NeXT '040 Cubes

From: Pete Turnbull <pete_at_dunnington.u-net.com>
Date: Thu Mar 22 02:46:02 2001

On Mar 21, 21:22, Tony Duell wrote:
> >
> > On Mar 21, 10:07, MTPro_at_aol.com wrote:
> >
> > > << First, put a 50-ohm terminator on that connector, so the NeXT
thinks
> > it is connected to a live (but very small!) network. >>
> > >
> > > Is this something I can pick up at say Radio Shack? Does it looks
kind of
> > like a little metal cap that fits on the coaxial connector?
>
> Wait a second. Standard ethernet has a 50 Ohm terminator at each end of
> the cable. so the DC resistance between the core and the shield is 25
ohms.

Yes, but the standard method of terminating a single (isolated) point is to
use ONE terminator (maybe not for Vaxstations :-)).

> According to all the notes I have on how ethernet transceivers really
> work, the transmitter is a current source, which develops a voltage
> across this impedance. The receiver is a voltage sensor. If 2
> transmitters send at the same time you get a greater-than-normal voltage
> on the cable, which is what is detected as a collision.

> But of course, _1_ transmitter will develop twice the voltage across 50
> ohms as across 25 ohms. Which means having 1 terminator on the network
> (however short -- even just on the network connector) will generate
> 'collisions' all the time

I've not looked that closely at the internal circuitry of a transceiver,
but I can assure you that that's NOT what happens with any transceiver I've
tried. Having no terminator at all does indeed make many transceivers
believe there are constant collisions, but I've always found one to be
enough to prevent that, providing it's directly on the transceiver. The
official method for SGIs is to put a microtransceiver with a single
terminator on the AUI connector.

> You should use a non-inductive resistor here. Some cheap ready-made
> terminators are anything but...

Yes, and it should also be around 1/2 - 1 watt. Many terminators use 1/4
watt and get away with it, but the nominal voltage is 10V (and hence from
P=V^2/R, and allowing a 50% duty cycle, you want 1W).

> > types: those with a characteristic impedance of 50 ohms, used for test
> > equipment, thin Ethernet, etc; and those with a characteristic
impedance of
> > 75 ohms, mostly used for TV and video signals).
>
> IIRC, the central pin is of slightly different diameters (IIRC the 75 ohm
> one is thinner). If you put the thin-pinned type of plug onto the other
> type of socket, it won't always make good contact. If you do the reverse
> you can damage the socket contact.
>
> And of course putting mixed connectors together will cause a (small)
> mismatch which can be a bad thing.

Yes, they are different, and I *think* that's the right way round :-)

-- 
Pete						Peter Turnbull
						Network Manager
						Dept. of Computer Science
						University of York
Received on Thu Mar 22 2001 - 02:46:02 GMT

This archive was generated by hypermail 2.3.0 : Fri Oct 10 2014 - 23:34:04 BST