Overlapping data on Subnet with AB7072-B

Hey there, I have problems with overlapping data on the subnet. I have five rectifier devices connected on the subnet transmitting and receiving some control and status bit as well as values. In normal condition (only connected - no rectifier is on duty/charges) the communication is fine and I can talk to any of the five devices.
But if I start one of the devices (in the moment it is only device 1 - all others are working), which means that the rectifier start charging load with up to 100A the bus communication is influenced.
The Logger shows that there are sometimes two data packages overlapping. We installed a oscilloscope and could see, that the time between the start of two data packages went up from 250ms to around 800ms.

Shield of the bus is grounded on the device side and on the coupler side. 120Ohm resistor between pin 8 and 9 is installed. connected pins are 5 (GND) and 8 and 9.

Anyone have some input for this behaviour.

Hello @Barnetta ,

It’s not clear what you mean by overlapping data. Do you have two devices that start talking on the bus at the same time or if you have EMI/EMF.

What protocol is being used? If it is modbus I would not expect there to be cross talk. unless some mistake is made like two devices set to the same Node ID.

From your second paragraph it seems like you might have an EMI/EMF issues. Either the current is coupling onto the data lines or their is a magnetic field from the system interfering. Ensuring the data lines are not near or around any power lines will help. I would also recommend using shielded twisted pair wiring. In addition to the 120Ω you may need a voltage bias, using 470Ω to pull the + line up to 5vDC and the - line to ground. Here is a diagram showing the wiring.
image

Hello @deryck_hms,
thanks for your fast reply and your suggestions.
The overlapping data blocks I mean are as following:

normal data package:
0x02 2 .
438 2 0x30 48 0
439 1 0x39 57 9
440 2 0x30 48 0
441 2 0x34 52 4
442 1 0x30 48 0
443 2 0x31 49 1
444 1 0x30 48 0
445 2 0x31 49 1
446 1 0x46 70 F
447 2 0x31 49 1
448 1 0x04 4 .

data package where two data packages merged together.

461 3 0x02 2 .
462 2 0x30 48 0
463 1 0x41 65 A
464 2 0x38 56 8
465 1 0x34 52 4
466 1 0x30 48 0
467 2 0x32 50 2
468 1 0x30 48 0
469 2 0x30 48 0
470 1 0x30 48 0
471 2 0x30 48 0
472 1 0x04 4 .
473 1 0x34 52 4
474 2 0x30 48 0
475 2 0x31 49 1
476 1 0x30 48 0
477 2 0x31 49 1
478 2 0x46 70 F
479 1 0x31 49 1
480 2 0x04 4 .

I checked the suggestions you already made and I can only see that we are missing the 470Ohm resistors so far.

220916 - HMSD Support - 100ms - M02.cfg (16.0 KB)

Hope it worked to attach the configuration and also the log file.

220919 - 12-59 - logbook - M02 - start Rec 1.txt (30.6 KB)

Additionally I attached two pictures, where I made some measurement of the bus and saw the increased time between two data packages after I started the charging process of the first device.


Hi Barnetta,

Can you be more specific on what you are looking for help with? What exactly is not working as expected? Nothing looks wrong in the config, log or in your capture.

I don’t think you have an issue with EMI/EMF if you did i would not expect to see valid data in the log. From the log it looks like the communicator is sending a message and another device is responding back. This might be what you are seeing in the easements. When you only see one “data packet” this might just be the communicator sending a message