Noise issue

Hi all,

I’m getting electrical noise coming from my 8 channel relay board. The relay board is solid state and isolated from the arduino (according to their specs). When I run my program with the relay board disconnected my LCD is working fine. When it is connected, (even just starting up) I’m getting noise going to the LCD. I have plenty of bypass capacitors going along my power rails of my breadboard including a small ceramic cap very close to the lines that run to the LCD pins 0 & 1. I have a RC Snubber attached to the only relay that is turning on upon startup and yet i’m still getting feedback.

What am I missing?

Heres my parts list:
Arduino Mega 2560
8 channel Solid state Relay: Amazon Link

RC Snubber: Part Dist. Page
DataSheet PDF

Wiring schematic is attached. RC Snubber and filter capacitors are not shown on the schematic yet.

Any ideas?

Could try a separate DC supply for the arduino and LCD maybe. Have 1 supply for handling the relay side of things. And 1 supply for the arduino (and LCD).

Do you still have the issue when the relay board is connected to the Arduino, but the loads are disconnected from the relays?

Jobi-Wan:
Do you still have the issue when the relay board is connected to the Arduino, but the loads are disconnected from the relays?

I do not. When power is shut off to the solenoids the relays control the board runs fine without any LCD issues. I suppose i'm confused cause I put a RC Snubber on the one relay that does come on during startup and it still does it. Very odd. The relay is driving 24VAC so its not a huge amount of power, i'd say less then 1A probably around half an amp max, I haven't measured it but I can defiantly if needed.

Did you try a snubber across the valve coil as well as the relay contacts?
Also a 240V might be too "hard" for 24V, I would have chosen a 120V (or less) part.

We are talking about EMI here, its not power thats the issue but the extreme dV/dt events on un-snubbed
relays (arcing basically). You shouldn't be switching highly inductive loads like solenoids without
snubbing, otherwise there'll be fast 100--1000V spikes zinging around causing mayhem.

You need the correct snubber RC values for the load and frequency for an AC circuit, ie large enough
capacitor to reduce the voltage spike enough, but not so large as to leak too much power (its a compromize).

Perhaps MOV / TVS snubbing would be better, just choose a device higher than the peak voltage of the
24Vac (say 40 to 50V).

And shielding the sensitive circuitry can be of some benefit too.

Blade2021:
I do not. When power is shut off to the solenoids the relays control the board runs fine without any LCD issues. I suppose i’m confused cause I put a RC Snubber on the one relay that does come on during startup and it still does it. Very odd. The relay is driving 24VAC so its not a huge amount of power, i’d say less then 1A probably around half an amp max, I haven’t measured it but I can defiantly if needed.

I guess that your problem is not when the relay comes on but when it goes back off.
Your load is inductive, and just disconnecting it can give you all kinds of positive and negative voltage spikes.

I am by no means an electronics expert, but I had an issue a while ago that I suspect was similar, but it was not using AC. The solution was to put 2 zener diodes in series in opposite direction across the load:
–|<|—|>|–
For 24V, I guess you should use 33V or 36V zeners. High enough that they don’t start conducting under normal circumstances.

Edit to add: This is a picture of the circuit that caused me issues. The symptom was that the MCU would reset.

edgemoron:
Did you try a snubber across the valve coil as well as the relay contacts?
Also a 240V might be too “hard” for 24V, I would have chosen a 120V (or less) part.

I have not tried that yet but I will defiantly do that before ordering any more parts. True, What is your opinion on this alternative I found on mouser.

http://www.mouser.com/ProductDetail/KEMET/PMR205AB6100M100R30/?qs=sGAEpiMZZMv1cc3ydrPrF85GIl804cbNu4sd0CEBJ%2FM%3D

MarkT:
We are talking about EMI here, its not power thats the issue but the extreme dV/dt events on un-snubbed
relays (arcing basically). You shouldn’t be switching highly inductive loads like solenoids without
snubbing, otherwise there’ll be fast 100–1000V spikes zinging around causing mayhem.

You need the correct snubber RC values for the load and frequency for an AC circuit, ie large enough
capacitor to reduce the voltage spike enough, but not so large as to leak too much power (its a compromize).

Perhaps MOV / TVS snubbing would be better, just choose a device higher than the peak voltage of the
24Vac (say 40 to 50V).

And shielding the sensitive circuitry can be of some benefit too.

I found this: Mouser If you have any good recommendations I’d gladly take them. The MOV/TVS would still go across the relay contacts correct?

Jobi-Wan:
I guess that your problem is not when the relay comes on but when it goes back off.
Your load is inductive, and just disconnecting it can give you all kinds of positive and negative voltage spikes.

I am by no means an electronics expert, but I had an issue a while ago that I suspect was similar, but it was not using AC. The solution was to put 2 zener diodes in series in opposite direction across the load:
–|<|—|>|–
For 24V, I guess you should use 33V or 36V zeners. High enough that they don’t start conducting under normal circumstances.

Edit to add: This is a picture of the circuit that caused me issues. The symptom was that the MCU would reset.

