David Duffy (AVD) wrote:
> Michael Watterson wrote:
>   =

>> Ruben J=F6nsson wrote:
>>   =

>>     =

>>>> We make DMX LED controllers - the DMX front-end is a 65lbc179 that =

>>>> feeds a fast opto-isolator.  The front end is powered with an =

>>>> isolated DC-DC converter.
>>>>
>>>> The 'lbc179 ** REQUIRES ** that you constrain its input signal to =

>>>> within its common-mode range.  The easiest way to do that is to =

>>>> require that the installer uses shielded cable to transport the DMX =

>>>> signal, and to connect the drain wire at both ends.
>>>>
>>>> We don't have ground-loop problems because the DMX front-end is =

>>>> completely isolated from the rest of our circuitry.
>>>>
>>>> dwayne   =

>>>>       =

>>>>         =

>>> So the shield is not connected to chassie/earth?
>>>  =

>>>     =

>>>       =

>> It would obliviate the point of isolated DMX if it was.
>>
>> This is the point I'm making. In fact if you are going between buildings =

>> it might even be illegal in some cases to interconnect the earths.
>>   =

>>     =

>
> But that's the point we're making.  You're not interconnecting the =

> earths. You're connecting the common reference.  That reference may or =

> may not be mains earth.  There's no requirement for it to be earth - it =

> just needs to be a common reference.
>
> In the case of isolated RS485, do you have a common reference (often the =

> shield)? Yes!  Is it mains earth? Not if both ends are truly isolated.
>
> As long as the I/O of only one device is earthed and the rest are =

> isolated, there's no problem. Ideally they are all isolated from mains =

> earth so that a single fault (common contacting earth) will have no ill =

> effect.
>
> So, we're all agreed that differential signals (RS485 for example) =

> travelling between places of differing earth potentials should be =

> isolated.  That isolation doesn't negate the need for the common =

> reference conductor, it just means the I/O sections are floated with =

> respect to earth.
> David...
>
>   =

However if both ends are isolated, you will find that a centre tapped to =

local 0V termination at both ends and "earthing" screen at one end  will =

work. You don't inherently need a 3rd connection. The 3rd connection, if =

used carries no signal,  it's just a primitive (and unsafe if the gear =

is on different supplies without isolated I/O)  method of limiting =

common mode voltage.

Capacitance is potentially higher with a screen, but only if the twisted =

pair is otherwise in free air. Assuming a correct impedance of =

transmission line correct driving and termination, then the capacitance =

per meter affects the distance. For longer distance you need fatter =

spacing between twisted pair and screen and also fatter core copper and =

fatter twisted pair insulation to reduce the loss.

If  everything is done properly a screened twisted pair is superior for =

long distance as there is less crosstalk with other cables and less =

noise/radiation. The  reason phone lines and typical Cat5e isn't =

screened is cost. Cat5 bundled with other function cables or  outside =

cable trays/trunking  (earthed)  is  best done with STP rather than USP. =

Less noise pickup, radiation to Radios and crosstalk. Ideally only earth =

screen at "central" patch panel, never both ends.



-- =

http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist