Please Note This forum exists for community support for the Mango product family and the Radix IoT Platform. Although Radix IoT employees participate in this forum from time to time, there is no guarantee of a response to anything posted here, nor can Radix IoT, LLC guarantee the accuracy of any information expressed or conveyed. Specific project questions from customers with active support contracts are asked to send requests to support@radixiot.com.

I'm trying to read and write to a PLC via a gateway using modbus4j. I've gotten the read function to work, but I'm not getting anywhere on the setValue() method. I've tried various variations of this:

IpParameters params = new IpParameters(); params.setHost("192.168.0.100"); params.setPort(502); //create modbus TCP master ModbusMaster master = new ModbusFactory().createTcpMaster(params, false); master.init(); int regValue=0; regValue = regValue | 1 << 0; ModbusLocator loc0 = new ModbusLocator(1,RegisterRange.HOLDING_REGISTER,0,DataType.TWO_BYTE_INT_UNSIGNED); master.setValue(loc0,8);

Also the getValue() method only works sporadically maybe 30% of the time I call getValue() I get the response I expect, and the other times I get all 0s out.

System.out.println(master.getValue(33,RegisterRange.HOLDING_REGISTER,70,2));

Am I misunderstanding some aspect of the get and set value methods? Thanks.

The line is not in the CVS code. Perhaps it still is in the src jar file. Try using the repo source instead.

Hello,

I'm currently working on a MODBUS TCP project. Right now, I'm able to read out and write to the network/sensors.

My next task is to automaticly read out the values of the sensor. The sensor are sending some kind of push notification with the current value to the network (every 20 seconds).

Is the Listener function the correct way of getting these automatic transmissions?

greets

(sorry for my bad english)

Why not trying to poll your slave withj a command line tool like "modpoll" ?

once you'll be sure that everythning is allright with it, then switch back to Mango.

Also use the log functionality in Mango (check the Log I/O Box in the modbus Data Source) and make a tail -f on the modbusIO-1.log file

Hi,

I'm facing a problem. My slave doesn't support function 0x16 so I would prefer mango to use function 0x06 instead.
i've notice that Mango uses it when writing a integer value to a register but uses 0x16 to write a "BIT" value (The Data point is declared as a holding Register, but a simple bit si addressed)?

Does anybody knows why ?

isn't it possible to force use of function 0x06 to when writing to only one register ?

Thanks, i'm not good in java and i cannot imagine put my hand in the code ... :-(

nico

Examples :

below I try to write value 1 to bit 1 of holding register 7 into slave 3

2014/04/07-18:50:42,920 O 03160007fffd0002e23a
2014/04/07-18:50:42,965 I 0396512f9c

My slave responds with an error because it doesn't handle function 16

Below when i try to write integer value 2 to same register 7 (my locator is then declared as a 2 byte integer value)

2014/04/07-19:06:49,813 O 030600070002b828
2014/04/07-19:06:49,848 I 030600070002b828

... Then it works

Hi.

I made a software to get some values from my PLC and it is certainly ok.. but I get response times mostly of 50 ms (although somethimes they vary between 9 to 60 ms). This was my first time working with an industrial protocol and I'm not sure if this results are "the expected ones" or if I'm doing something wrong. I'll leave some images of my results using the CurrentTimeInMillis() method.

Thanks.

Attachment: download link

Hi;
For a term project I need to connect to a PLC over Modbus Serial.
I've searched and used for many different Java Modbus libraries but none of them worked. Modbus4J seems to be working. At least no error on initialization state.

Should've turned to RTU mode, my bad. It works fine now.

Adding to this discussion: the listener.start() calls into RtuSlave.start(), which immediately calls SerialSlave.start():

in SerialSlave:

@Override public void start() throws ModbusInitException { try { serialPort = SerialUtils.openSerialPort(serialParameters); transport = new StreamTransport(serialPort.getInputStream(), serialPort.getOutputStream()); transport.start("Modbus4J SerialSlave"); // <--- This line throws the exception } catch (Exception e) { throw new ModbusInitException(e); } }

The line that throws the NullPointerException is the third line, the transport.start("Modbus4J SerialSlave").

This calls into the StreamTransport class of seroUtils.jar, which we do not have the source code for.

Would it be possible for me to get this source code so that I can debug this problem?

Hi mlohbihler
I recently figured out how to write several registers in one request but I can't find how to change data types too, is it even posible?
Thank you.

Hi mlohbihler
That's my doubt. How can I write a bunch of registers at once. I was looking at WriteRegistersRequest and WriteRegistersResponse but I don't know.
Thank you.

Has something to do with RTU over TCP, I think.

Hi, I have a little doubt with setParity parameters.. I wanna know how to set non parity and I don't really see how.

Is it 0 = none, 1 = odd, 2 = even, 3 = mark, 4 = space? thx

No, not at this time but it may be made available some time in the future.

