Relays and Microcontroller reset issue

[jrz126]

Whats the value of the pullup resistor on the reset pin? If it isn't pulled up hard enough, some noise could cause a reset.

Also, there should be a register in the pic that would indicate the reason for the reset occurring.

 

[emer dauz]

i think you need inductance filter.

 

[ljcox]

How are the untis powered?
Do they all share one power supply?
OR
Does ech have its own supply?

What type of power supply are you using? eg. a plug pack.

As I understand your problem, when a relay is operated or released on one unit it can cause a reset on another.

Therefore, a transient pulse must be coming through thwe power supply or the data link.

Have you tried disconnecting the dta link from one and see if it still resets when a relay is operated or released on another?

Have you tried operating one unit from a 12 Volt battery in order to isolate it from the power supply? This would prove whether the pulse is coming via the supply line.

As emer dauz implied, you may need a low pass filter on the 12 Volt supply lines.

 

[alvinyow]

the pull up resistor is 4.7k

 

[alvinyow]

How are the untis powered?
Do they all share one power supply?
OR [no]
Does ech have its own supply?
yes, each using own 12Vdc adapter]
What type of power supply are you using? eg. a plug pack.
[regulated switching power adapter 12Vdc 3A - similar to notebook adapter]
As I understand your problem, when a relay is operated or released on one unit it can cause a reset on another. [yes, it does happen, not every switch, but it does happen quite frequence]

Therefore, a transient pulse must be coming through thwe power supply or the data link. [i do not have the scope. it become quite difficult for me to troubleshoot the problem]

Have you tried disconnecting the dta link from one and see if it still resets when a relay is operated or released on another?[did not tried this]

Have you tried operating one unit from a 12 Volt battery in order to isolate it from the power supply? [it is isolated since i use one adapter for each module] This would prove whether the pulse is coming via the supply line. [hmmm......]

As emer dauz implied, you may need a low pass filter on the 12 Volt supply lines.

[any pointer?]

[attached is the detail drawing and the pcb layout ... ]

 

[nsaspook]

Your 5 and 12 volt grounding paths on the PCB from the 2803A (9) and pic (19) share a long narrow twisty path that's a prime suspect for ground bounce when the relay switches.

 

[MikeMl]

Yes, you have a ground loop!!! The negative power supply return net from the PIC/analog stuff should have been keep entirely separate from the negative power supply return net from the relay coils/relay driver IC. Those two separated ground trace nets should have only one cross-connection on the PCB. The only place where the two nets are tied together is at the negative terminal of C8.

 

[ljcox]

I agree with nsaspook & MikeMl.

Is it possible for you to make temporary changes to the PCB to eliminate the ground loop and separate the low & high current gnd tracks as MikeMl suggested?

However, I don't see how this would cause one unit to reset another. As I wrote above, I feel there must be some coupling through the power line or the data line.

 

[alvinyow]

Would you think that this will be a better modification to solve my problem?
Appreciate all comments.

 

[nsaspook]

I agree with nsaspook & MikeMl.

Is it possible for you to make temporary changes to the PCB to eliminate the ground loop and separate the low & high current gnd tracks as MikeMl suggested?

However, I don't see how this would cause one unit to reset another. As I wrote above, I feel there must be some coupling through the power line or the data line.

I agree that it might not be the only problem. I really don't like the design of the AC part of the relay switching, with no zero crossing or snubbing on the contacts it's a monster EMI generator that IMO should be completely separated from the controller for reliable operation with this PCB design. I would ground or set to outputs any unused pins on the uC to get a little more immunity from noise and take a hard look at the xtal circuit for dropouts during AC switching arcing.
I've had my own problems with relays and controllers not being friendly, so I know what a pain it can be.

https://www.electro-tech-online.com/custompdfs/2010/08/AN2764.pdf
https://www.electro-tech-online.com/custompdfs/2010/08/AN2321.pdf

 

[ljcox]

This is good advice. But it may not be necessary to move the relays to another PCB.

I have just completed a garage door opener controlled by a PIC.

It switches the motor on/off using relays mounted on the same PCB as the uP.

However, I connected a resistor in series with a capacitor across the relay contacts and have had no reset problems.

The values of the R & C need to be chosen according to the load characteristics.

 

[alvinyow]

what about MOV? can MOV make some changes?

 

[mrdauz]

reset suddenly occurs when spike appear around circuit due to insufficient supply current. i recommend u use switching mode power supply

 

[alvinyow]

the 12Vdc adapter i'm using is a switch mode power supply.

 

[ljcox]

what about MOV? can MOV make some changes?

You need to understand what you need to do.

When the relay contacts open, there is a large & rapid change in current & voltage.

This creates a wide spectrum EMI. So what you need to do is to slow down the change.

A MOV will certainly limit the maximum voltage but it will do little to slow the transient.

Whereas, a RC circuit will slow the transient and thus reduce the EMI.

I suggest that you do an internet search for "suppressing motor transients", "transient suppression", motor transient suppression, etc.

Please let us know what you find.

 

[mrdauz]

alternatively u are able to separate the supply between controller and load with optoisolator like PLC had.

 

[ljcox]

I have attached an extract from my Garage door opener circuit for your information.

Note that the R & C values need to be selected according to the load characteristics.

 

[mrdauz]

All uC ports are isolated with optoisolator. Do not use common power supply between controller and load

 

[alvinyow]

i had tried with MOV today. It suppress some spikes and bring the controller to a more stable state.
So far had not found any reset problem. And neighbour module does not affected with the switching too.

 

[alvinyow]

alternatively u are able to separate the supply between controller and load with optoisolator like PLC had.

Will try this in my coming trial.