/Gateways/CanBus/Fiber-Optic-CAN-bus-Extenders

On the UV issue, would it also not relate to the amount of movement being done?  One could possibly use flexible metal conduit if there is not complex repetitive movement, and it sounds like your situation might not be this.....just lots of back and forwards passes with slight height adjustment each time. 

cheers, Aus

[Ausman]

It has also just occurred to me to ask why you need to use 120V?  In some ways it implies that you're plugging into mains to run your control gear. If you're running the trucks, can't you use power derived from them?  Or are you doing this and use 120V stuff as it is easily available to you?  (You Yankees who need far bigger cables because you use 1/2 the volts us sensible (?) people do!)  🙂

cheers, Aus

[Flex727]

21 hours ago, Ausman said:

(You Yankees who need far bigger cables because you use 1/2 the volts us sensible (?) people do!)  🙂

On a trip to Northern Ireland a few years ago I was using one of those portable immersion heaters and was absolutely shocked at how fast it boiled the water. I had never used one before and set about getting one when I returned home. Yep, here in the U.S. they are near useless.

[Ausman]

I know, I know, off topic, but all that copper you guys are hogging!  In theory you use twice as much to do the same job.  In a place like the USA that adds up to a huge amount.  Or perhaps you don't have as much "safety" capacity on your lines and routinely run much the same cross sections as us?  What are your current ratings for 2.5mm² and 4.0mm² say?