RS485 Communication disturb in run time

[vrajesh]

In one project we are using ST1480ABDR (Transceiver) and Baud rate is 115200.

 

One master and 19 Slave connected in star topology.

 

Please find attached reference termination resistor on both side, one at master and only one at slave.

 

Run time our communication disturb. we are using shielded non twisted cable for communication.

[Tesla DeLorean]

Is it more resilient if you use twisted-pairs?

[vrajesh]

AS of now twisted pairs not possible. 

 

Even cable length is max 2.5 meter. so in differential signal communication having 2.5 meter max cabling is fine.

 

What you say?

 

[Tesla DeLorean]

Fine yet not working.. that's a bit of a quandary.

Have you looked at the signals at both end with a scope? What does this disturbance look like exactly? How exactly does it present in terms of communications? Does that change if you drop the baud rate?

Some reason RS485 complaint cabling wasn't used? Who made that decision? Have they debugged this installation with you?

Not possible, or not convenient? Can you test on a smaller scale? Perhaps a handful of nodes?

[Peter BENSCH]

@vrajesh Have you thought about why in your picture "on board termination.png" fig. 6c = star topology is NOT recommended?

You wrote that you have a termination resistor on the master and only one on the slave. However, it is not clear whether:

1. only a single slave has a terminator resistor and the other 18 do not, or whether
2. all 19 slaves have one

Do you realise that you can't get any meaningful communication in either case?
In relation to the above questions, the following cases arise:

1. your bus is most likely flooded with reflections, which is why @Tesla DeLorean asked for an scope picture
2. your transceivers each see the parallel connection of all termination resistors

Regards
/Peter

[Andrew Neil]

---

@vrajesh wrote:

One master and 19 Slave connected in star topology.

 

---

And you supplied this picture:

It tells you that the star arrangement is a bad configuration - so why do you do that?

Why are you surprised when it does, indeed, give you bad results?!

 

And why did you not use twisted pair - as every reference suggests you should?!

eg, 

https://www.csimn.com/CSI_pages/RS-485-FAQ.html#:~:text=Most%20importantly%2C%20it%20must%20have%20a%20twisted%20pair%20for%20the%20two%20data%20lines.%20You%20will%20not%20achieve%20the%20noise%20cancellation%20that%20RS%2D485%20was%20intended%20to%20provide%20if%20you%20do%20not%20use%20twisted%20pair 

 

https://www.analog.com/en/resources/technical-articles/rs485-cable-specification-guide--maxim-integrated.html