shielded cable problems

[buzzp]

We have a current transducer that has a 4-20mA output. We are using twisted shielded cable. The length of wire is around 75 feet at most to the display device. When the shield is left ungrounded, we get the proper reading on the meter. When the shield is grounded, we get a triangle looking wave (AC from 90-120Hz up to 55V, so I am told) imposed on the 4-20mA signal. Only one end of the shield is grounded when this happens. The cause is not from where the cable is routed because we have temporarily ran wire on the floor with the same result. We have metered the cable to the pairs and there is no short to the pair. I do not have a cable part number yet but it seems there is something radiating in the vicinity of the cable. There are four exact systems and only two are doing what I described (these two are located by each other and the other two installations are about 50 feet away).
My thoughts on the potential cause is:
1. something is radiating in the vicinity of the cable, but grounding one end should take care of it
2. the impedance of the shield to the pairs is such that the radiated frequency causes a low impedance to the pair from the shield causing the problem (this seems unlikely to me)
3. sounds like it could be a second order harmonic issue but not familiar with sources of second order harmonics
4. Ground is likely a noisy ground but I am not familar with the plant layout
This is a power generation facility and the area this is in is the governor/exciter systems. Any advice or comments? I was trying to decide if a filter on the input to the meter would help anything like a ferrite core but think the frequency is to low for something like this.
Sorry for the vague info but I will post more info as it becomes available.

 

[buzzp]

Perhaps we are in a resonant condition?
I think it is a noisy ground situation. I am told we have a "ground mat" and this can not be the cause. I am just learning about generating stations but everything is pointing towards some problem with the ground.

 

[ScottyUK]

Buzzp,

How many poles does your generator rotor have? Would a four pole rotor account for the 100Hz pickup? Two poles would give 50Hz, so...?

Ferrite core at this frequency would have no measurable effect, so don't waste your money.

 

[buzzp]

The generator itself has 32 poles (225rpm), hydro.
I did not think ferrites would help any.
I am still compiling information and will likely make a trip there to get all the details. They claim the same interference is on the signal wires with the shield disconnected from ground but it displays the correct value. I think they are measuring from ground to the signal wires and not across the signal wires itself otherwise the display should still be wrong.

 

[pablo02]

How is the speed of the generator measured -- is this old enough to use a tach-generator (small 4 or 5 pole generator usually found in the same area as the governor -- the voltage appears to be right) -- could this be the source?

which end are you grounding? the end farthest away or the end closest? you should try grounding the other end instead, or you might even try grounding both ends for this application... with one end ungrounded, you can possibly build up enough voltage on the other (end not grounded) to affect your readings...

if you do have a tach-generator, you might try shielding it's wires...

 

[jbartos]

Suggestions to buzzp (Electrical) Aug 19, 2003 marked ///\\We have a current transducer that has a 4-20mA output. We are using twisted shielded cable. The length of wire is around 75 feet at most to the display device. When the shield is left ungrounded, we get the proper reading on the meter. When the shield is grounded, we get a triangle looking wave (AC from 90-120Hz up to 55V, so I am told) imposed on the 4-20mA signal. Only one end of the shield is grounded when this happens. The cause is not from where the cable is routed because we have temporarily ran wire on the floor with the same result. We have metered the cable to the pairs and there is no short to the pair. I do not have a cable part number yet but it seems there is something radiating in the vicinity of the cable. There are four exact systems and only two are doing what I described (these two are located by each other and the other two installations are about 50 feet away).
My thoughts on the potential cause is:
1. something is radiating in the vicinity of the cable, but grounding one end should take care of it
///Up to certain frequency, e.g. 1MHz. Above it, more grounds are needed. Consults textbooks on EMI.\\2. the impedance of the shield to the pairs is such that the radiated frequency causes a low impedance to the pair from the shield causing the problem (this seems unlikely to me)
///Yes, correct.\\3. sounds like it could be a second order harmonic issue but not familiar with sources of second order harmonics
///Yes, it might; however, chances appear to be very remote. One well-known source of second harmonic and even harmonics is the half-wave rectifier.\\4. Ground is likely a noisy ground but I am not familar with the plant layout
///Power system ground, e.g. grid, is considered noisy. The instrument or computer ground separate from the ground grid is considered quiet ground.\\This is a power generation facility and the area this is in is the governor/exciter systems. Any advice or comments? I was trying to decide if a filter on the input to the meter would help anything like a ferrite core but think the frequency is to low for something like this.
///The power ground grid may be very noisy.\\Sorry for the vague info but I will post more info as it becomes available.

 

[buzzp]

There is another issue with loss of communications on a RS485 comms line in the same area. This is leading me to believe there is a noisy ground. I am trying to find what type of set-up they have on the network as far as terminating resistors, 2 or 4 wire, if using a signal ground, and any isolation. I believe the two problems, incorrect meter display and loss of comms are caused from the same problem.
I will find out how the speed is measured. I believe this has been upgraded.