Automation system install - Cable shielding efficiency test

[dedshaw1612]

In my industry of installing automation systems (robot cells, conveyors, ect) in factories, some customers have a spec that requires electricians to run a test with a special fluke device on the ethernet cable runs to make sure that they are good. I forgot the name of the device, but it is very expensive. The electrician places one device at one end and plugs the RJ45 adapter of the cable into it, and then the other RJ45 into the main device. It runs through some analysis checks and determines how good the cable run is. This is basically a good indicator to show whether or not it should be replaced so people aren't scrambling down the road trying to figure out weird issues.

I was curious if there was such a device made for determining how well the shield is working for a cable. There have been times in the past i had to interface with a device and the manufacturer spec was that I had to use a shielded cable. I'm sure some of you are aware the pains and the rabbit hole you go down when you start talking about ground loops or just ineffective shielding in general.

Sometimes the solution isn't all inclusive or easy, and instead of buying one outright I have to find part numbers of a spool of shielded cable with the correct OD that will fit into a terminating connector that has the right number of pins etc etc.

When i turn this design over to some install electricians, or even a prebuilt one, it would just be nice if I could run a test on the cable after installation that says "yep, this defiantly attenuates noise and does not have a ground loop per spec bla bla" Does such a device or robust technique exist? I am not even too concerned if i have to constantly reconfigure it with different connectors for each application like an M12 5 pin, M12 8pin, etc.

If the device or technique does exist, ideally it doesn't involve soldering. Just some terminals i can terminate some wires and flip a switch and run the test to call it day.

Thanks in advance

 

[gar]

210214-2212 EST

dedshaw1612:

Who do you work for, and where? What plants, and kinds of signals?.

A twisted pair for signal wiring will be less susecptacle to induced currents (voltages) from uniform magnetic fields over a number of twists.

A Beldfoil shielded wire pair, or single wire, will reduce capacitively coupled signals.

A magnetic shield will reduce magnetically induced signals.

Having a ground point at only one end is advantageous.

Using one or more isolation transformers can be useful.

.

 

[dedshaw1612]

i work in detroit MI and I used to work for a major robotics company as a controls engineer. This was over a year ago so some of the details are escaping me. I did Hardware design, plc programming, etc and we were doing some small upgrade for a luxury car manufacturer. It was a cavity wax process , in some places they use an expanding foam to reduce rust, here it is a very runny wax that hardens over time.

We outsourced the end-of-arm-tooling (EOAT) to a company that specializes in doing EOAT work of all kinds. Their in-house product was a proprietary design that used their own protocol/bus to send signals to various spouts to extend or retract , that then injected the wax inside the car body. They specified to me that i needed to use a shielded m12 8-pin cable. Trying to find something that was robot rated and shielded that was one piece was a very difficult task so i basically had to find the parts and design it from scratch.

while doing the design i just second guessed myself a bit on some of the things. If it can only be grounded at one end, one has to make sure that during install one of the connectors has to make sure the ground is continuous all the way back to the panel where the panel it originates from, and that at the end device, breaks the ground signal.

I can't keep my eye on electricians all the time, all i can do is give them drawings and direction. It would be nice if there was some sort of test i could do afterwards to verify.

 

[paulengr]

So Fanuc continues to eat itself.

Don’t attempt to disconnect shields at the ends. Just ground or don’t ground the ends as needed.

 

[hbiss]

some customers have a spec that requires electricians to run a test with a special fluke device on the ethernet cable runs to make sure that they are good. I forgot the name of the device, but it is very expensive.

That would be a certifier.

If the device or technique does exist, ideally it doesn't involve soldering. Just some terminals i can terminate some wires and flip a switch and run the test to call it day.

A simple continuity test would tell you all you need to know. I don't know if anybody makes something to test for it but off the top of my head all you would have to do is disconnect the cable, short the shell at one end to the other conductors then at the other end measure the resistance between the shell and the conductors. I'm sure somebody could make up a test adapter to make this test easy.

As for shielding effectiveness, that's a function of the cable design and what you do with it at each end.


-Hal

 

[gar]

210215-1706 EST

dedshaw1612:

It seems you may be talking about CAT-5 or equivalent cable. Also you may not be talking about long lengths. You need to be more specific about the function of the cables, lengths, and signals being transferred. What are the noise sources around the cables? What range of frequencies are being transferred? Are some simply analog?

I can transfer 9600 baud signals on CAT-5 cable over several thousand feet. Higher baud rates reduce the useful cable length.

If you need good noise immunity use fiber.

hiss's suggestion of an ohmmeter test is a starter.

Some of my waveform samples are at:

BETATRONICS&reg, Waveforms for CAT-5 Cable used for RS232 to RS232, communication, Long Cable, High Baud Rate, RS-232, fast, I232, vendor, supplier, manufacturer, industrial

Waveforms for CAT-5 cable used for RS232 to RS232 communication, Long Cable, High Baud Rate up to 115.2 kbaud.

www.beta-a2.com

These were mostly with an unterminated line.

.

 

[petersonra]

The testers are not that bad price wise. For the most part you need a cable tester that verifies the individual conductors have been run between the correct pins. Some of them have some other functionality but for run of the mill ethernet cables it really is not necessary to do much more than that if using good cables, good connectors and good tools to install the connector to the cable. The tool part is often overlooked. Can't tell you how many times I have seem field electricians trying to make up ethernet cables with a $10 tool they bought at Walmart.

