"Peter L. Peres" wrote:
>
> >supplier
>
> I don't have one off of my head, because I normally buy off the shelf or
> from catalogs. But, by looking at the ethernet card I have in my hands
> now, I can see a large DIP transformer hybrid directly connected to the RJ
> plug. So practically ALL ethernet cards have a transformer coupled output.

I don't know of any Ethernet cards that don't have a transformer-
without a transformer (or capacitor isolation, at least), an ethernet
card would never be able to meet the common mode noise rejection
specifications.  The transformer is generally capacitively coupled to
ground- I've seen 2KV rated capacitors in that application.

> Try to find the exact make of the cards you are using and contact the
> manufacturer to know the specs. He will be able to tell you how much their
> magnetics insulate. The hybrid on my card is made by Delta (it's a cheap
> NE2000 clone card). If the manufacturer does not cooperate, try another. I
> am quite sure that they have heard this before.
>
> The ethernet signals for BNC are a very nice rectangular waveform (for a
> short network at least) with high = GND and low = -2V (load is about 5
> cards). The balanced RJ type of signals are twice that if I'm not wrong.

100BASE-TX should be +/- 1.00 V into a 100 ohm load, if not, they aren't
compliant with the IEEE specification.  I'm pretty sure 10BASE-T is the
same way.

According to the IEEE 802.3 specification (a 1200 page tome I have on my
desk for my day job) the isolation requirements of the 10BASE-T physical
layer are as follows: (transcribed by me, so any typos are my fault).

(IEEE 802.3-1998)
14.3.1.1 Isolation requirement

The MAU (medium attachment unit) shall provide isolation between the DTE
Physical Layer circuits including frame ground and all MDI (Media
Dependent
Interconnect) leads inclusing those not used by 10BASE-T.  The
electrical
separation shall withstand at least one of the following electrical
strength
tests.

a) 1500 V rms at 50 Hz to 60 Hz for 60s applied as specified in Section
5.3.2 of IEC 60950: 1991.
b) 2250 Vdc for 60s applies as specified in Section 5.3.2 of IEC 60950:
1991.
c) A sequence of ten 2400 V impulses of alternating polarity, applied at
intervals of not less than 1 s.  The shape of the impulses shall be
1.2/50
us (1.2 us virtual front time, 50 us virtual time of half value) as
defined
in IEC 60060.

There shall be no insulation breakdown, as defined in Section 5.3.2 of
IEC
60950: 1991, during the test.  The resistance after the test shall be at
least 2 Mohm, measured at 500 Vdc.

---

I don't think it is required to work afterwards, it just can't present
dangerous voltages to the operator during/after the test.

Matt Bennett

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.