ULTRAMAT 23 Measuring Problem: 0 values are displayed at the outputs on the DCS

Having split from its previous thread it is not apparent in this thread that the problem is the analyzer's 4-20mA output operates OK when the analyzer is turned on, then fails 20 minutes later.

There is one possibility that I have encountered with output field wiring. 

If you're turning the analyzer on and the process starts at the same time, everything starts 'cold' and heats up over time. A loose electrical terminal screw could make a good electrical connection when its cold, but as it heats up, the metals expand and a loose connection creates an open circuit. The test for this situation is to substitute some other 4-20mA signal, like from a source or calibrator and if the signal maintains it integrity by giving valid readings more than 20 minutes, half an hour or an hour, then the problem is not in the wiring, the problem is on the analyzer end.

My experience

I started my controls career doing SCADA data collection with analyzers for serveral years and then subsequently moved to process instrumentation, where 2-wire loop powered circuits are the norm. 

My observation from back then is that analyzer design (amongst various vendors) seemed to focus very strongly on the process measurement itself but attention to the output circuitry neglected, the output always seemed to be just an 'add-on' to make a marketable product.  

My experience back then was that the outputs on analyzers frequently failed, even though the measurement system continued to function.  Given the cost of an analyzer, usually 4x - 100x the cost of a process field instrument like a pressure transmitter, it seemed to me to be  failure of the overall design to not provide a more robust output, like isolated serial digital outputs, choice of 2 or 3 wire 4-20mA, or isolated voltage outputs.

I saw a clear difference in dealing with 4-20mA, 2-wire, loop powered field instruments  (like pressure and temperature transmitters).  The nature of 2-wire loop powered process instrumentation drives that industry segment towards electrical isolation as a standard.  My experience shows that the output circuitry regulating the analog loop current, on the whole, is robust, as evidenced by the nature of  failures over time where it is the measurement that fails, not the output circuitry.  The output can signal a "fail-safe" condition, indicating that the output is working but that the measurement can not be trusted. 

My experience does not identify your exact problem, but my dated experience with analyzers would point toward a failing output.   

Unless your plant is already geared up for Profibus, I suspect that fixing the 4-20mA output will be far less expensive than getting a DCS to talk Profibus.