I have ran a successful code on my Arduino and connected all Hardwares correctly, three of which communicates via I2C protocol (LCD, ADS1115, DS3231) and an SD card module (SPI). I have tested with my computer and everything works okay. On connecting it to the designated machine (MIG welding machine), while the machine was working, the LCD screen sometimes goes off, or it goes blank or shows gibberish characters, and also, the Arduino Board freezes.
I just learnt that this is a case of Electromagnetic Interference. Please how do I go ablout eliminating this?
At the moment, I have ordered decoupling capacitors, 10nF, but i am not sure it will be enough as i am dealing with more of communication.
Dealing with EMI can be frustrating and depends on circumstances. Being in the vicinity of a high current device like a welder can be quite challenging. You may want to try having all of your electronics in a GROUNDED / EARTHED metal box, but don't forget about cooling and getting wires in there. The Ground / Earth wire should be short, heavy, and preferably braided.
vaj4088:
Dealing with EMI can be frustrating and depends on circumstances. Being in the vicinity of a high current device like a welder can be quite challenging. You may want to try having all of your electronics in a GROUNDED / EARTHED metal box, but don't forget about cooling and getting wires in there. The Ground / Earth wire should be short, heavy, and preferably braided.
Thanks, I actually noticed that
So would putting all my components in a metal enclosure that is properly grounded suffice?
What about the I2C devices, would they be shielded by the enclosure? Or do I need a braided cable for that too?
Ceejay90:
So would putting all my components in a metal enclosure that is properly grounded suffice?
The answer is: it depends.
Try it. If that isn't enough, try putting your I2C devices in a grounded metal enclosure also. I don't know your setup so it is hard to give specific advice. I2C devices are not meant to be placed very far from their master - a foot or two, perhaps, but not more. More may work sometimes but it can be bad, especially in a high EMI environment. In a high EMI environment, it is best to keep everything close and short.
There are other things that you can do. Each additional thing is probably more difficult and more expensive. You just keep adding and adding until you get the performance that you want. Notice that I wrote "adding." You almost never subtract what you have added.
EMI can be conducted (getting in via signal or power wires) or radiated (getting in through waves in the atmosphere). A metal box reduces radiated EMI but does nothing for conducted EMI. Feedthrough capacitors may be helpful for power wires (because they reduce conducted EMI) but can limit speed on signal wires. Your 10 nF capacitors are probably not feedthrough capacitors.
Try it. If that isn't enough, try putting your I2C devices in a grounded metal enclosure also. I don't know your setup so it is hard to give specific advice. I2C devices are not meant to be placed very far from their master - a foot or two, perhaps, but not more. More may work sometimes but it can be bad, especially in a high EMI environment. In a high EMI environment, it is best to keep everything close and short.
There are other things that you can do. Each additional thing is probably more difficult and more expensive. You just keep adding and adding until you get the performance that you want. Notice that I wrote "adding." You almost never subtract what you have added.
EMI can be conducted (getting in via signal or power wires) or radiated (getting in through waves in the atmosphere). A metal box reduces radiated EMI but does nothing for conducted EMI. Feedthrough capacitors may be helpful for power wires (because they reduce conducted EMI) but can limit speed on signal wires. Your 10 nF capacitors are probably not feedthrough capacitors.
Thanks for your advice.
I understand that for decoupling, one should use a 10nF ceramic capacitor. Now, this feedthrough capacitor, what is the specified rating? Is it dependent on application, or is there a standard rating? If so, what is it?
On connecting it to the designated machine (MIG welding machine), while the machine was working, the LCD screen sometimes goes off, or it goes blank or shows gibberish characters, and also, the Arduino Board freezes.
How is it connected? Could you provide a circuit diagram?
There might be opportunity to reduce interference here - opto isolation, MOVs, ferrite cores, etc.
Note that there's electronic enclosures available with emi/rfi coating or shielding. There's also emi shielding spray paint available.
dlloyd:
How is it connected? Could you provide a circuit diagram?
There might be opportunity to reduce interference here - opto isolation, MOVs, ferrite cores, etc.
Note that there's electronic enclosures available with emi/rfi coating or shielding. There's also emi shielding spray paint available.
An opto-isolator is one way to isolate digital circuits but is not so good for analog measurements. An isolator for current measurements would be a current transformer, a Hall effect device, or something else.
