False Spikes in Seismic Vibrations

Currently we are facing serious issue in our control system. Secondary air fan is equipped with seismic vibration in each bearing. The problem is whenever we start work in field in any transmitter which is not related with this vibration, we observed spikes in vibration probes in our control system. The common thing between any transmitter and this vibration probe is only DCS, and all other thing have no any relation like JB, cable and as well as far from each other. Please advice what is the possible reason of this spikes.

Urgent response will be highly appreciated.
 
W
I would want to ask the following questions, and I am sure that the bright people on this formal might be able to help you better troubleshoot your problem if they knew more about the problem.

1. What kind of signals do your field transmitters and vibration probe put out and send to the DCS?

2. Is any of this wireless?

3. What kind of cabling are you using, e.g. shielded twisted pair?

4. Are your shield and instruments grounded properly? Are the shields grounded at only one end? Do your probe shields share anything with each other or other instruments? Improper grounding is probably the most common issue that causes strange problems

5. Do the field transmitter’s signal and the vibration probe (box in between for vibration probes?) go to the same DCS box?

6. Is this a new problem? On an existing system? New System?

7. Does it happen on all your vibration probes? With the same magnitude and frequency no matter which instrument you work on?

8. Can you affect the amplitude or frequency of the spikes by what you are doing in the field? Is the amplitude large enough if it was a true signal you would feel it?

9. Are the spikes the same amplitude and frequency on all your problems?

10. Are all the field instruments doing this or a group of them?

11. What exacting are doing to the field instruments to initiate this problem?

12. Does disconnecting your vibration probe cable shield change anything for that probe?

13. Are any of the field instruments and probe signals routed be the floor in your equipment room?

14. Have you physically traced the probe wires to see where they go?

15. Did this start after a thunderstorm or other large storm?

16. Are the filed instruments in a building or in the open like in a refinery?

17. Does it do it in both night and day? How about with different technician teams?

18. Have you ever seen these kinds of spikes before? If so, what caused them?

19. What have you done to troubleshoot the problem?

William (Bill) L. Mostia, Jr. PE
ISA Fellow, FS Engineer (TUV Rheinland)
Winner of the 2018 ISA Raymond D. Molloy Award
Sr. Safety Consultant
SIS SILverstone, LLC
http://sissilverstone.com
 
>1. What kind of signals do your field transmitters and
>vibration probe put out and send to the DCS?
Pressure, temperature, level, flow 4-20mA and radial vibration through proximities probe direct on DCS without BN3500 plus seismic vibration direct to DCS

>2. Is any of this wireless?
No wireless

>3. What kind of cabling are you using, e.g. shielded twisted pair?
Shielded

>4. Are your shield and instruments grounded properly? Are
>the shields grounded at only one end? Do your probe shields
>share anything with each other or other instruments?
>Improper grounding is probably the most common issue that
>causes strange problems
shield ground at DCS eng through clean earth...probe only two wire output however multi core cable to DCS include shield cable which is connected in DCS...we have two types of grounding clean and dirty...all instrument shield connected through clear earth and panels and structures connected through dirty earth..

>5. Do the field transmitter’s signal and the vibration
>probe (box in between for vibration probes?) go to the same
>DCS box?
There are Junction boxes in field which is different whereas in DCS there are three panels for field JB.

>6. Is this a new problem? On an existing system? New System?
Yes new problem on an existing system

>7. Does it happen on all your vibration probes? With the
>same magnitude and frequency no matter which instrument you work on?
Only seismic type mean eddy probe type

>8. Can you affect the amplitude or frequency of the spikes
>by what you are doing in the field? Is the amplitude large
>enough if it was a true signal you would feel it?
Amplitude is large and sometime goes beyond 20mA which shows bad in DCS and sometime within range...

>9. Are the spikes the same amplitude and frequency on all
>your problems?
Amplitude of each probe is different

>10. Are all the field instruments doing this or a group of
>them?
all eddy current probes only

>11. What exacting are doing to the field instruments to initiate this problem?
replacement of any transmitter..open wire due to replacement

>12. Does disconnecting your vibration probe cable shield
>change anything for that probe?
Never disconnecting shield...

>13. Are any of the field instruments and probe signals
>routed be the floor in your equipment room?
Routed through cable tray

>14. Have you physically traced the probe wires to see where they go?
To JB direct and from JB to DCS through cable tray

>15. Did this start after a thunderstorm or other large storm?
It’s difficult when it’s start but long time ago..

