Well, I was taught that i never ground shielded audio cable on both ends as it may create ground loops that cause interference to the signal. I could not find proposition for UTP cables.
How do you do it? Do you ground on both ends?
Thanks, but I already saw that as I stated.
as I stated.
you did not state that anywhere. to avoid grounding loop you need to follow method 2 and read the note on the bottom of the page.
I usualy have ground on both ends, as I use metal cases and they get grounded just because they are mounted on another metal surface or because of power supply grounding.
Thus, I tend not to connect shield to ground on both ends of UTP cable. Iwonder if I should continue this practice or connet to ground no matter what. How does ground loooping influence signals through UTP cable?
Follow the same practice you use to do with balanced audio cables: shield is grounded only at one side to avoid ground loops.
Still a balanced system with its own possible CMRR figures and the balanced configuration itself will reject wideband signals & noise.
But low frequencies (50/60Hz to a couple MHz such as long & medium wave radio signals) will be rejected only by twisted pair CMRR figures.
This thin aluminum foil have almost no effect below 5MHz. Balanced audio cables used to have dual braid to improve this at low frequencies (except those thin ones that was used for short runs inside racks or between them on the same floor).
Regards;
mramos, thank you. That is kind of discussion I wished to start.
You are not doing it right. If you use any kind of grounding on the tower, it’s enough to connect only one end of the shielded cable. If however you are not connecting the board to ground, the shielded cable must be connected to the shielded ethernet ports on both ends, and then the shielded cable will serve as a grounding cable.
About the grounding loop - the 50Hz noise (electromagnetic induction) which is created because of the loop is basically not affecting the 100/1000 mbit ethernet lings, you just have to take in account that there will be additional heating of the ethernet shield in the cable.
Also important to always use FTP cable or S/FTP cable.
Follow the same practice you use to do with balanced audio cables: shield is grounded only at one side to avoid ground loops.
please don’t give wrong advice if you are not sure. your advice is only valid if the board is grounded.
May be I express myself wrongly.
I assumed that to talk about loops is mandatory that you already have a grounding path - so - when you connect both ends at data cable you’ll create it (the loop). If the board is not grounded there is no reason to talk about “loops” due to the lack of a second ground path.
At the wiki there is the second picture where board is not grounded e.g devices that have built in antennas (SXTs??) or antennas where pigtail came from the fedder structure which is isolated such as some grid reflectors (grid + groove?). In those cases you can, better - you must - use cable shielding as ground path / electrostatic drain path if those devices do not provide a ground terminal.
About loops affect or not ethernet links I’m not sure, depends on. We faced some years ago a situation where several cable modem controllers was at one building floor and had it’s own ups & etc using one phase of the electric mains. One floor bellow there was several Cisco switches and a Siemens STM16 SDH/ADM plugged to another UPS using another mains phase. All grounded of course.
When certifying this instalation there was random data errors, packet loss & etc. They only stopped when both floors started to share the same phase. May be the CMTs ethernet have some “issue” leaking this phase difference (eg varistors at the input eth isolation transformers grounding the data lines to protect the transformer itself) but there was a huge 120V RMS 60Hz sine (vy low current anyway) swinging each data pair from -60 to +60 V. Twisted pairs low frequency rejection is ~60dB.
Regards;
thanks for clarifying your point.
By the way, if you use POE to connect outdoor device, isn’t it already grounded by the minus pair of the cable(brown, using TIA/EIA-568-B standard)? So maybe there is no reason to use STP cables to ground the outdoor device?
The dc power supplies are isolated from ground due to the decoupling of the transformer.
The transformer is then grounded directly to the power companies ground.
The only reason for grounding loops would be poor conductance of the ground, ie. Relying on the bolted sections of a tower to ground.
Ultimately a 1 AWG wire should be run the full length of the tower for proper grounding, and ground loops should be avoided.
OK, I read several terms and explanations now in this topic. To keep this understandable for all of us dummies can I have some explanation of the following:
“ground loop” - What is exactly mend by this?
“CMRR figures”? (Ramos) - What does it stand for?
“When certifying this instalation there was random data errors, packet loss & etc. They only stopped when both floors started to share the same phase. May be the CMTs ethernet have some “issue” leaking this phase difference (eg varistors at the input eth isolation transformers grounding the data lines to protect the transformer itself) but there was a huge 120V RMS 60Hz sine (vy low current anyway) swinging each data pair from -60 to +60 V. Twisted pairs low frequency rejection is ~60dB.” (Ramos) Can you explain this into plain English?
“By the way, if you use POE to connect outdoor device, isn’t it already grounded by the minus pair of the cable(brown, using TIA/EIA-568-B standard)? So maybe there is no reason to use STP cables to ground the outdoor device?” (ONdrejSkipala): Do you mean because the minus pair of the cable is attached to the earth of the board?
“Ultimately a 1 AWG wire should be run the full length of the tower for proper grounding, and ground loops should be avoided.” (tgrand): What stands “AWG wire” for? And why should such a cable run the full length of the tower if the tower itself is (usually) of metal?
We have radio cards where it is not clear which connections are to be used for grounding. I asked the question on the forum but never got an answer…
Some combined devices are hard (SXT) or impossible (Groove) to ground.
