RS485 burning up when connecting to thermostat

[hemanteda]

RS485 burning up when connecting to thermostat/relay module

I'm trying to connect a micro-controller based RS845 node to a relay controller (PIC micro + MAX485). The relay controller is mains powered and supplies 24V AC to nodes (has internal 220 to 24V AC transformer). My node is powered from this 24V AC, I'm using a bridge rectifier +LM2596 based circuit to get 5V DC output. Here is what happens,

When I power my node from a separate transformer of the same type and rating (220VAC to 24VAC) and connect just the RS485 high & low wires to the relay controller, all works well.

If I power my node from the relay controller but leave the RS485 unconnected, output voltage is still 5VDC and nothing heating up. Then as soon as I connect the 485 connections, the PIC starts to receive a lot of null chars, everything starts to heat up, and one or the other component blows up.

I'm in desperate need of some expert advice, thanks

 

[paulfjujo]

(has internal 220 to 24V AC transformer

is it a transfo with separate coils or an auto-transfo
link bewteen primary and secondary...
test insulation between input and output of transfo..
what about the earth and 0V of 5V power supply ?
do you have to respect Phase and Neutral on main AC power ?

 

[xenos]

The easy way:

Of course you could use opto isolation of some kind, or a signal transformer or an 1Watt isolation power supply for your transeiver.

The hard way is to understand what is happening:

1) Are you sure the relay controller supplies for your cirquit 24Vac or you just took 24Vac from it's internals?
2) Can you add some kind of schematics with the connections you make and the relay controller power supply (or a block diagram)?
3) Does the relay controler has a bridge rectifier or some other form of regulation? eg with two diodes and common ground?
4) It is very important to understand how is the relay controller powered and how your cirquit.
5) It would be better if you could get DC from the relay controller with common ground with its MAX485.

 

[KlausST]

Hi,

You don´t write about GND wiring.

For RS485 you needm to connect the signal lines and the GND line.

***
for a first try (without damaging a RS485 device) just connect both GND with a 1k resistor.
If it gets hot or even burns there is a problem with non isolated power supplies.

If the voltage across (AC and DC) is below 1V than it should be no problem. Then you can connect the signal lines too.

Klaus

 

[asdf44]

Hi,

You don´t write about GND wiring.

For RS485 you needm to connect the signal lines and the GND line.

***
for a first try (without damaging a RS485 device) just connect both GND with a 1k resistor.
If it gets hot or even burns there is a problem with non isolated power supplies.

If the voltage across (AC and DC) is below 1V than it should be no problem. Then you can connect the signal lines too.

Klaus

This.

You have a problem with ground potentials and it's easiest to analyze and solve that directly.

RS-485 is differential, but not isolated. That means that both boards need grounds that are at the same or nearly the same potential (though some RS-485 chips allow operation with wider common mode or fully isolated voltages). The very best way to ensure this is to have a ground between them.

 

[hemanteda]

@paulfjujo
it is a step-down transformer, no link between primary and secondary... I'm attaching a block diagram of the setup



@xenos
1) the relay controller supplies 24VAC for pumps/valves etc
2)attached block diagram
3)the 24VAC supplied from the relay board is AC not rectified, however i do have a bridge rectifier on my board before the LM2596 section
4)the diagram shows the connections
5)DC from the relay board is not provided, have to use 24VAC only


@KlausST
I have not connected the grounds of the 2 boards! I thought that 485 only needs 2 wires, I'll try your suggestion and report back

If this is ground related what would be the easiest way to solve without connecting the grounds? The problem is that the relay board doesnt give out a ground... just 2 wire RS485

Also given that I'm only using one pic, one max485 and one max232 is there a simpler way to power my board then using the LM2596 circuit?

Thanks everyone, really appreciate it

 

[xenos]

Re: RS485 burning up when connecting to thermostat/relay module

The 24Vac rail you got is probably different from the rail that powers the relay board, so you have voltage between the grounds.
So, if you cannot connect to the DC of the relay board and have common ground,

you can have:

1) https://www.maximintegrated.com/en/products/interface/transceivers/MAX1480A.html
or
2) power your board with this:https://www.mouser.com/ds/2/141/oej-sc-wc1224-22984.pdf
or
3) use an one to one transformer (24Vac to 24Vac) and power your board. This is an economic (but space consuming) approach.

Of course there are a lot of other solutions like rs485 driver+transformer, isolated 1Watt smps and optocoupler driving the rs485 IC etc

 

[KlausST]

Hi,

There is a problem, because both boards take the same AC power lines, so they are not isolated. On the other hand it is not clear what is the GND potential at the one board and the other board.
They are not isolated, but they don't use the same GND either.

The GND to GND voltage and current is not predictable.

To solve this I'd use a small (ready to buy) DC/DC converter to generate a true isolation. Install it at the board with the lower current consumption.

***
You say there are only two wires, no GND connection...
You need at least a protection circuit for the RS485 ICS and a resistor circuit to keep signal levels within the specified input range (often +/-7V)

Klaus

 

[DraganM]

Send you good doc for RS485 and solutions for noise, short circuit, safe, ... good rs485 communication, in attachment

 

[hemanteda]

Re: RS485 burning up when connecting to thermostat/relay module

@xenos Thanks for the help, I'll look into MAX1480 solution

@KlausST You were absolutely spot on about this. I tested the GND to GND difference, its about 9volts!