We have (39V) vibration transducers that send 150mV/in/sec@80hz. We've been having problems lately so I contacted the manufacturer of these 4-131-0103 sensors and found out that they recommend annual calibration testing of the sensors, which they can do. The problems I'm seeing are biased and fluctuating signals. I'm concerned about problems with the wiring and controller. I'm curious if it's possible to use a function generator to test the cabling / controls. For example connect a 0.75V RMS, 84Hz test signal at the local JB and/or AIO card, and see 6in/sec on the display. I did read the procedure that has you using the NATD and the HSAA card, but I'm hoping there's another way, Though I think we actually do have the NATD/cables hiding around here somewhere.

 

It was done (using a frequency generator to simulate seismic vibration pick-up output), but I don't think it was as good a simulation as it might have been (for what reasons I never learned). Yes; there is the NATD and cable arrangements, but if I recall correctly that doesn't test the wiring from the field devices to the input terminal board to the controllers.

I would be more interested to know when this problem started. After a maintenance outage?

Is it happening on all the seismic inputs (turbine; load gear; generator)?

There are several seismic vibration pick-ups in the Load Compartment, and it gets pretty hot in there (just about everywhere in the world!). The ones in the load tunnel (on the #2 bearing) are continually exposed to a LOT of heat.

I've also heard of pick-ups not being properly secured to the bearing caps/Load Gear boxes (by using medium thread locker AND stainless steel safety wire (sometimes the socket head cap screws get lost when assembling or disassembling and reassembling the pick-ups and ge (good enough) screws are used that don't have holes for the safety wires)--and people STILL don't put thread locker on the threads when the safety wires are not installed. A screw is a screw right? Except for certain applications.

Some machines used plates bolted to the tops of the bearing caps/Load Gear boxes to mount the seismic vibration pick-ups on. Those plates need to be FIRMLY bolted--and the bolts need thread locker (preferably the stronger thread locker) AND some kind of locking plate--NOT JUST LOCK WASHERS.

And, they do wear out over time, especially the ones in the Load Compartment/Load Tunnel.

Finally, especially in the Load Tunnel/Load Compartment, the sheathed cables used to connect the seismic pick-ups to a nearby junction box (outside the Load Compartment) get damaged, and they also are exposed to a lot of heat and vibration--especially if the cables aren't properly clipped and restrained along their length. Many people also don't take proper care when installing the cables--both when connecting them to the pick-up (using stainless steel safety wire AFTER making sure the connection is firm (but not "gorilla tight" as a former colleague used to say)) and when terminating them in the junction box. There is a shield drain wire which should be properly terminated--either to the shield drain wire of the interconnecting twisted, shielded pair or to ground in the junction box (the typical factory termination was to the interconnecting cable shield drain wire).

We don't know how old the machine is or how well it's been maintained. A periodic review of ALL terminations--not just the vibration pick-up wiring--in ALL junction boxes is sometimes very "informative" and also preventative, and also magically eliminates some intermittent, nuisance alarms and even trips. I don't recall which I/O cards the seismic vibration pick-ups eventually are connected to in the controller racks, but if this is happening to ALL the seismic vibration inputs it's kind of hard to think it's happening to all of the I/O cards in the three controller racks (I'm presuming this is a TMR Mark IV), but it COULD be the AIO.n module on the right side wall of the Mark IV control panel enclosure. (The Mark IV Speedtronic Elementary will show all of the modules/ribbon cables/I/O cards involved--it was an EXCELLENT document!!!) That AIO.n module is common to all the speed pick-up inputs (if I remember correctly).

You could try using a frequency generator, or using the NATD card & cables. I'm a little fuzzy on the math in your suggested test routine; 0.75 VAC is five times 0.150 VAC/in/sec, so wouldn't that be 5.0 inches/sec, not 6.0 in/sec? (Maybe just a typo.?.?.?) I'm not sure what the voltage level should be, either.

