Good explanation by Lawerence. He touched on it, but to expand a little as it helps trouble shooting. The link light comes on when the receive side knows there is link. If you only get a light on one end of a link, you know which pair to look at, though it could be the port is bad.
Also, I believe the IEEE spec would say no active or passive devices on the wire other than connectors. What you are wanting to do will mess up the capacitance and inductance of the wire and that can cause data errors. Also, on many devices, disabling the port does not drop link. Link is a physical characteristic, not a logical one. Often if the port is power, it will link regardless of the administrative state of the port. Cameron Crum wrote: > That is the answer I was looking for. We have these multi-poe boards we > designed and had a bunch manufactured ... just passive devices that take > an input voltage and spread it across 9 ethernet ports with two of the > ports switchable between the input voltage and 12V. The signal side of > the ethernet ports go to mirrored ports on the other side of the board > to plug into a switch/router. I was thinking that if there was an easy > way to sense the connection, I could throw in an XOR chip and a few > small relays to make a cheap remote power cycle per port by simply > disabling the port on the switch or router on the signal side of the > board. Since the switch chip is involved, it becomes a much more complex > and expensive part. > > Cameron > > > On 3/11/2010 2:38 PM, Lawrence E. Bakst wrote: > >> The link LED and all other LEDs for Ethernet Jacks/Connections are driven by >> the Ethernet PHY chip or the Ethernet chip itself the PHY is integrated. >> >> Link is turned on by the PHY sensing the LIT (link integrity test) in >> 10BaseT which I believe has become part of the auto-negotiation protocol in >> later standards. This is part of the Layer-1 (Physical Later) protocol in >> the spec. >> http://en.wikipedia.org/wiki/Autonegotiation >> >> So to be clear it's not just a LED hooked up to one of the wire via a >> resister or some analog hack like that. The PHY knows that their is another >> PHY on the other side of the cable and if the PHY sees the other PHY it >> turns on the LINK light. PHYs often provide other lines to show collision, >> speed, and duplex and these can be tied into other individual LEDS or >> bi-color LEDs. >> >> If the link lights are on at both ends the connection is good. It still >> might be the case that a duplex mismatch or bad auto-speed negotiation could >> cause problems. Both of these problems show up from time to time, especially >> on older gear. For both cases the cure is often to fix the speed or duplex >> on one side and that prevents the auto-negotiation from failing. >> >> One cause of not getting a link light is that a MDI/MDI-X mismatch. Most >> newer chips have auto MDI/MDI-X which prevents the problem in most cases. >> >> leb >> >> At 12:52 PM -0500 3/11/10, Robert West wrote: >> >> >>> Yeah, but which circuit? The transmit, receive or maybe the unused pairs? >>> >>> That got me wondering also. >>> >>> Anyone know what pair triggers the light??? >>> >>> Bob- >>> >>> >>> -----Original Message----- >>> From: wireless-boun...@wispa.org [mailto:wireless-boun...@wispa.org] On >>> Behalf Of Justin Wilson >>> Sent: Thursday, March 11, 2010 12:15 PM >>> To: WISPA General List >>> Subject: Re: [WISPA] Ethernet LEDs >>> >>> Simple terms it's the completion of a circuit. >>> >>> --- >>> Justin Wilson<j...@mtin.net> >>> >>> On Mar 11, 2010, at 11:29 AM, Cameron Crum<cc...@dot11net.com> wrote: >>> >>> >>> >>>> This may be a little out there, but does anyone know what causes the >>>> "link" light to show on an ethernet jack when the cable is plugged in? >>>> Is it as simple as just attaching an led to one of the signal wires, >>>> or >>>> is there some logic in there. Just curious. >>>> >>>> >>>> --- >>>> --- >>>> --- >>>> --- >>>> -------------------------------------------------------------------- >>>> WISPA Wants You! Join today! >>>> http://signup.wispa.org/ >>>> --- >>>> --- >>>> --- >>>> --- >>>> -------------------------------------------------------------------- >>>> >>>> WISPA Wireless List: wireless@wispa.org >>>> >>>> Subscribe/Unsubscribe: >>>> http://lists.wispa.org/mailman/listinfo/wireless >>>> >>>> Archives: http://lists.wispa.org/pipermail/wireless/ >>>> >>>> >>> ---------------------------------------------------------------------------- >>> ---- >>> WISPA Wants You! Join today! >>> http://signup.wispa.org/ >>> ---------------------------------------------------------------------------- >>> ---- >>> >>> WISPA Wireless List: wireless@wispa.org >>> >>> Subscribe/Unsubscribe: >>> http://lists.wispa.org/mailman/listinfo/wireless >>> >>> Archives: http://lists.wispa.org/pipermail/wireless/ >>> >>> >>> >>> -------------------------------------------------------------------------------- >>> WISPA Wants You! Join today! >>> http://signup.wispa.org/ >>> -------------------------------------------------------------------------------- >>> >>> WISPA Wireless List: wireless@wispa.org >>> >>> Subscribe/Unsubscribe: >>> http://lists.wispa.org/mailman/listinfo/wireless >>> >>> Archives: http://lists.wispa.org/pipermail/wireless/ >>> >>> >> >> > > > > -------------------------------------------------------------------------------- > WISPA Wants You! Join today! > http://signup.wispa.org/ > -------------------------------------------------------------------------------- > > WISPA Wireless List: wireless@wispa.org > > Subscribe/Unsubscribe: > http://lists.wispa.org/mailman/listinfo/wireless > > Archives: http://lists.wispa.org/pipermail/wireless/ > > > -- Scott Reed Sr. Systems Engineer GAB Midwest 1-800-363-1544 x4000 Cell: 260-273-7239 -------------------------------------------------------------------------------- WISPA Wants You! Join today! http://signup.wispa.org/ -------------------------------------------------------------------------------- WISPA Wireless List: wireless@wispa.org Subscribe/Unsubscribe: http://lists.wispa.org/mailman/listinfo/wireless Archives: http://lists.wispa.org/pipermail/wireless/
