Author Topic: is it possible to damage Ethernet NIC with RS-232 signal  (Read 2336 times)

0 Members and 1 Guest are viewing this topic.

Offline m4rtin

  • Regular Contributor
  • *
  • Posts: 91
is it possible to damage Ethernet NIC with RS-232 signal
« on: September 28, 2015, 08:27:58 pm »
Let's say that I connect a DE-9 <-> RJ45 cable in a way that DE-9 connector is connested to (Cisco) network device serial(RS-232) port and RJ45 connector is connected to laptop NIC port. This should mean that laptop NIC might receive +3V to +15V or -3V to -15V signals on some pins while normal Ethernet on physical layer is between -2.5V to +2.5V. Is there a risk to damage NIC with 6 times higher voltage than nominal? Is it possible to damage serial port? Has anyone done that in practice?
 

Offline bitslice

  • Frequent Contributor
  • **
  • !
  • Posts: 493
  • Country: gb
Re: is it possible to damage Ethernet NIC with RS-232 signal
« Reply #1 on: September 28, 2015, 09:34:27 pm »
Ah, silly Cisco and their RJ-45 console port,
ethernet signals go via a transformer before they get near the electronics, so it will be OK (AFAIK)


 

Offline amyk

  • Super Contributor
  • ***
  • Posts: 7321
Re: is it possible to damage Ethernet NIC with RS-232 signal
« Reply #2 on: September 28, 2015, 11:58:33 pm »
Ah, silly Cisco and their RJ-45 console port,
ethernet signals go via a transformer before they get near the electronics, so it will be OK (AFAIK)
Yes there's a transformer (the "magnetics") but too much current will burn it out. Of course this is entirely dependent on how it's wired, as you may end up with the voltage harmlessly between two of the pairs (which is essentially what PoE does).
 

Online Kleinstein

  • Super Contributor
  • ***
  • Posts: 8852
  • Country: de
Re: is it possible to damage Ethernet NIC with RS-232 signal
« Reply #3 on: September 29, 2015, 04:43:51 pm »
RS232 outputs are current limited (at least they should be if by the standard, and all chips a know do limit the current). This way the RS232 is protected from shorts too. The current is likely not enough to damage the ethernet input transformer. Also input impedance on ethernet is much lower, so there will not be a high voltage.

So normaly there should not be any damage, but I have noch tested or in detail looked at the ethernet specs.
 

Offline suicidaleggroll

  • Super Contributor
  • ***
  • Posts: 1453
  • Country: us
Re: is it possible to damage Ethernet NIC with RS-232 signal
« Reply #4 on: September 29, 2015, 05:39:46 pm »
I agree that it should be alright.  The RS232 spec dictates that outputs should be able to drive into a short indefinitely, so that side should be fine, and the current limit on RS232 should be low enough to not damage the transformer on the RJ45 side, assuming the cable was even wired to pump current through one of the pairs.
 

Offline Cliff Matthews

  • Supporter
  • ****
  • Posts: 1871
  • Country: ca
    • General Repair and Support
Re: is it possible to damage Ethernet NIC with RS-232 signal
« Reply #5 on: September 29, 2015, 08:27:50 pm »
The only damage to Ethernet I've seen In 30+ years of IT from Buss-tag to Racal-Milgo sync modems to working at Gandalf-Infotron and Intel-Dialogic, is by lightning. The interface is magnetically coupled and often potted - it used to look like a fat IC chip

Now many of the magnetics are hidden inside deep RJ-45 jacks

Liitlefuse has quite a few good references http://www.littelfuse.com/~/media/electronics/design_guides/esd/littelfuse_ethernet_protection_design_guide.pdf.pdf
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf