TIG Welder crashes my arduino.

Other Hardware General Electronics
1

I made a linear slide for welding in straight lines with the TIG welder. The slide is controlled by an arduino uno, it has an optical encoder at the end of the screw, an LCD display for displaying movement parameters (distance, speed etc) and the circuit has a 5 amp relay that starts and stops the welding process.

Everything went fine when I tested it home. At the beginning of the program the relay clicks, then the slide starts moving at the set speed, for the set distance, then the slide stops and relay disengages. Perfect.

When I tested it in the workshop, with the TIG welding machine attached, the second the welding machine creates the electric arc, the LCD on my control box gets screwed up, and the arduino stops running the code. Nothing gets fried, I only have to reset it. But...bottom line is...the welding machine makes it crash.

I tried plugging the control box in another socket in another room in the house and pushed the weld button on the welding machine in the other room. The lcd and arduino don't get messed up.

So...what could it be? The circuit is powered via an industrial power supply:

and it can keep operating even if I unplug the power supply and plug it back 1/2 seconds later, so the caps inside it are really enough to keep the circuit working even if the power completely drops for a few miliseconds, while the TIG welder starts.

So...what could it be?

It's worth mentioning that there are quite a few wires going back and forth from the encoder to the arduino box, from the box to the welding machine (the relay wire) , and a couple of wires going from the arduino box to the ends of the slide for detecting when the slider has reached the end of the run.

I'm thinking maybe all these wires act as an antenna and catch some of the electromagnetic pulse generated by the welding machine when it starts, and that causes the screen and arduino to get "confused" :slight_smile:

But I really don't know if that's the case, and I don't know where to start with toubleshooting the problem.

Any ideas are very appreciated!

2

An sketch showing connections (and phisical layout: distances, etcetera) would help (to help).

Regards

3

I don't have photos right now, and I'll go to the workshop in the weekend. I was hoping to get some ideas before I get there.

4

But from memory, here are the wire lengths:

Arduino Control box - Optical encoder - 1mt
Arduino Control box - End switch - 1mt
Arduino Control box - Slider motor - 1mt
Arduino Control box relay - TIG Welder - 2mt

The control box cannot be more than 2 mts away from the welding machine as it's currently configured now.

5

It is probable (i'd say almost sure) that, as you guess, interferences are ruining the project (for the moment; sure we can fix it).

Shielding is a first measure to take: just by installing shielded cables (without connecting the shield anywhere -for the moment) in the signal (encoder, motor; in general the lesser the current a cable drives more sensitive it is to interferences) cables should improve (eventually solve) the problems.

The welding machine has enormous (for the problem we try to fix) currents: they generate a big magnetic field (carry a compass with you to verify it) that certainly interfere the other currents in the neighbour, the small current ones being the most affected ones. The shield diminishes the interference a lot.

Try it

Regards

6

Good thinking, the electromagnetic waves smash the Arduino.

Normally, for indoor use, a circuit with an Arduino should not have a higher impedance than 10k. That is a safe value. The Arduino pins are not protected and go straight to the chip, but in most cases it will work.

In a car, it is a noisy environment, and special care should be taken for voltage spikes and so.

That arc (and the current through the wires and transformer) is even worse. The circuit should have a lower impedance at every point, and you should use a lot of shielding. Use shielded cables and perhaps a shield around the Arduino. That is not a gurantee it will work, perhaps the Arduino will be okay, and the cheap display might go nuts with that arc. I think I remember that a cheap LCD display could go nuts with a strong RF signal.

(while I was writing this, vffgaston also mentioned the shielding and currents)

7

Try setting all your unused digital pins to INPUT_PULLUP. In fact even the ones you are using should probably have either have pullup or pull down resistors on them.

8

If your TIG welder is a decent one then it'll use an HF (high frequency) start circuit which generates a high frequency/high voltage to ionise the gap air and hence initiate the weld current flow. This output of energy covers a very wide frequency spectrum so in all probability it is the HF transmitted power that is killing your system.

