RJ45 splitter for Ethernet - possible ?
Bookmark this page:
 Yahoo!
 Google
 Windows Live
 del.icio.us
 digg
 Netscape
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Posted by News Reader on March 29, 2008, 2:15 pm
Please log in for more thread options >
> Well, technically ULF is seen at the receiver, but it is > too slow and neutralized by balanced signalling to make > any difference in the detection process. > >> If you were correct, we would not observe the minimum
>> separation distances from EMI sources that we do, when >> installing network cabling. >
> There are many other reasons for separation. First and foremost > is the dreaded electrical inspector who will often insist upon > it (for electrical "safety") and her word is law. Second, some > AC loads get very HF noisy (arc welders, certain motors and X-ray > machines are frequently cited) and inject high freq noise onto AC. > This noise can be troublesome. > >> We keep data cabling away from AC wiring (i.e.: 60 Hz), and
>> if we have to cross paths with AC, we ensure that we are >> perpendicular to the AC wiring, and never parallel to it, >> due to the effects of induction. >
> A clean AC draw (incandescent lighting) may produce some 60 > Hz induction, but it is harmless and easily rejected by the > receiver. Similarly the telephone ring 88V, 20Hz, unless > you have an antique phone with contacts on the bell armature. > Even then the protocols are robust and will retransmit. > >> Best Regards, News Reader
>
I'll yield to your superior knowledge of the subject matter, and adjust my top-posting behavior. > Please refrain from top-posting. The convention is
> quote-and-respond which makes following threads and > side-comments easier. > > -- Robert > Best Regards, News Reader | ||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Posted by P.Schuman on March 29, 2008, 2:42 pm
Please log in for more thread options News Reader wrote: > Robert Redelmeier wrote:
>>> If you think a high current, low frequency signal cannot
>>> impact the receiver, you are incorrect. >>
>> Well, technically ULF is seen at the receiver, but it is >> too slow and neutralized by balanced signalling to make >> any difference in the detection process. >> >>> If you were correct, we would not observe the minimum
>>> separation distances from EMI sources that we do, when >>> installing network cabling. >>
>> There are many other reasons for separation. First and foremost >> is the dreaded electrical inspector who will often insist upon >> it (for electrical "safety") and her word is law. Second, some >> AC loads get very HF noisy (arc welders, certain motors and X-ray >> machines are frequently cited) and inject high freq noise onto AC. >> This noise can be troublesome. >> >>> We keep data cabling away from AC wiring (i.e.: 60 Hz), and
>>> if we have to cross paths with AC, we ensure that we are >>> perpendicular to the AC wiring, and never parallel to it, >>> due to the effects of induction. >>
>> A clean AC draw (incandescent lighting) may produce some 60 >> Hz induction, but it is harmless and easily rejected by the >> receiver. Similarly the telephone ring 88V, 20Hz, unless >> you have an antique phone with contacts on the bell armature. >> Even then the protocols are robust and will retransmit. >> >>> Best Regards, News Reader
>>
>
> I'll yield to your superior knowledge of the subject matter, and > adjust my top-posting behavior. > >> Please refrain from top-posting. The convention is
>> quote-and-respond which makes following threads and >> side-comments easier. >> >> -- Robert >> >
> Best Regards, > News Reader | ||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Posted by glen herrmannsfeldt on March 29, 2008, 5:36 pm
Please log in for more thread options News Reader wrote:
(snip) > The issue is induction, and induction results from current, not voltage.
> If you think a high current, low frequency signal cannot impact the
> receiver, you are incorrect. V=L di/dt. It depends on current and frequency. Low di/dt also results in low inductive coupling. > If you were correct, we would not observe the minimum separation
> distances from EMI sources that we do, when installing network cabling. I believe ethernet can work with common mode 240VAC on the line. (Be careful around it, though.) I was going to test it someday but still haven't done it. > We keep data cabling away from AC wiring (i.e.: 60 Hz), and if we have
> to cross paths with AC, we ensure that we are perpendicular to the AC > wiring, and never parallel to it, due to the effects of induction. Differential signaling is pretty good at keeping the signals apart. Except for the case of arc welders (mentioned here in the past) it is rarely a problem. -- glen | ||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Posted by on March 29, 2008, 5:59 pm
Please log in for more thread options >
> The point is that Ethernet is a low voltage, low current implementation. So is POTS, save for ringing. > The issue is induction, and induction results from current, not voltage.
