Control Transformer

Status
Not open for further replies.

travish

Member
Location
North Carolina
I have a machine at work that is having noise issues. one of the manufactures of a motion controller suggested the possibility of ungrounding the control transformer. Is there a reference in the code that says that 120vac control transformers in a industrial machine has to be grounded. 250.20(B)1 is the only thing I can find and my voltage is less than 150v to ground anyway, it also says "that supply premise wiring systems". I don't think this machine is a"premises"

Any help?

Travis
 

tom baker

First Chief Moderator
Staff member
OK if you unground the CPT, if there is a line to case fault, the case becomes energized and the fault won't clear.
I suspect its not a grounding issue but a mfg issue with their equipment. The next thing they may ask for is a ground rod at the machine.

And if you don't accept the machine is premise wiring, take a look at UL508 or NFPA 79 they cover your application.
 

don_resqcapt19

Moderator
Staff member
Location
Illinois
Occupation
retired electrician
Almost everything that is not a part of the machine itself is premises wiring. See Article 100. The control transfomer is a SDS and removing the grounding connection is not likely to make any difference in the operation of the equipment. You need to find the real cause of the problem.
 

travish

Member
Location
North Carolina
I would think the last paragraph uner the definition for premises wiring excludes the control cabinet from premises wiring

it is still early here, what does SDS stand for?
 

travish

Member
Location
North Carolina
NFPA 79 8.3 "control circuits shall be permitted to be grounded or ungrounded"

I don't see any reference to a voltage level that the control circuit should be less than.
 

wirenut1980

Senior Member
Location
Plainfield, IN
What are the symptoms you are experiencing that makes you think the cause is noise? Most of the time when I am troubleshooting control power related problems, the cause is voltage sags taking out a relay or PLC.

Does the machine shut down or display an error, does anything else happen at the same time? How often does it happen?

I would not unground the control power for the reason mentioned by Tom.
 

ptonsparky

Senior Member
Location
NE (9.06 miles @5.9 Degrees from Winged Horses)
Occupation
Electrical Contractor
For testing purposes, remove the bond. Just put it back when that solution fails to work out. Removing the bond is not going to cause current to flow from that machine to another or you. It is specific to that machine. An accidental or unintended fault of one of the secondary conductors just makes it a grounded system again. Realizing the fuses may now be in the grounded conductor.
 

travish

Member
Location
North Carolina
the symptom is the motion controller displays an error that says a "CPU watchdog failure has occured. This is generally a result of EMI". The fault happens very random. maybe a month or slightly longer between occurances. the last thing that I tried was disconnecting the DH485 communication from the Operator interface to the motion controller. The problem has not reoccured since but I did have 1 instance last week where we were running one length of cut, finished that order and started a new order (different lenght) and the controller was still cutting the old length. Had to go thru the length change process again and it started running correctly.

I know this forum is not for fixing machine problems, Just wanted to make sure that I wasn't doing anything unsafe or a code violations by disconnecting the Ground.
 
IME with embedded systems (not specifically PLCs), watchdog timeouts are almost always caused by internal problems, and usually because of bad design (software can't handle the interrupt load or otherwise goes off in the weeds, hardware doesn't filter or debounce the inputs or stabilize the power, that sort of thing). It's possible that adding a filter to the control power might help, but otherwise any change is a usually a wild guess.
 

petersonra

Senior Member
Location
Northern illinois
Occupation
engineer
I doubt the suggested fix will solve your problem. This kind of thing is often suggested, along with adding RFI filters, and fixing "bad" grounds.

None of these things ever fixes anything.

It is best to try and figure out what is really going on but many times that is very difficult, and it often ends up being a design flaw in the electronic hardware and the only people who can truely fix it won't.

I have spent many an hour trying to track down these kind of intermittant and elusive problems. A lot of times the end user just cannot afford to have me spend the time it takes to track them down so they just live with them.
 

james_mcquade

Senior Member
A cpu watchdog failure is an error indication that something has happened to the machine program.
it is an internal timer that is used to determine how long the program takes to finish one program scan.
this preset timer is reset at the end of the program scan. if it times out, you get the message you described.