Anyway, best of luck. The Mark* IV was--and still is--a pretty robust turbine control system, but it isn't easy to troubleshoot or troubleshoot with, and it doesn't like walkie-talkie interference. And, it's not a true TMR control system--but it's pretty close to a proper definition of TMR for its time. The Auxiliary Display was an AWESOME troubleshooting tool, and a working alarm printer (or logger) is essential, too. (There's NOTHING WRONG with the "old school" dot matrix line printer--line printers were the correct choice for the application and ribbons are still available. The old "roller" paper gatherer was a disaster, and people simply didn't understand the printer needed to be stationary and could be jostled when the printer drawer was opened and closed. But, it was--and still is--a fine turbine control system. The Mark* IV Migration (complete gutting of everything but the <PDM> and replacement with Mark* VIe components is pretty nice--but it's such a different operator interface and there is SO much information available with the HMI and ToolboxST Trender that it overwhelms anyone who knows/knew the Mark* IV (operators AND technicians). AND, the HMI configuration and consistent lack of documentation is horrid.

Thanks for reading this far. Best of luck! Let us know how things turn out for you!

 

I would be more interested to know when this problem started. After a maintenance outage?

Is it happening on all the seismic inputs (turbine; load gear; generator)?

After reconnecting after a major inspection, I saw one sensor acting a little different than the others when still shutdown. All of the others were about 0.25 in/sec which is what I’m used to seeing. BB2 (Bearing #1, 39V-1B) was reading about 0.5, but after startup it dropped down to about 0.1 to 0.2 like the rest of them. It continued to read high when shutdown and similar to most when running (0.1 to 0.2).

Months afterward, BB5 (Bearing #2, 39V-3B) went to 0.0 in/sec while running. It went to about 0.25 when shutdown. We replaced 39V-3B. Nothing changed. We then replaced the cable, and eventually ran external cable based on what we thought we saw at the time. When we ran the external cable, we did get the signal back but then it and BB4 are reading high and spiking.

BB4 now reads around 0.5 in/sec while shutdown, along with BB2 when shutdown, and BB5 reads normal while shutdown (after reading high for about a day). Which is what I’m showing in the graph below.



This is the other unit, like above, it's before and after a shutdown.

There are several seismic vibration pick-ups in the Load Compartment, and it gets pretty hot in there (just about everywhere in the world!). The ones in the load tunnel (on the #2 bearing) are continually exposed to a LOT of heat.

The hardest to work on need the most attention. That might be a universal law.

I've also heard of pick-ups not being properly secured to the bearing caps/Load Gear boxes (by using medium thread locker AND stainless steel safety wire (sometimes the socket head cap screws get lost when assembling or disassembling and reassembling the pick-ups and ge (good enough) screws are used that don't have holes for the safety wires)--and people STILL don't put thread locker on the threads when the safety wires are not installed. A screw is a screw right? Except for certain applications.

Some machines used plates bolted to the tops of the bearing caps/Load Gear boxes to mount the seismic vibration pick-ups on. Those plates need to be FIRMLY bolted--and the bolts need thread locker (preferably the stronger thread locker) AND some kind of locking plate--NOT JUST LOCK WASHERS.

And, they do wear out over time, especially the ones in the Load Compartment/Load Tunnel.

Lots of stuff to check on here, thanks. I don’t remember ever using Loctite on the screws, I know we’ve used the lockwire.

Finally, especially in the Load Tunnel/Load Compartment, the sheathed cables used to connect the seismic pick-ups to a nearby junction box (outside the Load Compartment) get damaged, and they also are exposed to a lot of heat and vibration--especially if the cables aren't properly clipped and restrained along their length. Many people also don't take proper care when installing the cables--both when connecting them to the pick-up (using stainless steel safety wire AFTER making sure the connection is firm (but not "gorilla tight" as a former colleague used to say)) and when terminating them in the junction box. There is a shield drain wire which should be properly terminated--either to the shield drain wire of the interconnecting twisted, shielded pair or to ground in the junction box (the typical factory termination was to the interconnecting cable shield drain wire).