Arduino inside a metal enclosure (diecast zinc box) and all signal leads shielded and "earthed" to the enclosure MIGHT resolve the problem.

The enclosure itself should then be connected to the same earth point as the welding machine.

9

Thanks for the suggestions guys, I'll have to think of a way to shield the cables and maybe ground the shield somehow.

The TIG welder is a pretty cheap model, and I'm not sure what frequency it operates on. Fact is...this project proved to be more work than I expected, but then again, all the projects are like that for me :slight_smile:

I'll do some work on it over the weekend and will come back and share the results. Thanks a lot!

10

capcaunu:
The TIG welder is a pretty cheap model

I wasn't aware that any such thing as a "cheap" tig welder existed. :money_mouth_face:

11

Hi, have you earthed the negative side of the DC output of the power supply.
Most of those SMPS leave the outputs floating, so at the moment you either have the arduino floating or the negative of the supply earthed somewhere that interference is prevalent.
And use a decent bit of wire to earth, the same gauge as the mains supply earth wire that is in the power lead.

Tom.... :slight_smile:

12

A welder is about the worst source of interference imaginable - if you can keep
all the sensitive electronics well away from it, and all cables between the sensitive
stuff and the welder should have careful RFI shielding and protection circuitry - then
that's a start. And don't run welder high-current cables alongside other cabling, separate
as far as possible.

If you have to have sensors near the welder / slide you will need excellent shielding
(quality shielded twisted pair(s) or coax) and probably opto-isolators too.

13

Shielded twisted pair wiring for everything, coax is not good enough. As has been said, metal box for the electronics and ground it. Single point grounds.

The shielded wire to the sensors should -not- be grounded anywhere but back at the Arduino.

14

I wasn't aware that any such thing as a "cheap" tig welder existed.

You can make "scratch start" machines fairly cheaply (but no aluminum welding). At work Faraday cages are built around the welders when welding in sensitive areas. The arduino may need to be put in a grounded enclosure similar to the power supply pictured.

15

The same thing happens to my lcd when I use my corded electric drill.
The only way I can think of to isolate your arduino is run it on batteries. Any connection to the GRID will
allow the interference (which isn't electromagnetic waves as much as it is spikes) . If you don't have a
local power factor correction box you won't be able to avoid it unless you run on batteries. I'm not even
sure the PFC unit would prevent it.

16

There are two ways the interference can enter your Arduino:-

  1. Through conduction along the wires
  2. Through electro magnetic pickup by the wires

While good shielding helps, you also need much more in the way of supply decoupling and input protection.
See
http://www.thebox.myzen.co.uk/Tutorial/Protection.html
and

17

Thanks for the replies guys, I did what you suggested and moved all the sensor wiring to a single screened cable, and covered the arduino box in aluminum foil.

It did not solve the problem yet, but I found an even bigger problem: the sockets in the walls did not have a proper ground so even the TIG machine is not properly grounded. I have to fix that grounding issue and then we'll see how everything else goes. But that's next weekend :slight_smile:

@Ken - considering there are tig machines worth tens of thousands...my $450 tig welder is definitely cheap :stuck_out_tongue:

I'll be sure to read Mike's links when I'm done with work today.

Have a great week!

18

Oh! Yeah, a bad ground will cause havoc, and is incredibly dangerous.

BTW, that aluminum foil around the Arduino is connected to the Arduino ground?

19

No, I haven't connected it to the arduino ground. I'm thinking it's probably a better idea to connect it to the socket ground, and try not to involve the arduino circuit in any electromagnetic noise capturing task.

Do you think it's better to connect to the arduino ground?

20

Yes, Arduino ground. Do you have copper foil ? So you can solder a wire to the foil.
But the Arduino ground is also connected to the computer via the usb, perhaps some insulation around the foil for protection.

The idea is that the Arduino don't get bad electro/magnetic pulses or noise. When the foil is connected to Arduino ground, the Arduino can only see it's own ground around it, so it feels happy :slight_smile: