Proper landing of shielded wire within cabinet

Sounds to me like you're getting EMF from a variable speed drive. That could be because your transducer (or transducer wiring) is too close to the motor or the cable feeding the motor. I've even seen noise introduced into the AC control power from drives without input reactors or an isolation transformer affect analog inputs. Is your instrument cable run inside a steel conduit? A shield can only reject so much noise.

currently the wires are only connected from the control cabinet to the machine for testing purposes within the shop, so they are in no conduit at all. although the is quite close to a slew of high voltage AC wire, i will have to rectify that tomorrow.

-are analog output signals just as susceptible to noise? i have an analog output card in the cabinet sending two separate 4-20ma signals to two separate vfds for speed control and watching the drives the HZ stays dead nuts when commanded.

-could i test with a voltmeter how much noise i am seeing on the signal wire? lift the shield from the cabinet, put it on one probe on meter, put other probe on meter to ground? or would i lift the signal wire from the transducer and use that while leaving the shield in place?
 
-are analog output signals just as susceptible to noise?

Yes.


-could i test with a voltmeter how much noise i am seeing on the signal wire? lift the shield from the cabinet, put it on one probe on meter, put other probe on meter to ground? or would i lift the signal wire from the transducer and use that while leaving the shield in place?

I'm assuming that you can see the effect of the noise on your analog input. An oscilloscope would be the ideal instrument for monitoring EMF levels, but you should be able to identify and correct the problem without using a scope.

Moving your signal wiring away from any AC power wiring is a good first step.
 
Yes.




I'm assuming that you can see the effect of the noise on your analog input. An oscilloscope would be the ideal instrument for monitoring EMF levels, but you should be able to identify and correct the problem without using a scope.

Moving your signal wiring away from any AC power wiring is a good first step.

My plan to solve the problem tomorrow is to first move the wiring. Second, then possibly get the transducer manufacturer on the phone and see if they recommend the unit being mounted to metal to act as a ground for the case, or if it can be installed as it currently is with the plastic “isolator” between it and the metal.
 
Instead of regular terminals for the shield, use dedicated shield terminals. See post #12.

The other thing you can do is to fight the noise at the source.
Do the VFD have an output filter ?
Are the motor cables shielded ? The shielded motor cable can be grounded in both ends (signal cables only one end).

edit: The isolator between transducer and machine mount could have something to do with avoiding ground loops.
IMO it would be a design flaw if the electronics inside transducer is not already isolated from the transducer housing. But may it is not isoloated, hence the need for the isolator.
 
Last edited:
Instead of regular terminals for the shield, use dedicated shield terminals. See post #12.

I’m new to all of this, control panel design and building even more so then programming and such. Are those just ground buss bars in that picture? I’ve never seen anything that neat in my searching for buss bars on the companies websites that we typically deal with.


The other thing you can do is to fight the noise at the source.
Do the VFD have an output filter ?
Are the motor cables shielded ? The shielded motor cable can be grounded in both ends (signal cables only one end).

I’m not sure, I would assume it does I will look today.
The motor cables, against my request, are not shielded.

edit: The isolator between transducer and machine mount could have something to do with avoiding ground loops.
IMO it would be a design flaw if the electronics inside transducer is not already isolated from the transducer housing. But may it is not isoloated, hence the need for the isolator.

I agree. But it was interesting to me to read one manufactures install instruction which specifically states to make sure the body is bolted to a grounded metal surface, and then the manufacturer that we use says no such thing one way or the other.
 
Are those just ground buss bars in that picture?
Yes. With specifically the purpose to terminate the shields. The difference is that with the clamp the shield has a much better connection to ground, as compared to twisting it together and sticking it in a regular terminal.

[..]one manufactures install instruction which specifically states to make sure the body is bolted to a grounded metal surface,
That one probably has isolated electronics. Grounding the housing will work the same as grounding the shield. But.... if the housing is grounded, then make sure the shield is only grounded at that end, and not at the other end.
Or (!) .. make sure the shield is not connected to the housing, and then ground the shield at the other end.
Both to avoid current loops.

[..] and then the manufacturer that we use says no such thing one way or the other.
Then in that one the electronics is probably not isolated.
 
I moved the signal wire away from the high voltage ac wires from when we just interconnected the machine to the control panel for in house testing purposes and that seemed to clean up the noise a bit.

It’s still an issue in as much as when the motor is running that’s moving the object that the transducer is attached to, you get a reading that’s about 2inches off from actual once said motor stops it’s output.

I liffed the shield wire and attached my meter to it, and the other side of the meter to ground. There is AC mV worth of noise on the drain wire, ranging from 60-160mV depending on which motors are running. So I assume that’s still my issue with my less then ideal 4-20ma reading into my PLC.

I will look into building my next machine cabinet where I require 4-20ma signals with the ground buss bar for shields as mentioned above to hopefully quell some if not all of th me issue I’m having now.
 
