Modbus problem

[Serega]

Good afternoon.

We have V 350 (Modbus Master) and V 700 (Modbus Slave 2). Modbus works on the RS 485 line and there are other Modbus devices (Slave) on the line. Master works well with Modbus Slave 1 device. But does not want to work with Modbus Slave 2 (V 700). At the same time, when monitoring data on the RS 485 line, there are no problems - slave 2 will respond to the master. I checked the settings of the ports V 700 and V 350. I tried to connect a laptop with Modbus slave simulator instead of V 700 - everything works. Tried to get modbus data from V 700 using laptop - everything works fine. For some reason, the V 350 and V 700 do not want to work with each other. Has anyone experienced something similar?

[Flex727]

Not an answer to your question, but why are you using a RS232 to RS485 converter?

Also, have you tried connecting only the V700 to the V350 without the others?

[Gabriel Franco]

I would  check: 

1. Terminal resistors, only at the ends of the network.

2.  Comms parameters, baud and so on.

3. Register numbers.

4. Delay time between readings of each slave.

 

[Ausman]

All the above from Gabriel and Flex.  Ditto from me on the 232 - 485 adapter....huh?  This may not work as expected given it's likely designed to be attached to a PC and run with specific drivers.  Do you know that the 350 can do 485 all on it's own?  If you don't, read about it in the install manual for the unit, with the illustrations of the Controller Port & Jumpers under the Pinouts heading.  And the associated question with using the 232 port is how are you connecting to the 350 to monitor things etc via Visilogic?  Does it have an ethernet card fitted?  If not, strongly suggest to do it this way, then your serial port is only for 485 use.

But to truly answer your questions, we need to see the insides of the Modbus Config blocks for both units.

 

[Serega]

22 hours ago, Flex727 said:

Not an answer to your question, but why are you using a RS232 to RS485 converter?

Also, have you tried connecting only the V700 to the V350 without the others?

I can't tell you the reason for using RS 232-RS 485 here. It was done a long time ago. I can assume that to isolate the field bus from the PLC bus. No, I have not tried to disable other devices on the bus. I'm going to make a note to try this.

[Serega]

22 hours ago, Gabriel Franco said:

I would  check: 

1. Terminal resistors, only at the ends of the network.

2.  Comms parameters, baud and so on.

3. Register numbers.

4. Delay time between readings of each slave.

 

1. Yes, there may be a problem with the resistors. I need to hook up an oscilloscope to find out. But it is unlikely that the problem is in them - there is data exchange over the bus.

2. I checked the port settings. I even tried changing the parity. It didn't work.

3. Everything is fine with register numbers. I checked. The Modbus Slave device responds to the request and sends data to the master. The master does not receive data.

4. On modbus slave 2, I tried to change the timeout from 0 to 100 (x10 msc). No result.

[Serega]

19 hours ago, Ausman said:

All the above from Gabriel and Flex.  Ditto from me on the 232 - 485 adapter....huh?  This may not work as expected given it's likely designed to be attached to a PC and run with specific drivers.  Do you know that the 350 can do 485 all on it's own?  If you don't, read about it in the install manual for the unit, with the illustrations of the Controller Port & Jumpers under the Pinouts heading.  And the associated question with using the 232 port is how are you connecting to the 350 to monitor things etc via Visilogic?  Does it have an ethernet card fitted?  If not, strongly suggest to do it this way, then your serial port is only for 485 use.

But to truly answer your questions, we need to see the insides of the Modbus Config blocks for both units.

 

Yes, I know that the 350th (V350-35-B1) can be configured as 485. But this was done a long time ago. I'll try to do it directly if I can't find a solution to the problem. 350 has two serial ports. One port is a modbus master, the second port is a modbus slave - it transmits data to SCADA. If necessary, I connect to the second port using VisiLogic . There is no ethernet port in 350, only EXP Port. Yes. I will try to post the port settings here.

[Serega]

[Serega]

Settings

[kratmel]

Slave 2 Baud Rate --- 9600  ≠  Master ---19200

[Ausman]

4 hours ago, Serega said:

There is no ethernet port in 350, only EXP Port.

Your 350 in theory only comes with one com port standard.  Extra ports are optional, be it 232 485 or ethernet.  In your case it appears you have another 485 port fitted. 

In addition to Kratmel's observation on baud rate, you have 232 as the chosen standard (top right) which is likely the culprit...change it to 485.  Also, in looking at the times in the data record,  to me there is also something wrong with your timeouts and retries.  Make sure you aren't getting a clash of timings/requests etc that completely throw things out of sequence.

Please note that another way you could talk much better b/n the Unitronics devices is UniCAN.  Extra cards needed, but easy to do and fast.  But I think some minor corrections here and there will fix your issues. I still ponder the converter!  What is the brand and model?

cheers, Aus

[Ausman]

Looking again at your calls, an observation to be aware of.  I have found that I need a few scans for more complex 485 reads to be done successfully, even though the system indicates things have finished via Function in Progress.  I have always put this down to buffer issues.  But other forum members successfully have comms working as you do, with calls in theory happening immediately when F i P finishes. If nothing of the above seems to fix the issue, maybe try incorporating a small delay b/n calls to see what happens.  I still think the timings shown seem a bit odd, as if something else is affecting the calls.  But it might also be what I've just written about.

Also, many members do not do multiple tries, just one attempt.   "If it don't work, it won't."  I personally do 2 attempts, and also monitor results to ensure things change as expected over time.

cheers, Aus

[Serega]

21 hours ago, kratmel said:

Slave 2 Baud Rate --- 9600  ≠  Master ---19200

Yes, this is my mistake. It appeared because I tried to restore settings from archived projects in the office. In this case, I missed the speed. In fact, the speed is 19200, not 9600. If I had different speeds, then slave 2 would not give an answer to the master.

[Serega]

6 hours ago, Ausman said:

Looking again at your calls, an observation to be aware of.  I have found that I need a few scans for more complex 485 reads to be done successfully, even though the system indicates things have finished via Function in Progress.  I have always put this down to buffer issues.  But other forum members successfully have comms working as you do, with calls in theory happening immediately when F i P finishes. If nothing of the above seems to fix the issue, maybe try incorporating a small delay b/n calls to see what happens.  I still think the timings shown seem a bit odd, as if something else is affecting the calls.  But it might also be what I've just written about.

Also, many members do not do multiple tries, just one attempt.   "If it don't work, it won't."  I personally do 2 attempts, and also monitor results to ensure things change as expected over time.

cheers, Aus

Okay, I'll try to follow your advice in a few days when I'm around the controllers.

Converter model https://expertdaq.com/en/product/ex9520/