[lug] testing for straight versus cross-over patch cables?

[D. Stimits]

J. Wayde Allen wrote:
> On Mon, 12 Aug 2002, Mr Viggy wrote:
> 
> 
>>Heh.  What annoys me is the IT guy that told me I could easily tap into 
>>the thinnet (RG-58) in the cieling to connect up an HP plotter.  Funny 
>>thing, after connecting it up, half the machines in the marketing dept 
>>go offline.  Seems that the 3 extra feet of cable I added was just too much!
> 
> 
> How exactly did you go about doing this?  The reason I'm asking is that
> the coax based (RG-58) thin net can't simply be tapped.  It has to loop
> through your computer and then back to the other machines downstream from
> you.  Basically, you took the downstream computers off-line when you cut
> the cable.  If you tried simply putting a "T" in the line and running a
> three foot extension to your network card, you effectively created a
> coaxial tuning stub on the line that modified the impedance of the line.  
> This can cause standing waves that will have rather unpredictable results
> (some machines keep working while those in the RF null fail).

Now you know why they suck at higher speeds. That is exactly the way it 
is done.

My guess:
The signal itself is probably *very far* from a sine wave...with square 
waves (infinite sum of odd harmonics) and complex waveforms you can't 
tune it anyway, high Q tuning destroys your ability to carry non-sine 
waves, it has to be very low Q. You add more pure resistance, and accept 
resistive losses and noise (in return for lower distortions on non-sine 
waves).

D. Stimits, stimits AT attbi.com

> 
> The tap would need to be something like:
> 
>                 
>    Thin net <--------------| |------------------> Thin net
>                            | |
>                            | |
>                            | |
>                             T
>                      ---------------
>                     | Your computer |
>                      ---------------
> 
> - Wayde
>   (wallen at lug.boulder.co.us)
> 
> _______________________________________________
> Web Page:  http://lug.boulder.co.us
> Mailing List: http://lists.lug.boulder.co.us/mailman/listinfo/lug
> Join us on IRC: lug.boulder.co.us port=6667 channel=#colug
>