Endress & Hauser Promass 83 Ethernet IP "NAN" Input

Hello All,

we have several ethernet IP promass 83 flowmeters from Endress + Hauser in our facility that we use to measure how much material we apply to each product going down the line. I am getting a "-1.#QNAN" input from one module for all the real inputs it provides. If im not mistaken, the NAN refers to "Not a Number." According to the PLC, the network connection is fine and dandy and not throwing any faults to the PLC. I am wondering if anyone has used these flowmeters before and seen anything similar. these are around 10 years old and were moved from a previous facility and those operations were brought here. I have also compared configuration files from a unit that is working via the web browser but nothing jumped out at me.

Thanks in advance,

L

is this a new installation, or is this a new symptom on a system that has been working in the past?


What are the values when read from the 83's webserver?

These flowmeters were working in the past, but I was never in that facility or saw them working. we basically are repurposing them for a similar but new application.

latchpad said:

These flowmeters were working in the past, but I was never in that facility or saw them working. we basically are repurposing them for a similar but new application.

Click to expand...

Is this the first time you are sending the values to a Logix PLC using Ethernet/IP?

Weird ! I don't have direct experience with E&H Promass with this network and that symptom.

You are correct that "NaN" means "Not A Number". It's possible that this device intentionally puts an invalid value into the FLOAT element if there is a configuration or diagnostic problem, rather than putting in a zero or holding the last state. The "actual system condition" code = -1 is an especially strong suggestion of a configuration or diagnostic problem.

To figure out what the value is, I would use a COP instruction to move one of those FLOAT tags to a SINT[4] array, and look at that array in hexadecimal radix.

Comparing those values to values you can read from the webserver might shed some light on how they're being handled.


With a proper EDS file and automatically named module-specific UDTs there shouldn't be a byte-swapping or other data transformation issue... but there might be.

They were used on an Ethernet IP network to a logix PLC in the previous facility. it was installed by our maintenance guys before I started this project. We have an identical line with 4 flowmeters which all work fine, and I am importing the logic to this PLC to get this display system up.

Thanks Ken, Ill try the SINT copy idea and see what it brings up. I figured that actual system condition being a -1 was odd too, as we have multiple flowmeters that show 1 for this value. But when I logged on the web browser for the device, it said the module was functioning properly. Weird Indeed.

If there are three units showing valid values in the Logix and this one showing NaNs, then I would suspect the one unit and/or the connections to it. Can you swap the ethernet cables between them?


What does the E&H web server say the values are? [whoops, our posts are crossing; it sounds like all seems normal on the web server, including the values]

Ken,

I copied the mass flow value to a SINT[4] array, and each element of the array shows 16#ff in the hex view. all the element bits of each sint tag are 1.

drbitboy, im going to connect in a minute and will post what I find.

Thanks for the help!

this is what I see in the web browser

Do negative flow-rates have meaning?


Those densities sure seem low ... the mass flow, volume flow, and density values' relationships are off by a factor of 10.


If the SINT array members are all 16#ff, then any questions about byte order can be saved for later, but as Ken notes with EDS et al. that is unlikely to be the issue. Anyway, [all 1s] is consistent with a quiet NaN (cf. here).


The IP addresses seem odd; is the netmask 255.255.0.0?

the negative flow rates indicate direction in the pipe I believe, which can be configured.

the net mask is 255.255.255.0 as is everything else on that network.

It appears this system is using all the questionable leftovers from the move here. there are 2 oddball profibus meters as well that ill have the pleasure of commissioning.

latchpad said:

the net mask is 255.255.255.0 as is everything else on that network.

Click to expand...

Hmm, assuming from that statement that there is only one network, then based on the IP configuration you showed earlier:





If the netmask is 255.255.255.0 (a.k.a. /24), then the two Ethernet interfaces are on two different logical networks: 192.168.212.0/24; 192.168.77.0/24.


So if the Logix PLC is on 192.168.212.0/24, then there is no way for it to make contact via 192.168.77.0/24, at least not via the same single LAN.






What does the web page look like on the flowmeters that are working?

Solved

I made some progress and am reading the values in the PLC now. I had to change some configuration settings in the configuration of the module definition in the IO tree. One drop down setting was set to IO Data (Without config) and I changed it to IO Data (With Config). I'm not sure why this fixed it, but it seemed to. The other meters we have are set to the "Without Config" setting and work fine.

Just following up with the thread in case anyone else is having this issue and happens to find it.

Cheers,

L