Most devices come standard supplied with power adapters or power splitters that are not made for grounding or use of shielded cable. What would be the risk in not using any grounding at all?
And does grounding not increase the risk of a lightning strike?
Using lightning arrestors, are they to be fit near the antenna or near the radio? (If antenna in top tower where board is at bottom).
On metal towers, still advised to run ground from top to bottom? What about the risk of lightning strikes? The better earth of the tower top to earth the higher the risk of lightning strikes?
When grounding boards in a Solar powered installation we have to be aware that some solar installations have to be grounded to positive (+) which than gives a definite short cut to the boards because they ground to minus (-).
Can the earth grid of a building (that comes with the electrical installation) be used for grounding antenna’s/radio’s or is it better to use separate? (Since these wires usually run parallel to the power lines would that carry inductive frequencies/magnetic pulse onto the earth we use for the radio installation? If not advices, what is than the best way to ground indoor installation?
Use of shielded cable; some say it has to be earthed on both ends, but others claim it is not needed. Some even say it still protects against interference (other task of the shield) even if not earthed.
I know about the “cage of farraday” but never learned about the need of connecting it to earth. Like a car is working as a cage of farraday but is definitely isolated to earth!
How does external frequencies (50/60Khz elec grid pulse, other sources) influence the Ethernet (10/100/1000Mb) connection?
In my quest to fight some virtual port flapping on one of my backhaul units (rb433AH+R52Hn) I had the antenna, the board and the radio connected to earth, used shielded cable with the shielded rj45 connectors and used power jack to supply board with 24Volt power.
Other end of the cable was attached to rb493AH which was properly grounded to same central earth pin.
In stead of virtual port flaps which disappearded after this I now had on a weekly base some REAL (!) port failures. It became so bad I had to do something. I took away the shielded cable and replaced it by unshielded cable. Since than (6 days now) no more issues…
One other rb433AH properly connected on metal plate grounded to grounded tower was working normally. When I replaced the R52H by a R52Hn to upgrade to mimo link and at the same time made a ground connect from that radio to the board the rb stayed dead… Only by disconnecting the ground wire of the R52Hn the board booted up and worked normal.
When board was booted and radio earth wire connected again it kept working! But after a (un-)scheduled reboot it stayed dead again…
Measuring the potential between the earth connection of the radio and the metal structure of the board casing I measured some 2 Volts? Where the hell is this coming from?
Bottom line:
A good explanatory Wiki document writing in plain English with explanatory pictures/drawings would make things clear for a lot of us and also might help understand why something the most weird behaviours and errors will pop up on our installations.
It should also go in to the differences and cause of ESD and lightning and interference and induction effects on radio’s and cables.
You wrote two pages of text, but there is already a Grounding manual that gives you all the information you need.
grounding loops are cause due to a potential difference from one ground point to another ground point.
This is due to resistance at the point(s) of connection.(ie. Bolt points of a tower)
To avoid grounding loops in audio many installers will solder connectors following crimping.
Oxidation on connectors also can create restistive forces.
Some metals are more suseptanble to this than others ( if our towers were made of gold we would have less chance of this phenomenon ).
AWG - American Wire Guage
1 AWG = 7.348mm diameter
When working on live power one is advised to always use a volt meter to measure the potential difference from electrical ground to a known good ground such as a copper water pipe, prior to starting any work.
Just because a tower is metal does not indicate it is a good conductor, especially across all of its bolting points, and especially because the metal is usually a galvanized type.
I would NEVER recommend using plastic standoffs.
some good information there, tgrand! thanks
so you agree some extra info is always useful. I’d rather spend 2 words too many to get clarification than guessing what someone writes is actually the same as I understand… Specially because not all English written here is explanatory by itself…
I know about the manual, it is not because it covers everything that I’m looking for extra explanations… 
(You should know me by now! 
)
http://www.soundonsound.com/sos/1994_articles/aug94/groundloops.html
http://en.wikipedia.org/wiki/Electromagnetic_compatibility
http://en.wikipedia.org/wiki/Common-mode_rejection_ratio
It is worth mentioning that you can have RF star earthing technique
Yes as lighting will take the path of least resistance to ground and in an area prone to lighting strikes perhaps the use of a de-ionizer unit at the very top of mast. Also if you use a timber pole as a mast then use a de-ionizer at the top and run a earth lead direct to ground earth, AP and Ptp equipment below the timber pole top conect to cabin earth ONLY.
If the cabin has a separate earth to the mast (metal as opposed to timber mast pole mast) then first have the earths tested as being effective then in the case of having two good earths, I would use the cabin earth for equipment in the cabin and on the mast use the mast earth , connect both ends of shielded cables from the AP to cabin earth and NOT connecting board earth to the mast earth so you will have two separate earths with no loops.
Ever wondered why there is inductive beads on AC power leads (50/60hz power leads) and inductive beads with de-coupling capacitors on DC rail voltages inside digital equipment running at MHZ or GHZ frequencies?
RF #Hash# suppression.
Not sure about electrical requirements elsewhere but here every connection from electricity supply company must have a certified earth point nearby before a connection will be allowed regardless of location.