>16. Are the filed instruments in a building or in the open like in a refinery?
Open

>17. Does it do it in both night and day? How about with
>different technician teams?
Both night and day observed with different member of teams

>18. Have you ever seen these kinds of spikes before? If
>so, what caused them?
Never

>19. What have you done to troubleshoot the problem?
Searching for solution
 
To summarize the situation as reported: (If incorrect, please correct.)

There is a Bently Nevada 3500 vibration system. The Bently Nevada system runs OK. It is not part of the problem. There are additional seismic eddy current probes that are powered by and connected to the DCS, but are not connected to the Bently Nevada system.

When a 2 wire loop powered field transmitter (xmtr) is disconnected from its DCS wiring creating an open circuit for that xmtr, the indicated output of one or more of the seismic eddy current probes spikes, which is sometimes an on-scale offset and sometimes offscale offset.

This has been going on for quite some time. No one remembers or has researched when it started.

Additional Questions

1. Are the spikes the same time duration of the DCS field transmitter (xmtr) open circuit? If the DCS xmtr open circuit is 5 seconds, is the eddy current spike seconds in duration? If the DCS xmtr open circuit is 2 minutes, is the eddy current spike 2 minutes?

2. Is the fault repeatable? If one opens PTxxx circuit, does the same eddy current probe spike?

3. Can the faults be mapped?
PTxxx faults eddy current probe YYY
TTxxx faults eddy current probe ZZZ

4. Do the seismic eddy current probes share the same DC power supply as the DCS xmtrs?

5. Do the seismic eddy current probes share the same Analog input cards as the DCS xmtrs? Same rack? Same grounding strip?

6. Do these seismic probes require bipolar or 'negative' DC power?

Speculation:

Disconnecting a mA DC current has little potential for inducing a transient spike, even in non-twisted pair cabling (could a DCS have been installed without twisted pair?) so I'm hoping the 'spike' is an offset of some longer duration - not a transient.

If it is an offset, then offsets have the "smell" of a common mode problem.

Non-destructive common mode problems typically manifest as an indicated offset of varying degrees from an on-scale offset up to an offscale (either direction because common polarity varies on the location) offset. Destructive common mode can fry the AI card due to excessive voltage.

It appears to me that the common factors between the DCS field xmtrs and the non-Bently Nevada seismic probes are
- DCS power supply
- analog input cards

But. I've never heard of removing a live signal creating a common mode problem.

Common mode problems typically sound like,
"I added a signal to the existing 3 analog inputs and as soon as I did, all the signals went off-scale. What's wrong?"

Adding a device with grounded point to an analog input can create an excessive common mode problem. So how does removing the typical isolted, non-grounded 2 wire loop powered field transmitter create a common mode problem?

Last question - what brand/model are these seismic eddy current probes?
 
>There is a Bently Nevada 3500 vibration system. The Bently
>Nevada system runs OK. It is not part of the problem.
There is no Bently Nevada system installed for vibration. 4-20mA type proximate used which is directly connected to DCS analog card.

>There are additional seismic eddy current probes that are
>powered by and connected to the DCS, but are not connected
>to the Bently Nevada system.
These eddy current type probes are loop powered from DCS which r directly connected to DCS analog input card.

>When a 2 wire loop powered field transmitter (xmtr) is
>disconnected from its DCS wiring creating an open circuit
>for that xmtr, the indicated output of one or more of the
>seismic eddy current probes spikes, which is sometimes an
>on-scale offset and sometimes offscale offset.
Transmitter disconnected from field mean from transmitter end and other end always connected with DCS yes correct sometimes spikes on-scale offset and sometime offscale.

>This has been going on for quite some time. No one
>remembers or has researched when it started.
Correct difficult to answer that which activities is the cause of this problem

>Additional Questions

>1. Are the spikes the same time duration of the DCS field
>transmitter (xmtr) open circuit? If the DCS xmtr open
>circuit is 5 seconds, is the eddy current spike seconds in
>duration? If the DCS xmtr open circuit is 2 minutes, is the
>eddy current spike 2 minutes?
Duration of spikes is normally one second up and one second down so total duration observed approximately 2 seconds whereas these spikes observed either opening of wire or termination of wire in field transmitter

>2. Is the fault repeatable? If one opens PTxxx circuit,
>does the same eddy current probe spike?
Sometime with different transmitters and sometime with different transmitter yes repeatable

