Correct Shield Installation

[Ken9876]

I have a job I take care of that uses 4-20 ma resolvers, recently I had a problem with one so I started poking around. I found that the shields from separate resolvers had been connected together at a junction box and at the plc cabinet. Shouldn?t these be separate? And what is the correct way of connecting the shields? I read grounded at one end; however resolver lit shows the shield grounded at the resolver and at the other end.

 

[gadfly56]

I have a job I take care of that uses 4-20 ma resolvers, recently I had a problem with one so I started poking around. I found that the shields from separate resolvers had been connected together at a junction box and at the plc cabinet. Shouldn?t these be separate? And what is the correct way of connecting the shields? I read grounded at one end; however resolver lit shows the shield grounded at the resolver and at the other end.

I'm not an expert on current loop devices, but I would tend to agree that the shields shouldn't be grounded at both ends. If grounded at one end, it can't matter if two or more are grounded at the same location. If grounded at both ends, I don't see how you avoid ground loops if all the shields are tied at a common point "A" and then to "B", "C", "D", etc.

 

[gar]

110708-1304 EDT

What is a 4-20 MA resolver? Is this a resolver with electronics, and the output converted to an analog 4-20 MA for angular position, or something else?

Are your shielded cables a twisted pair inside the shield?

Does the resolver work well without the long cable?

What kind of a problem is being exhibited?

How is the resolver being powered?

If you replace the resolver with a a 4-20 MA simulator does the problem go away?

.

 

[gar]

110708-1312 EDT

gadfly56:

Why or how would a ground loop affect a 4-20 MA loop? I am not saying it won't, but what is your logic that it would?

I would assume that at one or both ends of a current loop that the electronics would be floating relative to ground.

.

 

[Ken9876]

110708-1304 EDT

What is a 4-20 MA resolver? Is this a resolver with electronics, and the output converted to an analog 4-20 MA for angular position, or something else?

Are your shielded cables a twisted pair inside the shield?

Does the resolver work well without the long cable?

What kind of a problem is being exhibited?

How is the resolver being powered?

If you replace the resolver with a a 4-20 MA simulator does the problem go away?

.

Yes it's for angular position. It's powered from a power supply which powers two other devices. Not sure if the conductors are twisted inside of the cable. This device is causing a fault in the plc program, the program compares the output of this device to the output of another device. There are two resolvers only one selected at a time, the program compares resolvers 1 or 2 to another angular device. The system works fine with 2 but not with 1. As for the simulator we don't have one. I just started looking into this and was looking for additional info.

 

[gar]

110708-1536 EDT

Ken9876:

My guess is a problem with device 2.

Most multi-conductor shielded cables with pairs of wires have each pair twisted. Twisting reduces broad area magnetic coupling. In CAT-5 cable there is a difference in the twist pitch from one pair to another to reduce the likelihood of a mutual coupling.

If the transmitters (the resolvers) have their outputs floating and likely are not the current source, but just a modulator of the current, then the output current should not be greatly affected by grounding of the transmitter chassis. Thus, the transmitter ground potential should be able to be different from the receiving end without a problem in the current loop signal. This would mean an isolated DC to DC supply within the transmitter.

You can make a current loop transmitter simulator with a LM317 regulator, a resistor, and a capacitor. I just ran a test at about 5 MA. The current thru this load changes less than 0.1% from 3 V to 25 V across the load.

I used a 249 ohms 1/4 metal film current sensor resistor, and an LM317 in a T03B package. Use a 6.8 mfd 35 V tantalum across the input to output. With the resistor in place the output is the lower end of the resistor which also connects to the reference pin.

Something around 62 ohms for a current sense will put you near 20 MA. A 1/4 W resisitor is still suitable.

A 250 ohm 10 turn Helipot in series with a 49.9 ohm resistor would make an adjustable current loop simulator.

.

 

[gar]

110709-0825 EDT

Ken9876:

I put a higher resolution meter on the circuit this morning. This has to be considered short time stability as contrasted with hours to weeks.

From 3 V to 25 V the current reading was 5.001 or 5.002 mA. On this 20 mA range the meter resolution is 0.001 mA. So we are at the noise level of the meter, and have a variation of less than or equal to 1 part in 5000.

