0
\$\begingroup\$

I'm trying to transmit Ethernet through relays. I'm using relays because I need them to switch to another signal path, and I'm specifically going to use relays and not other switching devices so that things are fail-safe and still provide a path when my device controlling the relays is off. So, I wonder if anybody out here knows can I safely transmit 100 Mb/s Ethernet through relays? I'm a bit worried because the spec says that Ethernet must be transferred through 100 Ohm impedance (unshielded twisted pair).

  1. Will the signal be ok?
  2. Are there timing issues or any other issues with doing this?
\$\endgroup\$
  • 6
    \$\begingroup\$ What is the problem you're trying to solve? I bet there is a more reliable solution to your challenge. \$\endgroup\$ – jippie Oct 18 '12 at 17:15
  • 1
    \$\begingroup\$ Safety is not the issue. Data integrity and reliability are. \$\endgroup\$ – jippie Oct 18 '12 at 17:29
  • \$\begingroup\$ Maybe add the question whether you should be using one of those "HF relays". \$\endgroup\$ – AndreKR Oct 18 '12 at 18:38
  • 2
    \$\begingroup\$ That is a bad idea. It might work like suggested but it most likley will not work! Just get a managed switch. You can tell it were to route data. IE Port1 incoming goes out Port4 (mornings) Port2 (evenings) or if Port1/2 Fail use fail over Port 3. 4Port Cisco ones about 70quid.. on ebay even cheaper. Seriously.. lan over relays = serious issues. If you still insist you need to shield the relays and use as small ones as possible. Like SMD size on a PCB. Not those large ones used 1980's Ford cars. \$\endgroup\$ – Piotr Kula Oct 24 '12 at 14:27
  • 1
    \$\begingroup\$ This sounds so Token-Ring... passive hubs full of relays, a DC signal on the cable would pick up a relay to include the workstation in the ring. But I quite agree this smells like an XY problem. \$\endgroup\$ – Harper - Reinstate Monica Dec 18 '19 at 22:31
3
\$\begingroup\$

Ethernet is surprisingly robust and you can use surprisingly bad cables before it starts to fail. So to send data through would probably work quite ok.

But when you switch the relays you would end up with a disconnect and a reconnect and those could cause the connected devices to behave strange since you on a logical level "moved the cable" from one device into another in a ms or so. And I am not sure how they would react to that...

But I get the feeling that there is another solution to your problem that will work better, even if this probably will work.

| improve this answer | |
\$\endgroup\$
2
\$\begingroup\$

For the first question about safely, the answer is yes. If you are using Power over Ethernet, check the voltage ratings on the relay.

It will likely work if the relay is at one end or the other. You can check to see if autonegotiation passed or it had to back down to 10 Mb/s.

| improve this answer | |
\$\endgroup\$
2
\$\begingroup\$

(read this as an addendum to @Johan's answer.)

If the very short switchover time turns out to be a problem you can just add a second relay layer with a time delay only to the actively switched side (not the NC one). This way you can chose the switchover time freely.

Keep in mind that you're adding another bit of non-twisted "wire" to the Ethernet cable, thus worsening the transmission capabilities a bit more.

Add a diode, a big Capacitor (and some other components) and even the switch back to the NC state in case of a power failure can be delayed.

schematic

simulate this circuit – Schematic created using CircuitLab

| improve this answer | |
\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.