1. the plc has been told to execute a subroutine in an endless loop to the point that the program cannot ever finish.
2. a servo of motion controlled system has been told to move within a specified time and its not making it due to age or play in the
mechanisms.

since this is happening only once a month, i suggest you go online with the plc and see what the watch dog timer
is set to, what the scan time was when it faulted, and look at the associated fault codes. This will be your first clue as to what is going
on. On allen bradley plc's, the watchdog timer can be changed with the software.

for specific help on other plc's goto www.plctalk.net.

regards,
james
 

travish

Member
Location
North Carolina
James, it is not in the PLC, the motion controller is giving the fault. I wrote the PLC program for the machine. I give an enable signal to the motion controller and it sends a 0 - 10v speed referance to the servo drive (really a fluxvector VFD) this Motion Controller times 2 encoders so that a rotory cutter, cuts at the correct location. we are forming rebar and cutting to length. we built the machine here at work, and this Motion controler is the only thing I have had an issue with.

I do pretty decent with electricity and PLC's, that motion stuff is over my head. that is why I bought an "off the shelf" motion controller instead of doing the motion in my controllogix processor.

Bob, I put a RFI filter infront of the motion controller, as you said it didn't help. I have borrowed a scope meter and am unable to find any noise that means anything to me. but I am much more comfortable with meter than a scope. I like Voltage on/off. not there is a little spike in every third sinewave or something like that. electrictians like me shouldn't be looking noise problems.
 
Last edited:

GeorgeB

ElectroHydraulics engineer (retired)
Location
Greenville SC
Occupation
Retired
A watchdog timer error is, as others have said, an indication that the program "locked up". I'd review the motion controller program for a step that cannot be finished SOMETIMES. There is a chance that the system will keep a log of errors.

To echo, plctalk.net or the drive manufacturer's site would be good resources.

Good Luck!
 

ELA

Senior Member
Watchdog timeouts can also occur due to noise. Electrical transients can cause the program counter to vector off to unused or incorrect memory locations. Thus the watchdog times out.
Often times lifting a chassis ground can help with noise issues but are probably not the actual cause. If lifting the ground improves the situation then you may have a ground loop.
You mentioned disconnecting the RS485 improved the performance ? Is there a logic common in that line that is bonded internally to the equipment or chassis ground?
 

travish

Member
Location
North Carolina
I am not sure that disconnecting the DH485 made the preformance any better, I just have not had the fault since I removed it, but it has not been that long either so I am unsure about wheather or not it helped.

With my volt meter there is no connection of the common of the motion controller to the common of my PLC and 24vdc power supply. I have already been thru and either disconnected or used optical isolators to try and eliminate all possible sources of noise. I am using a relay card on my plc to send signals to the motion controller.
 

ELA

Senior Member
the symptom is the motion controller displays an error that says a "CPU watchdog failure has occured. This is generally a result of EMI".

That tells you that the motion controller is susceptible to EMI. A company I worked for years ago owned the design of a PLC that would watchdog all the time due to transient electrical events.
That was when we got serious about "designing for EMI". When a piece of equipment is not designed to withstand appropriate levels of EMI they will often suggest that the problem is in the grounding.
Do the specifications on the unit state their EMI immunity levels? If not then I would put it back on them.

EMI can enter a device in many different ways. Either conducted or radiated. Cables connected to the device provide a common entry point. The longer the cable the better "antenna" it makes (generally speaking).
Any chance that hand held radios are ever used in proximity to the equipment? The possibilities are many when it comes to EMI. I agree an electrician should not be the one responsible for troubleshooting EMI issues (not saying an electrician should not attempt it, just not be held responsible). There are lots of EMC consultants out there for hire. ( I say this but be cautious)...

We once had a field issue where the manufacturer of a piece of equipment was telling us that our grounds to their equipment were insufficient. They hired an EMC consultant who wrote up a nice 7 page document "proving" our grounds were bad ( too high of a resistance).
It turned out that he took resistance measurements with the power on to the equipment! I was very surprised by this.

I then spent several days on site monitoring via oscilloscope with 2 engineers from that company. When we observed a transient event they stated " look there is the transient caused by your poor grounds". I noted that the "transient" was not bi-polar which I commonly see when it is a noise event.

After capturing a few more events we found that their equipment had an intermittent open, on a wire in a cable, that was part of their motion control!

Best of luck to you.
 
Status
Not open for further replies.
Top