Re: [uCsimm] HV capacitors on the LAN transformer

From: Daniel Haensse (
Date: Wed Aug 15 2001 - 09:44:26 EDT

Tom, I asked some of my electrical engineer friends.
A) The the transformer is symmetrical.
B) The wave is supposed to be symmetrical on the twisted pair
C) The common choke has always the same potential (Vplus + Vminus = 0) if the
wave is symetrical.

Now by AC grounding it, we get a supression of common noise that is Vplus +
Vminus <>0 because every change away from ground will be pulled to ground.
So this is not for EMI protection of other devices, it is for a better
communication when the environment is noisy. For this reason the capacitor
is on transmit as well as the receiver lines. The pickup of the noise comes
from the long twisted pair lines.
For DC protection a resistor may be added parallel to the capacitor to allow a
DC discharge to shield.
The rating is 2kV to even withstand pick-ups from an impact of a flash
somewhere near the twisted pair line (not into the line), which will be of a
common mode nature and very short in nature.
So use ceramic capacitors if you whant to make it rugged, or leave it away when
you use it on a small network.

Thank you for the question, very interesting. I've learned a lot.


Am Mit, 15 Aug 2001 schrieben Sie:
> Daniel Haensse wrote:
> >
> > Am Mit, 15 Aug 2001 schrieben Sie:
> > > Folks, who can answer this...
> > >
> > > There is a LAN transformer on the ucsimm (VALOR- PULSE
> > > ENGINEERING part PT4153S). On most Ethernet designs,
> > > there are 2 1kV 0.01 capacitors connected to this
> > > transformer on the ethernet side. On the ucsimm these
> > > transformer pins are left unconnected. Question: what
> > > these capacitors normally for, why they are missing
> > > and is there any penalty for not having them?
> > Those capacitors will discharge to gnd if the voltage is above 1kV.
> > So if the LAN transformer can't stand the voltage anymore, on the uCSimm
> > it will discharge to the uCSimm and may blow the ethernet controller or
> > anything else that the high voltage enjoys to destroy.
> While it is true that a sustained potential above 1KV will *eventually*
> cause the capacitor to fail.. The EMF pulse will have destroyed the
> electronics on the other side of the magnetic long before the cap will
> have failed. No, the cap voltage rating is so that it will withstand
> the voltages placed across the wiring by standards organizations such as
> the U.S. Underwriters Laboratory. I had a circuit that had to comply
> with a U/L spec. that required a sustained application of 600volts
> potential without catching on fire. The standard was for industrial
> wiring installations where it was possible that the wiring could be near
> 440volt three phase supply lines and may be shorted against that wiring
> (due to accident / incompetance).
> There are far more effective ways of achieving the protection of EMP
> (lightning) and electrostatic discharges that burning a cap out! These
> range in effectiveness from: spark gaps, MOV, and transil + transorb
> arrangements. You have to keep in mind that the EMF potential of a
> pulse is very high and of short duration (in most cases) and would need
> a very high speed clamping circuit to absorb the pulse. The clamping
> circuit response time is usually in the sub-nanosecond range (as in the
> case of a Transil). A garbage-variety capacitor offers no such
> protection.
> One of the most effective EMP clamping circuits that I have designed
> with is a Transil, resistor, and Transorb:
> Device >---------------\/\/\/\/\/\---------------> Line
> | |
> | |
> | |
> | |
> | |
> transil transorb
> | |
> | |
> | |
> | |
> ------- -------
> / / / / / /
> The Transil is a low voltage device, typically withing the ABSOLUTE MAX
> ratings of the device to be protected. The resistor is a 10 ohm device,
> and the Transorb (gas tube, similar to neon), has a firing voltage
> rating of 70 volts. The circuit works as: when a pulse arrives, the
> transil clamps the overshoot to ground causing a voltage drop across the
> resistor, the transorb will fire at the 70 volt level and continue to
> fire (conduct) as long as current is present. The worst case is that
> you will blow the resistor off the board! The transorb is capable of
> withstanding 1500Watts/second.
> Nope, the cap is not for overvoltage protection...
> Regards,
> TomW
> --
> Tom Walsh - WN3L - Embedded Systems Consultant
> "Windows? No thanks, I have work to do..."
> ----------------------------------------------------
> This message resent by the list server


Daniel Haensse Dipl. Ing. ETH Klinik fuer Neonatologie UniversitaetsSpital Zuerich Frauenklinikstr. 10 8091 Zuerich Switzerland Tel: ++41 / 1 / 255 5342 Fax: ++41 / 1 / 255 4442 email: This message resent by the list server

This archive was generated by hypermail 2b30 : Sun Apr 07 2002 - 00:01:43 EST