You can get a decent cable tester for under $200 and a good crimping tool can be had for well under $100.

 

[hbiss]

For the most part you need a cable tester that verifies the individual conductors have been run between the correct pins. Some of them have some other functionality but for run of the mill ethernet cables it really is not necessary to do much more than that if using good cables, good connectors and good tools to install the connector to the cable.

Exactly. All you need is a $200 wire mapper (continuity tester). You don't need a certifier which does cost thousands. Really, the only time to use one is when the specs in the contract call for certification and a printed record of each run. And that's the time you sub the certification out unless you want to spend big bucks on something you may only use once.

Got me thinking now, I wonder if there is a wire mapper that will test the shield continuity on STP category cable.

Can't tell you how many times I have seem field electricians trying to make up ethernet cables with a $10 tool they bought at Walmart.

Yup.

-Hal

 

[dedshaw1612]

210215-1706 EST

dedshaw1612:

It seems you may be talking about CAT-5 or equivalent cable. Also you may not be talking about long lengths. You need to be more specific about the function of the cables, lengths, and signals being transferred. What are the noise sources around the cables? What range of frequencies are being transferred? Are some simply analog?

I can transfer 9600 baud signals on CAT-5 cable over several thousand feet. Higher baud rates reduce the useful cable length.

If you need good noise immunity use fiber.

hiss's suggestion of an ohmmeter test is a starter.

Some of my waveform samples are at:

BETATRONICS&reg, Waveforms for CAT-5 Cable used for RS232 to RS232, communication, Long Cable, High Baud Rate, RS-232, fast, I232, vendor, supplier, manufacturer, industrial

Waveforms for CAT-5 cable used for RS232 to RS232 communication, Long Cable, High Baud Rate up to 115.2 kbaud.

www.beta-a2.com

These were mostly with an unterminated line.

.

not really sure if you read any of my posts. It literally reads that i was using an m12 cable, and i wasn't asking for alternatives. When your force-fed a design and the mfg tells you to use an m12 8 pin shielded cable, you don't ask if you can use fiber instead.

 

[paulengr]

Exactly. All you need is a $200 wire mapper (continuity tester). You don't need a certifier which does cost thousands. Really, the only time to use one is when the specs in the contract call for certification and a printed record of each run. And that's the time you sub the certification out unless you want to spend big bucks on something you may only use once.

Got me thinking now, I wonder if there is a wire mapper that will test the shield continuity on STP category cable.



Yup.

-Hal

Yes, there are continuity testers with shield testing. I paid about $60 for mine. It’s a cheap one but I don’t do a lot of network pulls.

 

[paulengr]

not really sure if you read any of my posts. It literally reads that i was using an m12 cable, and i wasn't asking for alternatives. When your force-fed a design and the mfg tells you to use an m12 8 pin shielded cable, you don't ask if you can use fiber instead.

Easily defeated. Most testers come with only an RJ-45 female port. What to do if you need to plug it into a socket? Test two patch cords. In your case make up four “adapter” cords. Two with m12 plugs on one end RJ45 on the other, and two with M12 sockets and RJ45. Use the tester to verify pairs of cords. Now you have everything you need.

Allen Bradley started the goofy nonstandard connector nonsense to sell their special equipment at their hyper inflated prices. Well before they jumped into the game we already had RJ45 connectors with hoods that screwed into a matching bulkhead connector from Molex among others. So it takes standard crimpers to make up a cable and if you don’t have a cable handy or need to do some kind of troubleshooting, you don’t need an adapter. AB came along and with Belsen decided to be the rip-off artists they are. But when you put factory technicians free of charge in the plant to promote your product, it’s hard to stop this nonsense.

 

[petersonra]

Easily defeated. Most testers come with only an RJ-45 female port. What to do if you need to plug it into a socket? Test two patch cords. In your case make up four “adapter” cords. Two with m12 plugs on one end RJ45 on the other, and two with M12 sockets and RJ45. Use the tester to verify pairs of cords. Now you have everything you need.

Allen Bradley started the goofy nonstandard connector nonsense to sell their special equipment at their hyper inflated prices. Well before they jumped into the game we already had RJ45 connectors with hoods that screwed into a matching bulkhead connector from Molex among others. So it takes standard crimpers to make up a cable and if you don’t have a cable handy or need to do some kind of troubleshooting, you don’t need an adapter. AB came along and with Belsen decided to be the rip-off artists they are. But when you put factory technicians free of charge in the plant to promote your product, it’s hard to stop this nonsense.

M12 and M8 connectors are a lot more secure and environmentally sealed than any of the RJ45 plugs though. AB did not invent M8 and M12 connectors. They are a well established European standard with a lot of companies supporting it.

 

[paulengr]

M12 and M8 connectors are a lot more secure and environmentally sealed than any of the RJ45 plugs though. AB did not invent M8 and M12 connectors. They are a well established European standard with a lot of companies supporting it.

Agreed. But not TIA. And M12 and M8 aren’t designed for 250 MHz. They are meant for control signals. Even CAT 7 and 8 use basically RJ45 plugs....of course this is “abusing” the RJ45 standard too.

In sewage plants I coat the plug pins in silicone grease then plug it in for a gas tight seal. Old telecom trick.