DCS Analog Input Loops - Blown Fuses

Hi. I hope everybody is doing well. I have been experiencing a strange phenomenon at my site which involves the fuses on several DCS analog input loops blowing from time to time. We had a few isolated incidents here and there when we first commissioned the system but the issue has become more frequent over the past few days.

Before I start detailing what the issue is, what I have done so far and what I am planning on doing, here are some details about the system.

DCS - Honeywell Experion PKS
I/O Family - Series C, Universal I/O (UIO-2)
Fuse Rating - 50 mA, Quick-Blow or Fast-Acting

The Problem:
The issue normally involves analog inputs whereby a particular transmitter becomes NaN on DCS out of the blue. Technicians find the transmitter in question powered off in the field i.e. loop power becomes unavailable. Upon checking, they are greeted with a blown fuse so they go ahead and replace it to resolve the issue. They observed only 2 or 3 failures of this nature in a span of 11 months so they did not feel the need to escalate the matter to me. However, we had three such failures in the past 4 days, one of which involved a safety critical device so I have been asked to check things out. It must be noted that all the transmitters that have failed so far have been loop powered in nature and belong to separate multicores/trunks and I/O cards so this is not a common-cause failure. The only thing common between these transmitters is that they are all Fuji FCX series differential pressure or pressure transmitters.

What I Have Done So Far:
I had a feeling some of the shield wires were not properly terminated so I had it verified. The shield wires are properly connected on one end only. Similarly, the transmitters themselves seem okay based on tests we carried out in our workshop with a tested power supply and measurement setup.

What I Plan On Doing:
* Check the AC and DC voltage on each individual core of the loop with respect to ground to see if there are any stray voltages / ground loops.
* Check the health of the instrumentation ground.
* Check the insulation resistance of the loops in question.

Another thing that I am suspecting is perhaps that the fuses we are using here are rated too conservatively (50 mA for an operating range of 4-20 mA). The analog inputs are all loop powered in our case and use a non-incendive power supply. We generally use 250 mA to 1 A fuses for analog inputs at our site but the problem with the higher rated fuses is that sometimes the channel ends up getting compromised before the fuse gets a chance to blow. So, for this particular installation, we decided to go with a much smaller value to ensure channels are not compromised in the event of a short. If the tests I have outlined above don't yield any fruitful results, I am planning on replacing these 50 mA fuses with 250 mA ones.

So, at this point, I am suspecting three things,
* Ground Loops
* Stray/Induced Voltages
* Fuse Rating Too Conservative

Please let me know if this is the right approach and if you have any more ideas as to what could be wrong here and what else needs to be checked, please feel free to share. Thanks in advance.
 
The fundamental difficulties are
- the infrequency of failure
- not understanding the cause

I would be inclined NOT to uprate the fuse. Despite being fast acting, at some point the 20ma loop current has been grossly exceeded for a short finite time. They are the loop‘s best protection.

My thinking is more weather related - any electric storms / thunder / lightning co-inciding With failure ?

Failing that I would go back to basics - pick one loop out (that has failed in the past) and do a ‘sanity’ check.
 
The fundamental difficulties are
- the infrequency of failure
- not understanding the cause

I would be inclined NOT to uprate the fuse. Despite being fast acting, at some point the 20ma loop current has been grossly exceeded for a short finite time. They are the loop‘s best protection.

My thinking is more weather related - any electric storms / thunder / lightning co-inciding With failure ?

Failing that I would go back to basics - pick one loop out (that has failed in the past) and do a ‘sanity’ check.
I have an update. I checked the DC and AC components on each loop with reference to ground. The DC components in each case were in the order of 0.006 V while the AC components were 0.1 V so I guess we can go ahead and eliminate the presence of ground loops as a possibility.

So now it comes down to the following two theories,
* Radio Frequency Interference - Although the transmitters we are using were tested as per EN 61326: 1997 and should be good between 80 to 1000 MHz at a field strength of 10 V/m, I wonder if it is somehow getting coupled into the cables. But then again, the cables in question use twisted pairs and are shielded and the only RF transmitter big enough to be of any concern let alone cause a problem is the Motorola walkie talkie base station which operates in the VHF region... so not too solid on that one either. The only reason I am even considering this is because radiated noise from said walkie talkies has interfered with some of our gas detectors and flow meters in the past.
* The analog input channels each have an in-built short circuit protection mechanism (PTC re-settable fuses) which is set to trigger at 25 mA while the external fuses are rated at 50 mA which begs the question of why the higher rated fuse is blowing earlier. Both are supposed to be fast acting but perhaps the 50 mA one is more sensitive or maybe it is a low-quality fuse that is blowing erroneously. Wish Honeywell provided more information about these PTC re-settable fuses.

At this point, I am just spitballing. The OEM says we don't need to fuse our analog inputs so my manager is of the opinion I should get rid of them entirely. He believes we may be looking for a fault that might not exist in reality.
 
Interesting update ....

I would go for the 50ma fuse blowing before re-settable fuse tripping detecting over current in the loop not at the analogue input. Mentioned weather above but would suggest static as a possibility. Static voltages could be suppressed at the analogue i/p terminals but sufficient energy to remove the fuse.

not easy to diagnose; I still have an old Philips analogue multimeter which was prone to static voltages so we know they are out there.
 
"The analog inputs are all loop powered in our case and use a non-incendive power supply."
Sounds you are getting ground loop surges, from the field (motor starts, etc.).

The field grounding, can really show extreme voltages relative to the control room/DCS grounding.
We were seeing several hundred volts between various grounds at the time.

Analog wiring should only be grounded in the control room, unless you have analog isolators.
 
Top