SLC<->ML1500 Serial Comms

Hi Chaps,

We had to try and diagnose an issue today, with a SLC & 2x ML1500's connected via Serial Comms over a Radio.

One of the sites was not responding (2), which in turn was giving false information from the 2nd site (3).

The attached .rss file is a snippet of how the comms are done. I saw this, and thought that this is not the best way to handle this type of comms, especially over radio.

If I were to do it from scratch, I would of made each MSG instruction mutually exclusive.

Anyway, as this site has been running for 8+ years, with occasional issues if one of the sites failed, we had to get to the bottom of it.
My understanding of this code config, is that as the poll to Site 2 is first, then that MSG instruction will execute, then the 2nd instruction will be queued in the buffer, and sent when the port is available, unless of course it is possible to poll 2x nodes at the same time over a serial link?
If there is then a failure, or partial reply, it is possible that the successful poll registered as a failed poll, therefore not sending the data to the required destination??

When we first got to site today, we noticed that the Send data was working, as the remote sites were responding to commands from the master, but the data coming back was what appeared to be corrupted.

What I did, was disable comms to the site that was not responding, and the active site worked fine, without skipping a beat.

Our plan is to correct the code, so that each MSG instruction is executed exclusively (unless there is a valid reason why not ??), then diagnose the potential faulty radio at the site that is not responding, if indeed it is faulty...

Are my assumptions about what the potential issues with the buffer are, and the rectification of the polling sequence correct?

Thanks

The attached .rss file is a snippet of how the comms are done

this part seems to be missing

http://literature.rockwellautomation.com/idc/groups/literature/documents/rm/1770-rm516_-en-p.pdf this might help its the Df1 protocol and command set reference manual cannot locate the DH+ specific other than on the specific modules however this manual does cover some of it.

Haven't found the Df1 installation guide but here is the DH+
http://literature.rockwellautomation.com/idc/groups/literature/documents/um/1770-um022_-en-p.pdf

One thing to note is the Dh+ is token oassing so no two can transmit at the same time. I would checdk to see if the Df1 is also tokem passing.

mordred said:

The attached .rss file is a snippet of how the comms are done

this part seems to be missing

Click to expand...

Sorry, didn't notice & realise that .RSS files could not be uploaded. .zip file attached.

One thing to note is the Dh+ is token oassing so no two can transmit at the same time. I would checdk to see if the Df1 is also tokem passing.

Click to expand...

The channel config on the SLC is DF1 Half Duplex Master, whilst the ML1500's are DF1 Half Duplex Slave, which makes sense.
Based on this, I would assume that there is not Token passing

I think its best to only issue message instructions sequentially ( one at a time).

With what you posted you have all four going at once.

Mickey said:

I think its best to only issue message instructions sequentially ( one at a time).

With what you posted you have all four going at once.

Click to expand...

And when done this way, can cause all sorts of problems if a site/node does not reply, or fails.

It does confuse me as to why it would be done this way, would of made commissioning interesting...

I am going to re-write it anyway.

yes agreed fire them off individually. You might try using some error handling on failed messages before sending other messages.