I have to disagree. I’m not an expert by any means so don’t take this as being rude (absolutely not trying to be rude. and I do appreciate your opinion.) The LCD is displaying “odd” characters happen as soon as the relay kicks in the “on” position. It doesn’t go off in the scenario i’m “troubleshooting.”

Thank you all for your time and feedback, It is greatly appreciated!

Blade2021:
The relay is driving 24VAC...

The relays are made for mains power (100-240volt).
Specs of the actual relay state an absolute minimum of 75volt AC.
There are three versions of the relay. Some already have buildin snubbering.
Leo..

Hey Blade,
I had a problem with AC voltage spikes using an opto-isolated relay and used a MOV to clear things up, the EMI gets through the relay to the board even with the isolated relay.

Heres the post if you want to look, maybe this might be your problem.

What am I missing?

I suspect several things. I agree with Wawa that the relays are not rated for 24vac. This may be causing slow turn-on since the triac gate current is obtained from the supply voltage. This can make turning an inductive load on very problematic.

Unless you can verify turn-on with a scope (don't do it unless you understand the danger and the need for additional isolation), I would try using electromechanical relays for the 24vac voltage switching (with subbers across the contacts).

The other unknown is what your wiring looks like. If you wire things anything even close to the horrible way Fritzing draws schematics, you could be creating yet more grief for yourself with series connected GND points without regard to the current in those wires. Ideally, there should be one common ground point with all common GND's coming to that point in a star fashion, not wires that go in series from point to point to point.

Lastly, your LCD contrast pot is shown incorrectly. The correct pot wiring is in voltage divider mode to place a variable voltage on the V0 pin. The way the schematic shows it, it can take your +5 supply from the Mega to ground when rotated to one end. VDD only goes to the pots CW end, it is shown connected to both the CW end and the wiper.

ribbonman:
Hey Blade,
I had a problem with AC voltage spikes using an opto-isolated relay and used a MOV to clear things up, the EMI gets through the relay to the board even with the isolated relay.

Heres the post if you want to look, maybe this might be your problem.

relay isolation problems - General Electronics - Arduino Forum

I have to agree. I think snubbing is my best bet but just looking for recommendations at this point. I don't wanna spend money on a part that isn't "fit" for the values my circuit will be using.

P.S. Nice post! I actually grabbed some links from your OP from that thread. Had some very descriptive information that will defiantly help going forward.

Wawa:
The relays are made for mains power (100-240volt).
Specs of the actual relay state an absolute minimum of 75volt AC.
There are three versions of the relay. Some already have buildin snubbering.
Leo..

Yes, agreed, the relay board isn't "rated" for that amount of low voltage but I was unable to find a relay board for that low of voltage while staying within the same similar design and solid state. So kinda hoping that isn't the issue.

avr_fred:
I suspect several things. I agree with Wawa that the relays are not rated for 24vac. This may be causing slow turn-on since the triac gate current is obtained from the supply voltage. This can make turning an inductive load on very problematic.

Unless you can verify turn-on with a scope (don't do it unless you understand the danger and the need for additional isolation), I would try using electromechanical relays for the 24vac voltage switching (with subbers across the contacts).

The other unknown is what your wiring looks like. If you wire things anything even close to the horrible way Fritzing draws schematics, you could be creating yet more grief for yourself with series connected GND points without regard to the current in those wires. Ideally, there should be one common ground point with all common GND's coming to that point in a star fashion, not wires that go in series from point to point to point.

Lastly, your LCD contrast pot is shown incorrectly. The correct pot wiring is in voltage divider mode to place a variable voltage on the V0 pin. The way the schematic shows it, it can take your +5 supply from the Mega to ground when rotated to one end. VDD only goes to the pots CW end, it is shown connected to both the CW end and the wiper.

I wish I had access to a scope. (Long story short, the company I work for doesn't see it as a "priority" to invest in. So I have to tackle issues like this blindly. My wiring isn't the best by any means, but I've kept it as clean as I can. I have all my grounds attached to the ground rail of my breadboard which I think is "ok"? correct me if i'm wrong. Then from the ground rail to the arduino, so the ground rail acts as my common point. Again if this is bad practice please do let me know.

The pot is most likely just an issue in the diagram. It is tested and working properly the way its wired, with the adjustment screw controlling the brightness of the LCD. So I'll be fixing this on the schematic very soon.

---- Continuing discussion ----
What would be better for fast acting switching? MOV's or a RC snubber? The loads are switching very fast so speed is very important. The load is a 24VAC solenoid, and pulls 1.3 amps when switched ON.

EDIT:
Looking at DigiKey for MOVs and ratings. I found these two: MOV1 MOV2

The main difference between the two is the "Energy" rating. Can someone please explain this?

Can I suggest that you persuade your company to invest in a modern scope?

eg Rigol (there are many others) do some good ones for a few hundred quid, and you'll find it useful for many other things, and will know what you're doing.

Given the man-hours you'll save, it's not that expensive

Alan

The energy rating is in joules which a standard SI rating and when viewed from the standpoint of electricity, one joule is equivalent to the dissipation of one watt for one second.

Simple translation: the higher the energy rating, the more energy the mov can absorb before letting the smoke out. Notice that the 10j part is physically larger than the 3j part? Now you know why.