Siemens 353, ODC, Modbus

We have configured a 353 controller with an ODC block accessible via Modbus. We can access all the readable registers and the data retrieved is correct.

The customer would like to remotely change the setpoint of the controller which is Input S. The manual indicates this register is R/W but when the customer writes to the register the value does not change.

Currently Input S of the ODC is configured to receive it's value from a loop Setpoint block. I read in the manual that if Input CL in not configured changes to loop data can be made from a network command or the local faceplate. Input CL is unconfigured for this ODC block so the implication is the loop data, namely Input S, can be changed by either means.

Suggestions?

There's something implied in the statement about CL - I suspect it is inhibiting the network write to the setpoint. But it isn't clear to me what status change is needed to get network access.

@danw You are correct, the implication is there in the manual but it is not clear enough (to me) to assure me how it functions. We were able to speak with a Siemens tech and he confirmed leaving Input CL unconfigured allows for the data to be manipulated either locally or from the network.

This application is actually linking a BACnet network to the Siemens controller via a BACnet-Modbus converter. We now suspect something in the converter's configuration needs to be tweaked. The Siemens Modbus register we are trying to write to is 32-bit and we suspect the two address construct of the data from the converter is out of sync with what the Siemens controller will accept.

If you're off by one register it would likely reject the attempt because the floating point would span two different valid registers.

I'd recommend trying a generic Modbus master, like Modscan32, ModPoll, or Simply Modbus (or any of the half dozen others) to see if the 353 takes am FC 16 write function directly, without going through the converter/gateway. If it does, then you don't have to mess with any of the things the CL affects, you can focus on the converter/gateway.

Great suggestion, using a generic master. I was onsite with the devices last week and did use Simply Modbus to troubleshoot an issue with Honeywell devices on the same network. When we couldn't write to the Input S register on the Siemens and expected the problem was with the Siemens configuration. I wish I had tried Simply Modbus on the Siemens when I had the chance.

The RTA converter being used does allow viewing the raw data so I plan to talk the BACnet tech through it to see if we glean any information.

UPDATE: Today we were able to test a 353 controller with Simply Modbus Master simulator. Bottom line, it worked, as expected. What we learned is the function code is 16, the data type must be 32bit Float, and "high word first" does not work, i.e. the low word must be first. On the latter point with high word first a value of 50.0 was represented as 4248 0000 (in hex). Disabling high word first flips the bytes to 0000 4248. This was acceptable in the range for the setpoint we had configured.

In summary, this test was to verify the configuration of the controller and that it would accept a written value to Input S of an ODC block. The final test will be configuring the RTA Automation BACnet-Modbus converter to follow this same protocol so acceptable values are delivered.

Thanks for the update.

There are 4 variant formats (word/byte order) for an IEEE 754 floating point value. Two of the four are widely used, I've never encountered the other two.

Most OPC servers and gateway converters offer the choice of one of the two dominant formats. Sometimes you have to poke around, but they're there.

The issue with eliminating 16 bits of a 32 bit FP is that the FP format does not evenly divide up into 16 bits of mantissa, 16 bits of exponent.

I believe some 2, 3, or maybe 4 digit integers will work by eliminating most of the mantissa, but beyond that things get dicey.