>3. Can the faults be mapped?
>PTxxx faults eddy current probe YYY
>TTxxx faults eddy current probe ZZZ
No all eddy current probes observed spikes with any transmitter

>4. Do the seismic eddy current probes share the same DC
>power supply as the DCS
Yes Exactly same power supply because these r loop power

>5. Do the seismic eddy current probes share the same Analog
>input cards as the DCS xmtrs? Same rack? Same grounding
>strip?
Analog card and panel may be different yes but controller is grounding strip
Is same

>6. Do these seismic probes require bipolar or 'negative' DC
>power?
Please clarify this question

>Last question - what brand/model are these seismic eddy
>current probes?
Monitran MTN 1185
 
W
Nothing stands out. The cause of your problem must come through the air, through the wires, or through the ground and there must be something that is common to both systems that causes the coupling mechanism between the systems. There is a troubleshooting technique that I developed, which I wrote about in that book I wrote on troubleshooting (Troubleshooting: A Technician's Guide, Second Edition published by ISA). I call it “circling the wagons,” where you draw an imaginary “circle” around your instrument loop and then look at all things that cross that circle boundary to see what could cause the problem and then remove the connection and see if it solves your problem.

There is another troubleshooting method that you might try. This is to remove the instruments in a system and then add them back into the system until the problem reoccurs. This situation in not exactly applicable but I would also be curious to know if you disconnect all the fan vibration probes at the DCS and connect one spare probe in the DCS equipment room only, does the DCS still show the spikes in the DCS. Also, you might see if adding one field vibration probe back at a time affect the spikes. If adding one back, causes the spikes to return, then disconnect that probe and connect up one of the other probes in turn and see if they all have the spike problem (here we are looking for a connection point to the probes).

William (Bill) L. Mostia, Jr. PE
ISA Fellow, FS Engineer (TUV Rheinland)
Winner of the 2018 ISA Raymond D. Molloy Award
Sr. Safety Consultant
SIS SILverstone, LLC
http://sissilverstone.com
 
>>Last question - what brand/model are these seismic eddy current probes?
>Monitran MTN 1185

The link below shows the wiring diagram for a Monitran 1185, taken from a spec sheet on the web:

https://i.postimg.cc/28R3xtQ6/MTN-1185-wiring-diagram-0-V-common-is-labeled-screen.jpg

It's 3 wire, 4-20mA device, which uses a common for the 0Vdc and 4-20mA (-).

Note that the 3rd wire on the diagram is labeled "shield", which I assume means the shield/screen around the 2 signal conductors.

The drawing could be interpreted as using 2 conductor twisted pair cable with a shield (screen) and using the shield (screen) as the 3rd conductor.

Is that how your system is wired?

Or do you have 3 copper conductors for the signal and a separate shield/screen?
 
You are absolutely correct, these r three wires probe in which one is shield and all instruments shielding connected on DCS instead of field. signal cable is multipair cable which include twisted pair of two core and one wire is shield wire. There r individual shield with each pair.

>The link below shows the wiring diagram for a Monitran 1185,
>taken from a spec sheet on the web:
>
>https://i.postimg.cc/28R3xtQ6/MTN-1185-wiring-diagram-0-V-common-is-labeled-screen.jpg

----snip---

>The drawing could be interpreted as using 2 conductor
>twisted pair cable with a shield (screen) and using the
>shield (screen) as the 3rd conductor.
>
>Is that how your system is wired?
 
>The drawing could be interpreted as using 2 conductor
>twisted pair cable with a shield (screen) and using the
>shield (screen) as the 3rd conductor.

We have twisted pair with shield, in multicore cable each pair include 2 core for signal and one is shield.
 
Currently there was startup of plant in which once again I observed spikes in vibration. This time the cause of spikes was not the removal of core from transmitter. This time operation complain that when they rack-in one of the blower breaker. Spikes comes suddenly in all vibration probes of fan; these all are seismic (eddy current type) whereas other probes (radial) remain intact with no vibration. Problem increasing day by day but not find any useful solution...
 
I'm leaning towards the idea that using the shield drain wire as a third conductor is not such a good idea; twisted, shielded three-conductor cables would probably be best, especially if the cables are exposed to high-current and/or high voltage wiring/cabling in cable trays, cable trenches or conduits.

