How to Automate Industrial Welding Positioners with Arduino -- Episode 566

That's an impressive ground system.  It is a huge equipotential ground plane - similar to what goes under high voltage substations.  The hope is that if something goes catastrophically wrong, you are the bird on the wire.  No step potential (delta V between your legs).  Think "dissipating lightning strike" and it really makes sense.

A great grounding system didn't work against you.  Proximity transmitter to receiver worked against you.  At least you didn't have to worry about circulating ground currents.

Yes, you have to use them for every wire going into and out of the computer.

One thing I forgot to mention in the video is how the system was grounded. I'll try to explain.

When this shop was built, the guy that owns it built a steel mesh into the concrete with 3/4" threaded inserts at each intersection. He then tied the grounding for the building into that mesh, as well as tied it into the rebar in the concrete. He uses the threaded inserts to hold down equipment and projects, as well as using them for local grounding points. 

The welder has a separate ground output which we had tied into this system, and the control box was grounded via the electrical receptacle. 

-Kaleb

Hi DAB,

Sorry, I replied to this yesterday but it didn't save...

I did put ferrite beads on the multi-conductor wire from the pendant, but not on each individual conductor. Is that what you mean?

-Kaleb

Hi Kaleb,

Did you try ferrite beads on the Arduino inputs?

When I worked on putting devices on aircraft, we had a similar EMI issue, and the ferrite beads did wonders at attenuating the pulses.

The guy that I was working on this for was quite proud of his reused festool box for an enclosure and wanted to use that, haha... In place of that, he had the enclosure lined with a thick foil, which was grounded. The cables were shielded and grounded inside the enclosure. Maybe the steel box would have made the difference. However, after I removed all of the inputs from the control pendant, the arduino stopped resetting or freaking out when the welder fired.

I forgot to talk about the grounding in this building in the video. The guy that owns the shop built in some interesting grounding when he built it. All outlets and breaker boxes were tied to a common ground that was sunk into earth per normal spec, but then he tied that into a system of steel anchors he built into the concrete floor. It was a grid of steel something like 1m x 1m, with 1" threaded inserts at each intersection for him to bolt things down onto the floor. The ground was also tied into the rebar in the concrete. He claimed that his acts as a huge ground plane, as well as giving many points to tie in an extra ground. 

The welder was running 3 phase 480, and had a separate ground that was tied directly to one of those threaded inserts. The control box ground was just tied in through the 110v receptacle. 

Maybe this helped, or maybe it made it worse. Anybody have any ideas on that?

Thanks,
-Kaleb

KalebTheMaker  If you opt for a 12 or 24 volt relay systems, you should be sufficiently isolated from any induced EMF.  You've already proven that the 5V threshold still picks it up.  For my (few) industrial Arduino applications, I've standardized on 12VDC and have a dedicated buck converter for the Arduino.  The relay modules are all 12V.  Fly by wire, right?  Otherwise, Andy's suggestion for CAN should also be a solid solution.  You're already using the ESP32.

I think CAN would work.  It functions near vehicle ignition systems.  Those can be pretty noisy.  My first thought was MODBUS as the RS-485 also uses a differential pair.  You can also get shielded, twisted-shielded wire for high noise environments.  If it's a longer run, CAN is probably the best solution. 

For the interference from the welder refer to the operators manual for radio and tv or electronic interference or the welder manufacturer. The arduino should be in a steel enclosure with a solid ground connection, and all cables going in and out must be sheilded and the sheild must be grounded only on the end in the control box.

I worked in an electric motor shop where we built control boxes for industrial customers. We used PLCs but the interference problems are the same. I also went to welder repair school at the Lincoln Electric factory. 

An interesting project. I will follow him.