1791d Analog Expansion (again) + Hart

GREETINGS;
DOES ANYONE HAVE EXPERIENCE WITH BRINGING IN A 4-20 ma TRANSMITTER, THAT HAPPENS TO ALSO HAVE "HART" PROTOCOL INTO A 1791D-N4CV2X ANALOG EXPANSION MODULE? WE ARE USING THIS EXPANSION MODULE (WITH DEVICENET) AND HAVE A FOXBOR. DP TRANSMITTER TYED TO ONE A.I.

PROBLEM IS, THE SIGNAL IS ALL OVER THE PLACE, (SOMETIMES IN THE NEGATIVES, SOMETIMES NOT!). THIS IS WITH A "ZERO" INPUT VARIABLE TO THE TRANSMITTER.

HOWEVER, IF I SIMULATE THE INPUT, NO PROBLEM, THE AI IS STABLE AS A ROCK. LIKEWISE, WHEN I CHECK THE OUTPUT OF THE TRANSMITTER TO THE AI, ITS STABLE AS A ROCK.

THE ONLY THING WE CAN COME UP WITH, IS THAT THE "HART" FREQUENCY IS HAVING AN UNDESIRED EFFECT ON THE INPUT CHANNEL.

ON OTHER CONTROL SYSTEM CARDS, WE DO NOT SEE THIS PROBLEM, BUT THIS IS THE FIRST N4CV2X WE HAVE EVER TRIED. (YOU CANNOT TURN THE "HART" "OFF" ON THE TRANSMITTER)

I TRIED THE AB TECH SUPPORT AND EVEN HAVE A CASE NUMBER, BUT NO LUCK SO FAR. THIS MODULE IS NEW TO THEM AS WELL, ESPECIALLY ONE WITH 4-20 AND "HART" COMING IN.
ANY HELP WOULD BE APPRECIATED.
THANKS IN ADVANCE.

Generally the Hart communications on the 4-20 line is "transparent" to the analog signal receiver. The problem you describe sounds more like electrical noise to me. Check your signal wire shielding, make sure it is grounded at one end only. Look at the options on your analog input card, and make sure you have enabled filtering on the signal. You can also try putting a fat capacitor on the 4-20 mA, or see if you can go into the transmitter and suppress the Hart communications.

Thanks, Tom.
Youre absolutley correct about the wiring and shield grounding.

However, we already tried this with no effect. (the TX is only about 10 feet from the AI module, and the single pair goes to the TX from the JB directly.

We are grounding the shield to common ground in the JB ONLY, and cutting and taping the other (TX) end.

I also agree that the HART is normally transparent on EVERYTHING I have ever worked with in the past, so I really didnt think about it on this job. But, it sure looks like it is narrowing down to this.

As far as filters, the N4CV has 50hz and 60hz available, but neither does any good on this noise signal.

Also, we cannot turn the HART off on the TX side, but good point.

I thought about the capacitor, possibly a .1 or .5 microfarad? What would you recommend? (It wouldnt hurt to try this).
Thanks for the input.

Yes, the HART protocol is a screwed-up mess that can cause all kinds of problems. The worst is that you can't reset the transmitter parameters without an expensive communicator. Try putting a 250 ohm resistor in series with the output. If that doesn't help, go buy you a non-HART transmitter.

I don't work with HART much, but it makes sense that a PLC-style analog input is designed to handle much faster-changing analog signals than a DCS-style analog input, and might be subject to interference from the HART digital signal superimposed on the 4-20 mA loop.

Endress + Hauser just brought out a HART-equipped 1797 FLEX EX analog module that might be a more appropriate input type, and Spectrum Controls has ControlLogix HART-equipped input modules.

I'm not sure how to filter this out... HART uses a 1200 bits per second modulated signal at 950 and 2500 Hz for the Hi/Low bit levels.

More about HART:

http://www.romilly.co.uk/

Maybe an isolator might help. Such as those available from Moore Industries.

Mickey

Thanks, Mickey and Eddie for your replies.

An isolator was considered, but we will probably just use the TX somewhere eles and buy (converntional) one for this particular input.

And Eddie, your correct in the fact that the analog slaves and TU's for a process system are generally "slow", or percieved as such. So the PLC input is possibly just too darn fast.

I have been talking to tech support at Rockwell, and they are very interested in figuring this out, as this is a "new" module and maybe they havent tested with HART(?).

They are aware (now) that HART trasnmitters are still very popular, and just about every major (Tx) manufacturer has a HART version.

I am still going to try a capacitor on the input, just to see if I can filter this out. I will also update this tommorrow, as to whether it works or not, FYI.
Thanks again.
tj