How to detect Modbus Communication Failure at Modicon M580?

[RuchirSomani]

Hi,

I need to understand how to find/configured in Control Expert for M580 whether Modbus Communication with any Modbus Slave device is working or has stopped ?
Is there any fail/error bit/code to sense this ?

Right now as values gets transferred between M580 and PSO over Modbus TCP,  the value gets stuck in PSO if downstream devices stop communication with M580 for any reason.
Is it also possible to convert this value automatically to "ZERO" in case the Modbus communication with field device breaks.

[Hennie_Smith]

Good day,

1. Yes you can monitor the Freshness bit in the Device DDT variable which is created when adding modbus device.

Secondly to set the values to 0 if comms fail, open the NOC/CPU in the DTM browser and go to the Request settings of the modbus device. Each request can be set to 'Set to Zero' under the Last Value column. 

Regards,

Hennie Smith

[YONGHO-KIM]

1. I have implemented communication fault/normal status detection using the 'Gest' value from Read_Var.
The second word of the GEST value is 0 when communication is normal, and any value greater than 0 indicates an error.
Please refer to the attached PDF file for more details, and feel free to contact us if you have any questions.

[RuchirSomani]

Hi,
Thanks a lot for your reply.

1) For what you have shared with NOC card, is same feature available with NOM card as well as we are polling around 10 Nos. of Modbus Serial devices using NOM card in M580.

2) Also, we have observed randomly data of MFM(Multi Function Energy meter)getting stuck while using READ VAR block to poll MFM(Multi Function Energy meter) for 10 Nos. MFMs in one single loop.

What is further way to diagnose such issues  i.e. how to understand if query length is too high or if we need to increase delay period between 2 polls of Modbus - this is for Modbus Serial(RTU) communication only. Does M580 inherently supports protocol tracing ?

Looking forward to your support.

Thanks & Regards

Ruchir Somani

[ag56]

Hello Ruchir,

There are some key concepts to take into account when implementing READ_VAR function.

First, you must specify the timeout period in the third register of the GEST array. A value of 5 means 500 ms and should normally be enough.

Please note that the NOM is configured with its own default answer delay (10 * 100 = 1000ms) and 3 retries. The timeout set in the third register should be greater than the answer delay time X number of retries.

You can reduce the default answer delay and set the number of retries to 0 to speed out the process if needed.

Second, it is recommended to expose the READ_VAR EN-ENO pins, then negate the EN pin and connect bit 0 (Active) of the first GEST register to it. This will disable the block call when a communication session is active.

Third, for serial communication avoid using more than one READ_VAR in parallel for the same NOM channel. If you are polling several devices on the same line, you need to manage the order in which they are called. For this purpose I am attaching a DFB based on DATA_EXCH function which is designed to manage as many devices as required sequentially and cyclically by proper parametrization. This all-in-one solution DFB requires no extra code. However, I do not have it documented in English so you will probably need some help to implement it. If you are interested in testing it, I can provide some simple examples.

Best Regards.

[YONGHO-KIM]

1. The file I attached was applied to the BMXNOM card. Therefore, the same configuration can be applied to the NOM card as well.
2. Does the freezing issue occur when communication is lost with the device connected to the BMXNOM0200?
If so, please refer to the following link for more information:
https://community.se.com/t5/Modicon-PAC-Forum/BMXNOM0200-READ-VAR/m-p/487416#M1274
3. If the issue is not related to point 2, please increase the polling interval at the READ_VAR EN input as shown in the attached image.
4. Compare the GEST values between normal and abnormal communication cases.
5. If you have any further questions, feel free to reply.

[RuchirSomani]

Hello,

Thanks for your reply. We also feel that there is problem of proper sequencing as we are trying to poll in total 37 devices across 2 NOM cards.

We have also came across a scenario wherein we kept just one "READ_VAR" block for one particular channel i.e. one serial channel dedicated for 1 RS485 Modbus device and after loading the program it does not generate any Modbus Query (NO ORANGE LED INDICATION ON NOM CARD).
And as soon as we shift this entire "READ_VAR" logic to some different page of the logic section, it starts generating modbus query without any changes(just the change of page).

It shall be really helpful if you share some example how to carry out necessary time grading for each Modbus poll so as to avoid overlapping.

Meanwhile to get rid of RS485 polling issues, we also tried using Modbus Serial to TCP converter in between and then carry out polling using NOC Card. But there as well, we faced strange challenge wherein as soon as we create 2nd request, the data goes haywire. Please find attached photo & video for your reference.

Request you to guide how we can resolve this issue for NOC card as well.

[YONGHO-KIM]

1. The number of Modbus TCP devices that the BMENOC module can support is as follows:

   • Please check whether the number of devices exceeds the supported limit.
   • If you could upload the project file, we will review the configuration settings for you.
   • 

2. Regarding the BMXNOM module, please refer to the information I previously provided.

   • Again, if you upload the project file, we will check the configuration settings for this part as well.

[RuchirSomani]

Hello,
Thanks a lot for your reply. Please find our response.
1. We are trying to get data from only 1 No. TCP Server installed(as Modbus Serial to TCP Gateway Device) for which we shall at maximum create 13 Nos. of request with each request polling 96 nos. of 16-bit registers.

2. We have gone through your suggestion for BMXNOM card as earlier shared, but things are not yet particularly clear how we implement proper polling pattern for total 38 Nos. of Modbus Serial devices directly on 2 Nos. of NOM cards.
Please find attached our configuration for your further review where we have configured RS485 based serial polling for 38 Nos. of devices across 2 Nos. of NOM cards.