Yesterday I observed a few techs working on a pretty old (~20 years) Rosemount 1051 with old firmware. They were having a (possibly) unrelated issue where the DCS cannot read the HART digital variables after upgrading some hardware so they were checking to see if the transmitter was sending digital variables at all (4-20mA primary variable was working just fine).

A TREX was hooked up and they saw that they could not configure the HART digital variables (old firmware) but they could read sensor temp and other items that were obviously coming across via HART. To narrow down the reason for being unable to configure the variables to the communicator or transmitter, theydecided to hook up a laptop with AMS on it to the transmitter to see if theycould access those settings. As soon as the laptop was hooked up, the boiler feed pump theywere standing next to tripped (suction pressure transmitter on a running feed pump). The readout on the laptop was the expected value (~80 psig) but the DCS was now reading ~25 psig. When the laptop was disconnected from transmitter, the transmitter reading in DCS jumped from 25 psig to what the laptop was reading.

The laptop was connected using some sort of USB device with banana plugs and hook clips that were connected to the transmitter terminals (parallel to DCS).

Any ideas?

 

The DCS was reading pressure using HART? If so, it likely acts as a Primary HART master. If AMS is configured as Primary HART master, it would conflict with the DCS.

 

The DCS was reading pressure using HART? If so, it likely acts as a Primary HART master. If AMS is configured as Primary HART master, it would conflict with the DCS.

I'll have to look into that further... I am 99% certain that the DCS was reading the pressure using just the 4-20 signal from the transmitter. The DCS is configured to read other HART signals on the same cable but none of them were functioning at the time.

I do think the DCS acts as a primary HART master and the laptop caused a conflict but I don't understand why it would cause a step change in the current readout

 

1. What is a Rosemount 1051? I’ve heard of 1150’s, 1151’s, run across 2051 and 3051x but not a 1051.

2. How many HART master devices were connected at the same time?

DCS analog input reading the pressure via HART?
DCS AMS getting HART diagnostics passed through the DCS AI?
TREX?
Laptop with AMS/HART modem?

Was the TREX disconnected when the laptop was connected?

HART is supposed to support two HART masters, but not three. But even with conflicting HART masters, industrial digital communications generally performs data word and message diagnostics and 'checks' and will discard any data or message that appears to have been corrupted. I think it highly unlikely that adding a third HART communicator would alter any HART values without the corruption being detected and having the data discarded.

3. Does the DCS analog input provide loop power on the (+) terminal so that DCS(+) runs to xmtr (+), DCS (-) to xmtr (-), or are loop connections made to an external DC power supply?

 

I doubt I’ll ever get answers to my questions so I’ll go with some scenarios that might have happened.

1. Assuming the DCS AI uses the 4-20mA analog signal for its input value, then incorrect connection of the TREX might have inserted resistance into the loop circuit that altered the AI’s reported value.

A quick look at the TREX manual shows that the TREX does not use a banana plug adapter with a 250 Ohm resistor, like the Rosemount 275/375/475 or the Meriam 4150/5150/Yokogawa hand held communicators. The resistor or the adapter is used to develop sufficient HART signal; minimum loop resistance is a standard HART requirement.

Instead, TREX has 2 HART cable connections, and 3 internal resistors

​

​

Apparently, the resistors are switched in or out of the loop circuit depending which wiring connection is used AND how one answers prompt questions.

​

There’s two possibilities I can think of where putting resistance into the measurement loop could affect the DCS value.

If the internal dropping resistor in the TREX were connected in parallel across the DCS analog input’ resistance, then the I*R voltage drop that the AI uses as the ‘current value’ would be lowered (by Ohms Law, parallel resistances) by some amount, which would lower the AI’s reported pressure value from 80 PSI to 25 PSI.

If the 500 Ohm resistance were somehow added into the circuit, the total loop resistance might have exceeded the ability of the DC power supply voltage to drive current the required current through the loop.

There was a statement in the manual to the effect that the TREX issues a warning if it detects that the loop circuit is powered when using the HART + power connection. Apparently, one can ‘override’ the warning and proceed with TREX applied power. I have no idea what the impact of applying a 2nd power source does to the loop circuit.

The TREX is more versatile than its predecessors and with that versatility comes a much higher degree complexity that might have produced the altered process value seen by the OP.

 

Another explicit means of faulting the loop current - connecting TREX mA port in the loop. Page 8 of the TREX Quick Start Guide.
https://www.emerson.com/documents/a...ams-trex-device-communicator-en-us-171742.pdf