I am making a light controller for my house. As a safety mechanism I have included a relay for every high-voltage circuit to be able to remove the load from the electronic dimmer when the light is turned off.
The LCD display is crashing after a while and shows random characters. This happens only when one or multiple relays closes, and after about a second.
There are lots of information on the internet about this very issue. So I have been installing some decoupling capacitors on the power rails between 5VDC and ground, as well as experimenting with RC filters on the contrast and backlight pins on the LCD. All without noticeable results.
Before I do some major changes to my circuit I want to ask for some advise. Will most of the EMI be removed by replacing the relays to SSR with zero-crossing? What values for RC filters is preferred?
I have not made a full schematic for the circuit (as I should have had), but I have included a detailed photo where most is described: Dropbox - File Deleted
That might be true. The producer of the relays offers a range of different plug-in modules with a indicator LED and different types of EMI protections. I choose the one with a varistor. Can that do the same job as a fly-back diode?
No, it might have an appeasing effect but it doesn't do the job of the diode. The diode protects the rest of the circuit from the reverse current of the relay coil when it's switched off. The varistor may limit the voltage so no human risk rises but the circuit gets problems or dies long before that.
I am powering the circuit with a 5A 10VDC AC/DC converter. Then I have one side of the breadboard powered with 5VDC from arduino, and the other side of breadboard I have connected with Vin 10VDC to give enough juice to power the replays.
Have you checked the power supply voltage when this fault occours to see if your supply is overloading or near overload?
No, I have not checked that. But I think the power supply is oversized for this project. I have checked it several times if its too hot, but it remains chill.
Also, since you are dropboxing could you put a video of the display flicker there so we may see it.
I will definitely do that. It may take some time before I get it up, since I have a bussy schedule today.
The photo is of little use when it comes to diagnosing the potential problems but i have a few things.
Rule no1: We do not detect power overloading by testing if the power unit is hot. If its at that stage, pull the plug and run.
What i refer to is that the voltage can spike down when you activate a relay or may drag the voltage down to a level that does not cause the circuit to stop but will cause it to act erratically.
Having said that a power supply able to deliver 5 amps should be ok (assuming you are not drawing 5 amps).
The bread board that you are using is not ment for long term use. Eg. Trusting it beyond a couple of days is asking for it as the connections are made by pressing a wire contact between two metal plates that can loose their spring and corrode (not outdoor usable) leaving you with intermittents that break down more as you draw more current.
I also notice that you have the cables for the led lighting running right besides the bread board 10v rail. As the system is using PWM with voltages between 40 and 100 volts this will lead to what they call capacitive coupling between them causing a ripple on the 10v rail of up to 100v. This also goes for all the other wires.
When you set up a system it pays to arrange things so that all the cables that run either AC or Pulsing voltages at higher levels are physically seperated as much as possable or a grounded metal plate is fitted between them.
Any metal wire that carries a voltage that changes state will cause an electromagnetic field to form and collaps around it (a transmitter).
Any metal wire running in parallel with one or more of these wires will 'recieve' this field and convert it to a voltage.
Seperation is the first step, Shielding the second and filtering is the final thing you should attempt.
Hi,
Are we looking down on the box, or is the box on a wall.
If the protoboard is vertical it will not be very reliable as all your components will be trying to pull out.
As stated before, usng protoboard for long term projects is not recomended.
I also notice that you have the cables for the led lighting running right besides the bread board 10v rail. As the system is using PWM with voltages between 40 and 100 volts this will lead to what they call capacitive coupling between them causing a ripple on the 10v rail of up to 100v. This also goes for all the other wires.
The wires you are referring to is the dimmed power cables to the LED lightning, which are operating at 9v. I think the voltage might be lower as I dim the light down. So then the high voltage is separated as much as possible inside that box.
Seperation is the first step, Shielding the second and filtering is the final thing you should attempt.
All the components inside the box are mounted on a metal plate (isolated between) which is connected to AC ground. Is this the kind of shielding you are describing?
Hi,
Are we looking down on the box, or is the box on a wall.
If the protoboard is vertical it will not be very reliable as all your components will be trying to pull out.
As stated before, usng protoboard for long term projects is not recomended.
Thank you for telling me about this issue. I had a feeling that this was not so good for long-term use, but I thought it was ok as long at the board was standing still. The box is wall-mounted so I will have to look for a better long term solution.
Sumitsubo:
The wires you are referring to is the dimmed power cables to the LED lightning, which are operating at 9v. I think the voltage might be lower as I dim the light down. So then the high voltage is separated as much as possible inside that box.
All the components inside the box are mounted on a metal plate (isolated between) which is connected to AC ground. Is this the kind of shielding you are describing?
Thank you for telling me about this issue. I had a feeling that this was not so good for long-term use, but I thought it was ok as long at the board was standing still. The box is wall-mounted so I will have to look for a better long term solution.
Hi again,
I downloaded the manual for your light dimmer and it indicated that the voltages varied (set by dipswitch) from 40 to 100 volts depending on the current and then used PWM for dimming.
Take note of my comment about AC and Pulsing DC in wires, although to a lesser extent it still happens.
Elecrtomagnetic Waves are like light so if you could shine a light from one wire to the next then they are not shielded.
Unlike light these waves can be stopped or reduces only by a metal wall that is grounded.
Sorry but no nounting them to a flat metal plate is not shielded.
Just a bit more on those waves, they are made stronger by the voltage in the wire and by the current carried by the wire.
So a wire that carries a pulsing 12v at 1 amp will have a stronger field than a wire that has 6v at 1 amp.
This is a VERY SIMPLE explanation of what is called electromagnetic interference.
Finally as i mentioned before your protoboard/veroboard no matter how its mounted is just for prototyping. Not for continued use.
