Mike,

Thank you very much for the reply, but before we start to blame my SCADA system, can we examine a few other items first.  I don't claim to be a Modbus protocol expert, but I have interfaced a few modus devices in the past and my SCADA system currently interfaces to many devices and different companies that adhere to the Modbus protocol standard without issue.

An observation:  There aren't any errors detected in the message format itself.  Note, that a master request goes out and the T3610 slave answers immediately (less than 1 second) with a valid response.  I conclude that the content for those registers are incorrect. 

So here are a few questions to make sure we are on the same page.

1.  Are the addresses 40049 Temp, 40050 Relative Humidity correct?

2.  What is the delay required for a Comet Device ?   Note that I am currently waiting 15 seconds. however the results are the same whether I wait 1 second or 15 seconds.

3.  If I am able to show you the message received from the T3610 is correct in format, but the message content is as I have described alternating between a correct temp and RH reading and default values of 0 and 75, what will that indicate to you?  What would your next suggestion be?

GBrown

Jan,

Great, now we are getting somewhere, I had put those addresses in my first post. 

1.  Yes, I looked at page 26, and saw the address 49, 50, etc.  but thought that valid Modbus read registers must start with a 4XXX so I added the prefix.  Note that when I use address 49 and 50, the data is not valid when I read it, but when I use 40049 I do get the same value as the web page reports.  My Modbus interpretation that all addresses that start with 00XXX to actually be coil write commands, not a read.

So I will do two things.

A.  I will use a 3rd party ModbusTCP driver independent of my SCADA system to poll the T3610 and see the results.  I certainly believe that your Modbus software will work on your device. 

B.  I will post to the Modbus community site and get their interpretation of the 49 and 50 address versus 40059, 40050. 

C.  If I don't have clarity from actions A & B, I will use wireshark.  My network team, does not like wireshark being used so I tend to use it as a last resort.


2.   I didn't think the timing was an issue, so it is good to take that issue off the table.

3.  See response C.

GBrown

Jan,

Hmmm, I must have missed that memo.  I just looked around and can't seem to find reference to Modbus 4XXX registers as being obsolete or communicating to them in that manner as being obsolete.   

Can you point me to where this is described? 

GBrown

Jan,

I spent some time today digging into my actual problem with some research about the Modbus versions that you reference.

1.  It turns out that my problem of alternating values had nothing to do with Modbus implementations being obsolete or not.  If you think about it, if a value changes, it can only be if something is writing a change.  I had an undocumented tag writing to the same location at about the same frequency.  So, I changed the Tag location and it easily solved the problem.

2.  I really didn't know what to make of your obsolete Modbus claims, so I did a bit of research on that.  I looked at a few Modbus simulation tools and asked a couple of experts that I know.  Here is what they said.

"You'll perhaps get a lot of answers, but true Modbus numbers registers from 0 to 65535, and coils also are 0 to 65535.

The 4x/0X stuff is a notation used by early Modicon programming tool, where 0 was reserved for none/nop and the address was strictly 0001 to 9999. So 4x0001 was register offset 0, and 0x00001 would be coil offset 0. Eventually, when 64K memory became cheaper than 4x, they bumped this up to 5 digits, but it is still a Modicon notation. You'll find, as example, that SE now use the %MW0 for memory-word instead.

My suggestion to vendors is always just include two columns in the document, so your two registers would be 49 or 4x00050, plus 50 or 4x00051. This makes everyone happy."

I actually printed the protocol documents which was actually good timing as I will be teaching a seminar on this soon.  I didn't see anything in the documents indicating obsolete or bad practice to use 4XXXX registers (Maybe you can point out this page).  In either case, there is lots of both new and old Modbus equipment that uses the 4XXXX convention, so as the expert has said, I don't see why a vendor wouldn't automatically include both, it is really an easy thing to do.

So at the end of the day, my SCADA system was able to talk to your sensor using 40049, 40050 registers.  So I am curious now, does your system actually support 4XXXX registers as every other system I have ever seen does. or is my SCADA system so smart, it could figure out what to do even though I had used an obsolete address format?

GBrown