More than adequate for any testing you would do.

.

 

[Ken9876]

Gar,
Thanks for all the info.
I was thinking the problem with device 2 as well. One more question, each resolver has it's own cable. At a junction box and at the plc cabinet all the shields from each device were connected together. Do think this would cause a problem?

 

[gar]

110709-1138 EDT

ken9876:

I do not think shields are a problem for a number of reasons.

Current loop implies that one or both ends of the current loop are isolated from earthing (grounding to some common entity -- the earth, building steel, the EGC wire, etc.). Thus, you are dealing with a loop of wire for the current loop, a current source, a current modulator, and a current sensor. If this loop is grounded at most at one point, then ground path current can not affect the current in the loop. The shield is not part of the current loop path.

Noise currents can enter the the loop via inductive or capacitive coupling. By using a twisted pair for the current loop wiring this minimizes inductive coupling. The twisting does a phase reversal and thus cancellation of inductive signals covering a large area. Capacitance reactance coupling into the current loop is generally not large compared to the loop impedance. Thus, not likely a problem.

Bad wiring, bad sender (source of current modulation), bad current source, bad current sensor are all likely causes of the problem.

Using a constant load current device, like I described, in place of the present sender and if the sensed results were stable, then this would likely eliminate everything except the sender.

.

 

[gar]

110709-2217 EDT

In post #9 I used the term "current source". This should not be read to mean a constant current source. Rather it is really a voltage source, may not need to be a very good voltage source, that supplies whatever current the current loop requires.

If this voltage source and the current modulator were in the same box with two output terminals, then that box might be described as a controllable constant current source.

.

 

[Smart $]

... I found that the shields from separate resolvers had been connected together at a junction box and at the plc cabinet. Shouldn?t these be separate? And what is the correct way of connecting the shields?

I've always been instructed by reputable sources that 4-20mA shields are to be terminated only at supply end. And the more professional installations I've been involved with use "spaghetti" tubing to insulate exposed shield between cable jacket and terminal. Exposed conductors should be kept to a minimum. At the field end, most any means of isolating/insulating the shield can be used. I usually wrap the shield at the end of the jacket and wrap with electrical tape (heat shrink is nice if available). I don't cut the shield off just in case there is a problem in the future and the shield needs to be rung out.

Splices should be avoided as much as possible, but when spliced, I prefer they be made with terminal blocks using a individual feed-through terminal for each conductor and each shield. However, there are those that will say crimp-on butt splices, or even wire nuts are acceptable. Its all good until problems present themselves :roll:

 

[gar]

1107010-1156 EDT

Smart $:

Provide a good technical reason why your procedure is so important. I agree that you probably should not ground the shield at both ends, but if the shield had adequate cross sectional area, then I doubt that 10 A at 60 Hz thru the shield would induce sufficient current in the twisted pair inside the shield to produce an error in this application.

Non-magnetic shielding of a wire, in other words a copper or aluminum shield, does not shield internal wires to the shield from internal or external magnetic fields. Thus, normal shielding of a current loop pair of wires does virtually nothing to protect the current loop circuit from magnetic field interference. What is important is the twisting of the wire pair to minimize magnetic coupling. Capacitive coupling is a different story relative to the value of electrostatic shielding.

I am not suggesting that you should ground both ends of the shield of a shielded cable. What I am suggesting is that the problem of the original post is probably not related to where the shields are grounded.

What could happen and be a problem is if both ends of the shield were grounded at different points, and a large current flowed thru the shield causing overheating and melting of the insulation.

.

 

[Rockyd]

I've always been instructed by reputable sources that 4-20mA shields are to be terminated only at supply end. And the more professional installations I've been involved with use "spaghetti" tubing to insulate exposed shield between cable jacket and terminal. Exposed conductors should be kept to a minimum. At the field end, most any means of isolating/insulating the shield can be used. I usually wrap the shield at the end of the jacket and wrap with electrical tape (heat shrink is nice if available). I don't cut the shield off just in case there is a problem in the future and the shield needs to be rung out.

