# EMI effect on modbus communication casuses failure of system

**URL:** <https://forumautomation.com/t/emi-effect-on-modbus-communication-casuses-failure-of-system/1833>\
**Category:** Industrial Automation\
**Tags:** modbus\
**Created:** [March 16, 2017, 11:37am UTC](https://forumautomation.com/t/emi-effect-on-modbus-communication-casuses-failure-of-system/1833 "2017-03-16T11:37:45Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![Nikhil\_N](https://forumautomation.com/letter_avatar_proxy/v4/letter/n/51bf81/32.png) [@Nikhil\_N](https://forumautomation.com/u/Nikhil_N)\
**Post date:** [March 16, 2017, 11:37am UTC](https://forumautomation.com/t/emi-effect-on-modbus-communication-casuses-failure-of-system/1833/1 "2017-03-16T11:37:45Z")

</div>

I have a compressor and PC communicating in modbus. The distance between master and slave is about 1000ft. I want to analyze the effect of interference on the communication. In actual field we are facing some issues due to the interference. I want to analyze how much error coming due to the interference . Please suggest any methods. Already I have made bursts from relay chattering and coupled to data line. I could see some flush error in the log file. When I coupled white noise, I could see time out error for some instants. please suggest some more ideas

---

<div class="post-metadata">

**Author:** ![carl\_e](https://forumautomation.com/letter_avatar_proxy/v4/letter/c/94ad74/32.png) [@carl\_e](https://forumautomation.com/u/carl_e)\
**Post date:** [March 16, 2017, 11:14pm UTC](https://forumautomation.com/t/emi-effect-on-modbus-communication-casuses-failure-of-system/1833/2 "2017-03-16T23:14:55Z")

</div>

I assume that at 1000 feet it is an RS-485 network.

Disconnect the shield ground. Ground the shield at multiple locations along the length of the cable run to induce ground currents. Put a capacitor between shield and each drive line to induce noise directly onto the signal lines. Lay a long extension cord directly alongside the cable, directly next to it, to induce noise parallel coupling and run a universal motor tool with brushes, like a big drill or jack hammer. Or better yet, welding cable from an electric arc welder.

Some links to notes on RS-485

PROFIBUS: Grounding Tips, Shielding, Noise, Interference, Reflections, Repeaters and more

> **[PROFIBUS: Grounding Tips, Shielding, Noise, Interference, Reflections,...](https://www.smar.com/en/technical-article/profibus-grounding-tips-shielding-noise-interference-reflections-repeaters-and-more)**

AN-979 The Practical Limits of RS-485, Texas Instruments, May 2004

> **[snla042a.pdf](https://www.ti.com/lit/an/snla042a/snla042a.pdf)**
>
> 172.38 KB

Create a long stub (a long lead line from the multidrop drive lines to the device connections, or a branch in a star topology) that will create a reflection. (TI note has a section on stubs)

IEEE paper: Overcoming data corruption in RS485 communication Related Articles (purchased)

> **[Overcoming data corruption in RS485 communication](https://ieeexplore.ieee.org/document/501550/?reload=true&arnumber=501550)**
>
> Data corruption is a frequent problem encountered in RS-485 communication systems. Most of the time this happens due to transmission line effects, imbalance in impedance in balanced pair, inadequate shielding and grounding of cables.

Look at figure 5 to see if breaking the shield offers a way to induce noise.