Finally, I don't think this has been mentioned before, but do all of these problematic seismic pick-ups connect to the same input card of the DCS? If so, could there be a problem or an issue with the input card, or the slot in the I/O rack where the input card is mounted? I'm NOT a fan of swapping cards for troubleshooting, but sometimes it's the next positive step forward (though it can cause the replacement card to fail if the problem is NOT the card; best to make sure the field wiring and connections are all good before swapping cards).

And, I encourage the idea of disconnecting all of the seismic inputs and reconnecting them one at a time to see if any one is or could be causing a problem for all the others. As was suggested, connect one, if it's not an issue, disconnect that one and connect another, and so on, to eliminate individual pick-ups as the cause of the problem.

Finally, where is the power supply for the pick-ups coming from? The DCS input card, or a separate power supply?

If you're really at a loss for ideas and the problem is getting worse, you might want to contact the DCS manufacturer, and/or the seismic probe manufacturer, or the control system integrator that configured and commissioned the DCS for assistance. We can't see the drawings (schematics) for the equipment and field wiring and having someone on site with a "second" set of eyes might be best.
 
The dated appearance of the first wiring diagram I found (on a distributor web site) spurred me to check Monitran's factory web site.

Monitran's factory site has what is presumably the current model 1185 spec sheet which shows a conventional 2 wire loop powered 4-20mA accelerometer, not the 3 wire device whose power supply (-) and signal (-) is likely to be grounded through the use of the shield as a conductor.

http://www.monitran.com/Portals/0/datasheets/1185.pdf

The 2 wire model is conventional 2 wire loop powered wiring. The spec sheet has a wiring diagram at the bottom of page 2 where the mA meter symbol represents the analog input load (or it could be in the side connected to the power supply return).
https://i.postimg.cc/TwmzVMxg/2-wire-version-of-MTN1185.jpg

My advice: replace the 3 wire velocity transmitters with their equivalent 2 wire versions. You can use the shield/screen for its intended purpose and likely eliminate the common mode ground loop that is causing your false offsets.
 
> Your problem sounds to me like an elevated ground potential or floating.

Or

Mixed potential sources that don't share a common reference and as a result.
 
>And, I encourage the idea of disconnecting all of the
>seismic inputs and reconnecting them one at a time to see if
>any one is or could be causing a problem for all the others.
>As was suggested, connect one, if it's not an issue,
>disconnect that one and connect another, and so on, to
>eliminate individual pick-ups as the cause of the problem.

May be we can check by this method but plant is operational, therefore currently not possible.

>Finally, where is the power supply for the pick-ups coming
>from? The DCS input card, or a separate power supply?

24VDC loop power coming from analog input card.
 
Currently plant is operational and we can check this during turn around.

>There is another troubleshooting method that you might try.
>This is to remove the instruments in a system and then add
>them back into the system until the problem reoccurs. This
>situation in not exactly applicable but I would also be
>curious to know if you disconnect all the fan vibration
>probes at the DCS and connect one spare probe in the DCS
>equipment room only, does the DCS still show the spikes in
>the DCS. Also, you might see if adding one field vibration
>probe back at a time affect the spikes. If adding one back,
>causes the spikes to return, then disconnect that probe and
>connect up one of the other probes in turn and see if they
>all have the spike problem (here we are looking for a
>connection point to the probes).
 
From your response we understand that we have to removed third wire (shield) from JB, and we have to connect -ve of signal cable to ground? Or -ve of power supply to the ground?
 
>> Your problem sounds to me like an elevated ground potential or floating.
>
>Or
>
>Mixed potential sources that don't share a common reference
>and as a result.

May be chance that reference is not same..

I have one query. We have two types of ground in our plant, one is clean & other is dirty. Clean is for shield cables, i.e., for signals. Whereas dirty is for structure, panels, equipments. Please advice what u think these all ground should be connected with each other or should be separate?
 
I've been pondering this over the past couple weeks. I still think that the installed 3-wire configuration is producing a ground loop.

Both the installed 3-wire wiring device and the mfg's current 2-wire spec sheet use the same model number. I think it strange that an updated model would keep the exact same model number, but what if it is not an updated model but the same device?

The mfg's current 2-wire loop powered wiring diagram (link in previous post) shows an empty 3rd terminal (b). I wonder if the existing device could actually run as a 2 wire, loop powered field device but is wired as a 3-wire. If it can run as a 2-wire device, I suspect that the problems would disappear.
 
Top