I recently found a shield drain wire on 39V-1B on the other unit that had a broken terminal where it connected to the interconnecting cable and it did make a big difference.

In the load tunnel, i think it’s in 1/2” conduit most of the way.

We don't know how old the machine is or how well it's been maintained. A periodic review of ALL terminations--not just the vibration pick-up wiring--in ALL junction boxes is sometimes very "informative" and also preventative, and also magically eliminates some intermittent, nuisance alarms and even trips. I don't recall which I/O cards the seismic vibration pick-ups eventually are connected to in the controller racks, but if this is happening to ALL the seismic vibration inputs it's kind of hard to think it's happening to all of the I/O cards in the three controller racks (I'm presuming this is a TMR Mark IV), but it COULD be the AIO.n module on the right side wall of the Mark IV control panel enclosure. (The Mark IV Speedtronic Elementary will show all of the modules/ribbon cables/I/O cards involved--it was an EXCELLENT document!!!) That AIO.n module is common to all the speed pick-up inputs (if I remember correctly).

I was hoping to rule this all out with the simulated signal.

It was done (using a frequency generator to simulate seismic vibration pick-up output), but I don't think it was as good a simulation as it might have been (for what reasons I never learned). Yes; there is the NATD and cable arrangements, but if I recall correctly that doesn't test the wiring from the field devices to the input terminal board to the controllers.

You could try using a frequency generator, or using the NATD card & cables. I'm a little fuzzy on the math in your suggested test routine; 0.75 VAC is five times 0.150 VAC/in/sec, so wouldn't that be 5.0 inches/sec, not 6.0 in/sec? (Maybe just a typo.?.?.?) I'm not sure what the voltage level should be, either.

Yeah, it was a typo, that's the ticket (I can't even explain how I messed that up)! I think I’m going to give it a try, it doesn’t sound like I will cause harm by trying.

Anyway, best of luck. The Mark* IV was--and still is--a pretty robust turbine control system, but it isn't easy to troubleshoot or troubleshoot with, and it doesn't like walkie-talkie interference. And, it's not a true TMR control system--but it's pretty close to a proper definition of TMR for its time. The Auxiliary Display was an AWESOME troubleshooting tool, and a working alarm printer (or logger) is essential, too. (There's NOTHING WRONG with the "old school" dot matrix line printer--line printers were the correct choice for the application and ribbons are still available. The old "roller" paper gatherer was a disaster, and people simply didn't understand the printer needed to be stationary and could be jostled when the printer drawer was opened and closed. But, it was--and still is--a fine turbine control system. The Mark* IV Migration (complete gutting of everything but the <PDM> and replacement with Mark* VIe components is pretty nice--but it's such a different operator interface and there is SO much information available with the HMI and ToolboxST Trender that it overwhelms anyone who knows/knew the Mark* IV (operators AND technicians). AND, the HMI configuration and consistent lack of documentation is horrid.

Thanks for reading this far. Best of luck! Let us know how things turn out for you!

Sounds like solid advice. Some of it I wish I had earlier on. But there are still trade-offs.

Thanks for taking the time. I will keep you updated on how this goes.

 

Where did the trend recordings come from?

What is the data capture rate of the trend recordings? (Once per second? 4 times per second? ???)

No; you won't cause any harm by simulating with the frequency generator.

Cable, even armor-sheathed cable, in a metal conduit in the Load Tunnel is going to get hot. I've seen some armor-shielded cable have a very low temperature rating (the armor shielding has nothing to do with temperature rating; it's the insulation inside the armor shielding that dictates the temperature rating). Sometimes sourcing gets the lowest cost item(s) without checking all of the specifications; it looks exactly the same from the outside, but it ain't on the inside....

Thanks for the feedback. Waiting to hear more.

 

The trend is from the DCS (MAUI>>HMI>modbus>DCS) I know the HMI has a lot of 4 times per sec. and I think these are on that list, but I think the DCS has them closer to 1 time per second. We don't have as much history on the HMI, which is why the trends are from the DCS.
I'll update this when I get there, but this little project has some lead time associated with it.