Electrically isolate your network, idea/suggestion/solution

[mav52]

1 hour ago, The Computer Audiophile said:

Yep. 

Wow , unreal.  Takes all kinds.

[jabbr]

10 hours ago, Pink Faun said:

As you point out correctly all ethernet equipment has to include a certain level of galvanic isolation from the cable. But as you also point out there’re differences between the transformers used (the more expensive ones). Mostly used ethernet ports has integrated transformers in a very small space. With discrete ethernet transformer (like the LAN isolator) there’s less crosstalk and we can choosing modules with better common mode rejection figures than the integrated connector module.

 

 

Thanks for clarifying ... I have s quibble with the term “galvanic isolation” in general because interruption the direct electrical connection prevents only “DC” current flow whereas frequencies > 0hz conduct via parasitic capacitance and inductance (of course moreso at higher frequencies). Obviously 1Ghz + Ethernet signals pass readily across the transformers, so really the goal is to minimize common mode noise conduction.

 

I know y’all know this but not everyone here understands.

[Ralf11]

well, the next logical question is how best to suppress HF noise transmission via parasitic capacitance/inductance...

 

 

[jabbr]

3 hours ago, Ralf11 said:

well, the next logical question is how best to suppress HF noise transmission via parasitic capacitance/inductance...

 

 

 

There are textbooks written in this topic including : High Speed Digital Design: A Handbook of Black Magic https://www.amazon.com/dp/0133957241/ref=cm_sw_r_cp_api_i_DqOSCb0HX7W17 and https://www.amazon.com/dp/013084408X/ref=cm_sw_r_cp_api_i_dxOSCbEQQW9TH 

and Electromagnetic Compatibility Engineering https://www.amazon.com/dp/0470189304/ref=cm_sw_r_cp_api_i_czOSCb2FX6QRC

and Signal and Power Integrity - Simplified (3rd Edition) (Prentice Hall PTR Signal Integrity Library) https://www.amazon.com/dp/013451341X/ref=cm_sw_r_cp_api_i_FBOSCbFKF9401

 

... but basically reduce common and differential mode noise 👍

[fas42]

From personal experience, it is mighty hard to eliminate these sort of interference effects, completely - the rule is, the closer the system gets to full transparency, the more easily heard are the degrading effects - when they are there, versus when they're not there. In part, it becomes a game of diminishing returns, as one tries to erect another barrier - or one can luck on equipment that has a 'natural' resistance to these factors, your job is then easier.

 

One workaround is to shut down every bit of electrical activity in the vicinity; tough, if a neighbour is spewing out noise ...

[jabbr]

So above reading is the long answer — the short answer is first design the boards and interconnects to generate less HF noise, and then filter both differential and common noise using appropriate LRC networks, common mode chokes etc. You need to consider the shielding on cable’s as an extension of the ground plane of the transmitter etc.

 

Of course with fiberoptic Ethernet there is no parasitic capacitance nor inductance across the cable. 

[marce]

8 hours ago, jabbr said:

 

There are textbooks written in this topic including : High Speed Digital Design: A Handbook of Black Magic https://www.amazon.com/dp/0133957241/ref=cm_sw_r_cp_api_i_DqOSCb0HX7W17 and https://www.amazon.com/dp/013084408X/ref=cm_sw_r_cp_api_i_dxOSCbEQQW9TH 

and Electromagnetic Compatibility Engineering https://www.amazon.com/dp/0470189304/ref=cm_sw_r_cp_api_i_czOSCb2FX6QRC

and Signal and Power Integrity - Simplified (3rd Edition) (Prentice Hall PTR Signal Integrity Library) https://www.amazon.com/dp/013451341X/ref=cm_sw_r_cp_api_i_FBOSCbFKF9401

 

... but basically reduce common and differential mode noise 👍

Got all 3 books on the bookshelf in my office, also recommend:

https://www.amazon.com/Fields-Electronics-Understanding-Using-Physics/dp/0471222909

and anything RF/microwave.