I suppose that if you had an isolated power supply, you could make an ordinary measurement of current using an ADC with a serial output, and use an opto-isolator on this serial output. Seems a bit complicated when you have an Arduino with a built-in ADC to use.
vaj4088:
An opto-isolator is one way to isolate digital circuits but is not so good for analog measurements. An isolator for current measurements would be a current transformer, a Hall effect device, or something else.
I suppose that if you had an isolated power supply, you could make an ordinary measurement of current using an ADC with a serial output, and use an opto-isolator on this serial output. Seems a bit complicated when you have an Arduino with a built-in ADC to use.
I used an ADS1115, to get current readings from the shunt, wouldn't that be "isolation" in a way?
By "isolation", people are really talking about "galvanic isolation". The ADS1115 does not provide any of that.
Okay,
But I am in a tight spot now as I have imported about 10 of the ADS1115, as it stands, I can't return them to get a current transformer, is there another way?
Feedthrough capacitors are available in various capacities and voltage ratings. They are intended for getting signals (and power) through metal enclosure walls while reducing the amount of EMI.
A feed through capacitor is a three terminal device. The body is intended to be inserted into a hole in the enclosure, and the body makes good contact with the metal of the enclosure, possibly by soldering, possibly by being a press fit. This is one side of the capacitor.
The signal (or power) passes through on the other two pins of the feedthrough capacitor. One pin is outside the enclosure, and connects to the external signal or power. The other pin is inside the enclosure and completes the connection for the internal signal (or power). The two pins (one inside, one outside) are connected (almost zero ohms) to each other, and connect to the other side of the capacitor.
You choose the capacitance by using enough to reduce the EMI to a desired level but not so much that you degrade the signal too much (or damage the driver). This is easy for power (use a lot of capacitance) but may require experimentation for signals. The choice of voltage rating is the same as any other capacitor.
vaj4088:
Feedthrough capacitors are available in various capacities and voltage ratings. They are intended for getting signals (and power) through metal enclosure walls while reducing the amount of EMI.
A feed through capacitor is a three terminal device. The body is intended to be inserted into a hole in the enclosure, and the body makes good contact with the metal of the enclosure, possibly by soldering, possibly by being a press fit. This is one side of the capacitor.
The signal (or power) passes through on the other two pins of the feedthrough capacitor. One pin is outside the enclosure, and connects to the external signal or power. The other pin is inside the enclosure and completes the connection for the internal signal (or power). The two pins (one inside, one outside) are connected (almost zero ohms) to each other, and connect to the other side of the capacitor.
You choose the capacitance by using enough to reduce the EMI to a desired level but not so much that you degrade the signal too much (or damage the driver). This is easy for power (use a lot of capacitance) but may require experimentation for signals. The choice of voltage rating is the same as any other capacitor.
I hope that you do not have to go this far.
Hmm, that is a lot, I will experiment with the feed-through capacitor then.
What if I put decoupling capacitors on the A0 and A1 of the ADS1115, can that help?
The I2C bus on the ADS1115 is going to be difficult to use with an opto-isolator. The clock is (usually) unidirectional and can use an opto-isolator, but the data line is bidirectional. Unless you get an I2C buffer that separates out the two directions, it is going to be quite difficult to use an opto-isolator.
Even if you have an I2C buffer that separates out the two directions, you are going to need an isolated power supply and TWO opto-isolators (one for each direction).
This sort of planning normally occurs PRIOR to operating electronics in the vicinity of a heavily EMI-producing device. I am sorry that you are having to do it now.
vaj4088:
The I2C bus on the ADS1115 is going to be difficult to use with an opto-isolator. The clock is (usually) unidirectional and can use an opto-isolator, but the data line is bidirectional. Unless you get an I2C buffer that separates out the two directions, it is going to be quite difficult to use an opto-isolator.
Even if you have an I2C buffer that separates out the two directions, you are going to need an isolated power supply and TWO opto-isolators (one for each direction).
This sort of planning normally occurs PRIOR to operating electronics in the vicinity of a heavily EMI-producing device. I am sorry that you are having to do it now.
That AC to DC converter is great but I did not see anything on the data sheet about it being galvanically isolated. Maybe it is, maybe it isn't, but manufacturers usually want to get this sort of information onto the data sheet if the information exists.