Yes, and ringing is low current (about 50mA, max). When it's on Cat5 cable, it's on a twisted pair as is the ethernet signal and the twists are different, which almost completely eliminates any inductive coupling. What little is left is almost entirely common mode, which the balanced ethernet receiver is almost entirely immune to anyway. > If you think a high current, low frequency signal cannot impact the
> receiver, you are incorrect. Have you never heard of a low-pass filter? > If you were correct, we would not observe the minimum separation
> distances from EMI sources that we do, when installing network cabling. > > We keep data cabling away from AC wiring (i.e.: 60 Hz), and if we have > to cross paths with AC, we ensure that we are perpendicular to the AC > wiring, and never parallel to it, due to the effects of induction. There are a number of reasons for that advice, but the 60Hz AC is not one of them. The noise concern is for switching transients from things like motors and arc welders, which can easily generate noise in the MHz range which could interfere with the ethernet signal. Note, however, that AC wiring is *much* higher current than POTS ringing, and it's not on twisted pair cable. Nonetheless, Cat5 has incredibly noise rejection. People who have tried to provoke errors from induced noise have found it quite difficult to do. -Larry Jones You should see me when I lose in real life! -- Calvin | ||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Posted by News Reader on March 29, 2008, 7:07 pm
Please log in for more thread options lawrence.jones@siemens.com wrote:
>> The point is that Ethernet is a low voltage, low current implementation.
>
> So is POTS, save for ringing. > >> The issue is induction, and induction results from current, not voltage.
>
> Yes, and ringing is low current (about 50mA, max). When it's on Cat5 > cable, it's on a twisted pair as is the ethernet signal and the twists > are different, which almost completely eliminates any inductive > coupling. What little is left is almost entirely common mode, which the > balanced ethernet receiver is almost entirely immune to anyway. > >> If you think a high current, low frequency signal cannot impact the
>> receiver, you are incorrect. >
> Have you never heard of a low-pass filter? > >> If you were correct, we would not observe the minimum separation
>> distances from EMI sources that we do, when installing network cabling. >> >> We keep data cabling away from AC wiring (i.e.: 60 Hz), and if we have >> to cross paths with AC, we ensure that we are perpendicular to the AC >> wiring, and never parallel to it, due to the effects of induction. >
> There are a number of reasons for that advice, but the 60Hz AC is not > one of them. The noise concern is for switching transients from things > like motors and arc welders, which can easily generate noise in the MHz > range which could interfere with the ethernet signal. Note, however, > that AC wiring is *much* higher current than POTS ringing, and it's not > on twisted pair cable. Nonetheless, Cat5 has incredibly noise > rejection. People who have tried to provoke errors from induced noise > have found it quite difficult to do. > > -Larry Jones > > You should see me when I lose in real life! -- Calvin Excellent response. I stand corrected. Best Regards, News Reader | ||||||||||||||||||||||||||||||||||||||||||||||||||||
| Similar Threads | Posted |
| RJ45 splitter for Ethernet - possible ? | March 28, 2008, 11:56 am |
| Ethernet: which RJ45 pins? | August 9, 2007, 9:49 am |
| best RJ45 crimper under $100? | August 14, 2005, 3:27 pm |
| 1 RJ45 to 4 BT converter? | August 31, 2005, 3:18 pm |
| RJ45 connector question | July 8, 2004, 2:55 pm |
| RJ45 to 110 Cross-Connects | August 24, 2004, 1:13 am |
| DB25-to-RJ45 Adapters | December 17, 2005, 8:04 pm |
| RJ45 connectors with a larger "gate"? | September 22, 2004, 10:22 pm |
| RJ45 connectors with a larger "gate"? | September 23, 2004, 9:59 pm |
| Corrosion in RJ11 / RJ45 Connectors | April 14, 2005, 5:51 am |
| panel mounted RJ45 socket? | August 9, 2005, 11:25 am |
| tool to ease RJ45 termination | April 17, 2007, 11:13 am |
| Wiring on RJ45 wall socket | July 8, 2007, 6:22 am |
| RJ45 Cat5e cables - male one one end, female (keystone) on other | July 21, 2005, 3:48 pm |
| DIY Ethernet | September 10, 2004, 8:57 pm |
>> impact the receiver, you are incorrect.