Geoffrey Akien

In response to Richard's command (FEX didn't notify me):

1) This is a deliberate feature related to how Modbus works. If you provide a deviceAddress of 0, then this is a broadcast, and modbus devices do not respond to it. So if you wanted to broadcast then then (~~0) would evaluate as true and this section (encapsulating all of the code to do with receiving data) would be run. This does not impact the sending of commands because this is run before the if ~deviceAddress section. I agree that on reflection using a double negative is not the most readable way to have done it, but it does work :).

2. I think what you've done to get round that problem is definitely one way to do it. If all the devices on that two-wire bus are using modbus then you could filter out only the responses from that particular deviceAddress. The whole deviceAddress thing is designed into Modbus exactly for multidrop instruments on the same bus. More info here: https://wiki.univie.ac.at/download/attachments/7503909/2kcomms.pdf

I didn't test it on RS422 so you may have to edit the code to make it work for you.

The message is for your information, and not an error. Modbus allows you to send more than one command per communication, and different Eurotherm devices have different maximum amounts of commands. When it connects tempobjconnect tries to ask what device it is so it knows what the maximum is for future communications (it has a list of known devices). In your case it is either an unknown device, or it didn't respond at all to the request.

In response to Richard's command (FEX didn't notify me):

1) This is a deliberate feature related to how Modbus works. If you provide a deviceAddress of 0, then this is a broadcast, and modbus devices do not respond to it. So if you wanted to broadcast then then (~~0) would evaluate as true and this section (encapsulating all of the code to do with receiving data) would be run. This does not impact the sending of commands because this is run before the if ~deviceAddress section. I agree that on reflection using a double negative is not the most readable way to have done it, but it does work :).

2. I think what you've done to get round that problem is definitely one way to do it. If all the devices on that two-wire bus are using modbus then you could filter out only the responses from that particular deviceAddress. The whole deviceAddress thing is designed into Modbus exactly for multidrop instruments on the same bus. More info here: https://wiki.univie.ac.at/download/attachments/7503909/2kcomms.pdf

I have this code working with a Eurotherm 3216 over two-wire RS-485. Two things needed to be fixed. Neither has been rigorously debugged but it should at least be a helpful start.

1) The NOT operator (~) needs to be removed from the line "if ~deviceAddress" in tempobjcomm.m". I don't really see how anything could have worked in its original state ...

2) As it is a two-wire bus, Matlab receives all outgoing commands. I circumvented that by adding the line " "fread(serialObject, length(command));" immediately after the line described in 1). There may be a way to do this in the serial-port settings but this way seems to work fine.

I'm using the RS422. But it seems like the program is not working for RS422.

when I type "tempobjconnect(serialObject)" in the command window.

It shows this:
"Warning: Could not retrieve Eurotherm controller type - read and write commands will be limited to the minimum allowed for series 2000 controllers for this serialObject."

Thank you sir!!