Splices should be avoided as much as possible, but when spliced, I prefer they be made with terminal blocks using a individual feed-through terminal for each conductor and each shield. However, there are those that will say crimp-on butt splices, or even wire nuts are acceptable. Its all good until problems present themselves :roll:

Smart $ proves up to his name. That is exactly the way we do it. A wel designed system reduces a lot of troubleshooting in the future.

 

[Smart $]

1107010-1156 EDT

Smart $:

Provide a good technical reason why your procedure is so important. ...

Without doing a lot of research, I cannot provide any technical reason to counter your assertions. And I'm even doubtful if such research would be fruitful, much less worthwhile. Instrumentation experts, from manufacturers through engineers to field technicians, all say the best wiring practice is as I mentioned earlier. I have been on several large-scale projects where some shields were terminated on both ends, and/or shields were not insulated or they were connected together in splice boxes, and those systems always had problems. That's enough reason for me...

 

[gar]

110710-1544 EDT

Smart $:

I like nice good quality workmanship, and what you describe accomplishes that goal. But in the problem presented by Ken9876 would replacing the present wiring, or changing how the present shield was grounded solve the problem? I doubt that it would unless there is some damage to the wiring. You could probably use plain unshielded twisted pair wiring and get successful results. The shield as a shield may really have little effect on the operation of the current loop.

But I also want to point out that just because something looks nice, does not mean it will work correctly. I was called to the Chrysler Eldon Axle Plant back when the Germans were in control about quoting on a gage to measure case preload. The head process engineer was a German and had someone build a gage for them. It did not work and his whole criticism of the gage was that it did not look good. I could not talk to him about the basic theory of the gage which would never work no matter how nice the gage looked. No way would I build a gage for him, or do any work for him.

I can tell you many stories where manufacturers of CNC equipment will tell the customer that they need a good ground at the CNC to solve their RS-232 problems. There is no definition of a "good ground" or why this is a solution, and the solution of driving a ground rod at the machine may not solve the problem. Fundamentally, at least some people at the manufacturer, do not really understand the cause of the communication problem, and sometimes some of their suggestions seem to be solutions, but may be very marginal, and are not a solution at all in other situations.

.

 

[Smart $]

110710-1544 EDT

Smart $:

I like nice good quality workmanship, and what you describe accomplishes that goal. But in the problem presented by Ken9876 would replacing the present wiring, or changing how the present shield was grounded solve the problem? ...

I was not suggesting that it would solve the OP problem. I was only answering his question(s) about proper (or shall we say recommended) shield terminations and splicing practices, and that not following these practices is known to have caused problems on 4-20mA instrumentation systems.

So what would you say if he did revise the shielding as mentioned and it did cure his problem?

 

[gar]

110710-1719 EDT

Smart $:

So what would you say if he did revise the shielding as mentioned and it did cure his problem?

I would want to try to find out why. That means questions would be asked.

We need to know if the housing of the resolver assembly was earthed or whatever at its end, or was it floating from any ground path including thru its power supply. If the resolver assembly was isolated, then one might want to connect the shield to the resolver housing.

.

 

[Smart $]

110710-1719 EDT

Smart $:

I would want to try to find out why. That means questions would be asked.

We need to know if the housing of the resolver assembly was earthed or whatever at its end, or was it floating from any ground path including thru its power supply. If the resolver assembly was isolated, then one might want to connect the shield to the resolver housing.

.

Questions asked, for sure.

But if the resolver assembly was isolated, one might "want" to connect the shield to the housing, but the proper thing to do is ground it with a bonding jumper (the NEC actually requires it).

 

[gar]

110710-2238 EDT

Smart $:

Does a microphone carried around by a performer have to be grounded at the point of its use, vs grounded thru the microphone cable at the amplifier input?

.

 

[Smart $]

110710-2238 EDT

Smart $:

Does a microphone carried around by a performer have to be grounded at the point of its use, vs grounded thru the microphone cable at the amplifier input?

.

Actually, the microphone doesn't even have to be grounded. Yes, many are... through the cable.

I acknowledge the point you are attempting to make, and the matter is debatable. However, debating the issue will serve no purpose in the end, at least from my perspective, so I'd rather not. If you want to recommend grounding field instrumentation using the cable shield, be my guest.