Joel.

Hello,

so far getting byte values is working but I'd like some opinions on the byte order.

Should byte 0 be the LSB or MSB?

Modbus being big-endian the first byte in the register (register is a 16-bit word) is the MSB. The second is the LSB.
But in my PLC when I configure the variables the first byte value at a given address is the LSB, the second the MSB.

To summarize:

PLC configuration
BYTE first AT %QB616
BYTE second AT %QB617

MODBUS register 308 (word value)
second | first

modbus4j
should this be
batchRead.addLocator("second", slaveId, RegisterRange.INPUT_REGISTER, 308, DataType.BYTE, (byte) 0);
batchRead.addLocator("first", slaveId, RegisterRange.INPUT_REGISTER, 308, DataType.BYTE, (byte) 1);
or
batchRead.addLocator("first", slaveId, RegisterRange.INPUT_REGISTER, 308, DataType.BYTE, (byte) 0);
batchRead.addLocator("second", slaveId, RegisterRange.INPUT_REGISTER, 308, DataType.BYTE, (byte) 1);

I would prefer the latter.

Regards
jmbsps

Hey thanks :D. And I don't feel bad 'cause that code is really strange to me. I'd never figured it out by myself.

HI !

Did someone face the same problem than me concerning modbus logs ?
I'm facing a lot of trouble with my modbus and logging would be very usefull...

Bye

Hi, I'm not bkhashfeh but I have the same problem and I'd apreciate the help :D . Here is my code where I'm reading three registers per second (or trying).

package modbusmastertest;

import com.serotonin.modbus4j.ModbusFactory;
import com.serotonin.modbus4j.ModbusMaster;
import com.serotonin.modbus4j.code.DataType;
import com.serotonin.modbus4j.ip.IpParameters;

import java.io.FileWriter;
import java.io.PrintWriter;
import java.util.Calendar;

class MBMasterTest {
public static void main(String[] args) throws Exception {
// Modbus master settings
ModbusFactory factory = new ModbusFactory();
IpParameters params = new IpParameters();
params.setHost("192.168.0.1");
params.setPort(502);
params.setEncapsulated(false);
ModbusMaster master = factory.createTcpMaster(params, false);
master.setTimeout(333);
master.setRetries(1);
// CSV writter settings
FileWriter fwriter;
fwriter = new FileWriter("C:/Users/NESTOR/Dropbox/Siemens_S7-1200.csv");
PrintWriter outputFile = new PrintWriter(fwriter);
outputFile.println("Date,Sensor_1,Sensor_2,Sensor_3"); // Primera línea
Object []Register = new Object[3];
// Date settings
Calendar cal = Calendar.getInstance();
long elapsedTime = 0;
long time = 0;

// Cuerpo del programa long start = System.currentTimeMillis(); // Tiempo de inicio try { master.init(); while(System.currentTimeMillis()<(start+10000)) { time = System.currentTimeMillis(); for(int i = 0; i<3; i++) { Register* = master.getValue(1,40001+i,DataType.TWO_BYTE_INT_SIGNED); } System.out.println(Register[0]+","+Register[1]+","+Register[2]); System.out.println("" + ""); elapsedTime = System.currentTimeMillis()-time; cal.setTimeInMillis(cal.getTime().getTime() + elapsedTime); outputFile.println(cal.getTime().toString()+","+Register[0]+","+Register[1]+","+Register[2]); // Se escribe nueva linea en archivo } outputFile.close(); // Se cierra el archivo } finally { master.destroy(); //Se destruye el objeto master } System.out.println("" + ""); System.out.println("Tardo " + (System.currentTimeMillis()-start) + "ms"); }

}

I got help and figured it out. My PLC can't read encapsulated data.. at least with my current configuration I think.. It works perfectly.

Hi,

I just want understand what's the difference of the field encapsulated of IpParameters

Here the source code

// Wrap the modbus request in a ip request. OutgoingRequestMessage ipRequest; if (ipParameters.isEncapsulated()) ipRequest = new EncapMessageRequest(request); else ipRequest = new XaMessageRequest(request, getNextTransactionId());

I work with a plc wago with the UdpMaster.

Thanks

I'm not sure too.

But notice that the test with tcp to tcp, with rtu encapsulation enabled, has the same problem even if there is no real serial device, and only tcp is involed (see the 3rd post in the thread):

run:
Master TCP started (50ms, tcp rtu -> tcp rtu) ** TCP, both master and slave encapsulated, no serial device, still has the problem

May it me something related to this: http://mango.serotoninsoftware.com/forum/posts/list/532.page ?

The only solution we can figure out (to be tested yet), is to close the TpcMaster on receiving a timeout exception, and create a new TcpMaster.
This involves issuing only single commands and not using batch.

Regards