I would like to suggest some tests to help you determine the source of your problem.
1. When you move in both directions does the measurement always "jump ahead" (2 inches in the direction you are traveling) when you stop?
2. Presuming the VFD is moving the item you are measuring can you disconnect the motor from the movement or disconnect the analog sensor from the item. Then run the VFD and note any changes in the analog sensor reading.
3. Finally move the sensor as close to speed as you can by hand and note the readout. Try this with the VFD running and not running.
4. Again presuming the VFD is moving the sensor. Run the VFD as slow as you can and note the sensor error.
 
Last edited:
I would like to suggest some tests to help you determine the source of your problem.
1. When you move in both directions does the measurement always "jump ahead" (2 inches in the direction you are traveling) when you stop?
2. Presuming the VFD is moving the item you are measuring can you disconnect the motor from the movement or disconnect the analog sensor from the item. Then run the VFD and note any changes in the analog sensor reading.
3. Finally move the sensor as close to speed as you can by hand and note the readout. Try this with the VFD running and not running.
4. Again presuming the VFD is moving the sensor. Run the VFD as slow as you can and note the sensor error.


1. I’m 99.9 percent sure that is the case, I will double check Monday morning.
2. A VFD is moving the item that the string transducer is attached to yes. I can easily disconnect the analog sensor from the movement and then run the machine and see what noise comes into the analog card.
3. I can do this as well Monday morning.
4. The vfd currently runs at two different preset frequencies, 10 and 40hz, I can lower the lesser frequency and see what I get

Thank you for the ideas!
 
I would like to suggest some tests to help you determine the source of your problem.
1. When you move in both directions does the measurement always "jump ahead" (2 inches in the direction you are traveling) when you stop?

exactly

2. Presuming the VFD is moving the item you are measuring can you disconnect the motor from the movement or disconnect the analog sensor from the item. Then run the VFD and note any changes in the analog sensor reading.

there is the same amount of noise it seems. sadly dont have a scope to be absolutely sure.

3. Finally move the sensor as close to speed as you can by hand and note the readout. Try this with the VFD running and not running.

moved the sensor with the vfd by hand and it still read the same. i did also just move the sensor itself, without the vfd on and once i stopped the reading stopped dead on.

4. Again presuming the VFD is moving the sensor. Run the VFD as slow as you can and note the sensor error.

when i run less HZ from the VFD the noise is absolutely less.

i pulled the cabling that runs to the transducer out of the machine to get clear and away from any high voltage AC wire, cleaned up the signal a bit but still not good enough.
 
moved the sensor with the vfd by hand and it still read the same. i did also just move the sensor itself, without the vfd on and once i stopped the reading stopped dead on.

Well, that narrows it down to the VFD as the culprit (not that it was a surprise)

i pulled the cabling that runs to the transducer out of the machine to get clear and away from any high voltage AC wire, cleaned up the signal a bit but still not good enough.

If I remember correctly you requested shielded cables for the motor and you were denied. If your application needs a flexible power connection to the motor you will need (barring shielded cable) liquidtite and grounding lead from the motor to the drive, then to the ground buss.

Another preventive measure would be a signal conditioner/isolator
https://www.eurotherm.com/products/io-signal-conditioning/action-iq-ac-dc
https://www.automationdirect.com/ad...Voltage,_Current_or_Potentiometer_Input/FC-11
 
can you actually SEE the noise with an oscilloscope? i know you can use one to see noise on an encoder, but since this transducer puts out a 4-20ma signal and not just pulses like an encoder would can a scope still be a useful tool? how would one determine if the 24v signal could be "dirty", i assume that youre thinking that could be an issue since you recommended a signal conditioner for said instrument.

If I remember correctly you requested shielded cables for the motor and you were denied. If your application needs a flexible power connection to the motor you will need (barring shielded cable) liquidtite and grounding lead from the motor to the drive, then to the ground buss

correct, was told that "its not needed"o_O
there are 4 separate motors on this machine, all needing flexible power connections. i assume with the liquatite you would want me to run it from the motor to the junction box on the machine...junction box on machine to the main cabinet will be liquatite or conduit once installed on site. issue with the former is at this point that not possible, would need to do that earlier on in the machine build.



this i will order tomorrow from auto direct and see if that does the trick. ideally next machine i will be more convincing and get shielded motor cable in there to quell any of these issues before they become issues. Thanks all for the recommendations.
 

Similar Topics

Dear All Please help me about Below picture problem.PLC not communicating proper.
Replies
4
Views
161
Hi. I’m looking to replace a basic Shark power monitor with a newer one with communication options, probably PowerMonitor 5000. The existing one...
Replies
2
Views
965
hello guys, i<m wondering if any body knew a method to export a routine with its proper tags in rslogix studio please
Replies
2
Views
2,571
This may be a dumb question, but do you try to nail the VFD size on the head or go up a size? I can see the value in going up a size of VFD in...
Replies
13
Views
4,587
We've had an intermittent issue with this for years. The application is sensing a level, or more accurately the absence of level in smallish ~5-10...
Replies
13
Views
7,253
Back
Top Bottom