[time-nuts] TBolt vs Twisted pairs

[Chris Albertson]

Yes, of corse twisted pairs can handle fast signals.  Gigabit Ethernet
over cat-5 cable is a good example.

What I meant when I wrote the below is that a bare t-bolt PPS output
can't drive 100 feet of cat-5 cable when the far end is connected to
to typical 74LSxxx t/l chip.  The pulse get smeared.

Yes, you could make it work easy enough.  Place some RS-422
transceivers on each end and it would work just fine.   But what I
said is the t-bolt is not able to drive the cable.   I tried it using
a spare pair in a cat-5 cable that was already installed in the walls
and up two floor.  It was very unreliable.



On Fri, Jul 20, 2012 at 4:31 PM, Hal Murray <hmurray at megapathdsl.net> wrote:
>
> (From a month ago.)
>
> albertson.chris at gmail.com said:
>> Take my word for it, the T-Bolt is not able to drive a 100 foot long twisted
>> pair cable
>
> I don't think that's quite the right way to phrase it.
>
> What type of twisted pair were you using and/or what sort of setup did you
> try?  How well did it work and/or what were you expecting?
>
> Yes, you may get much better results if you use differential
> drivers/receivers.  But that's if you have common mode problems.
>
> -----------
>
> I remember, many years ago, when I got an interesting lesson in this area.
> The difference between junk twisted pair and good stuff was impressive.
>
> We were installing a T microwave link.  On T1, a 1 is a pulse, a 0 is an
> absence of a pulse during a bit slot.  Pulses alternate polarity to keep a DC
> balance.  T1 is 1.544 megabits/second or 647 ns per bit.  I don't remember
> the details, but the ballpark is a 200 ns pulse has to get through.  So the
> rise time has to be in the ballpark of 20-50 ns.
>
> We had to go a few hundred feet.  My first try with a spool of whatever I
> found in the lab was a joke.  The spool of good stuff that we ordered worked
> fine.  I'm pretty sure the good-stuff was Belden Datalene but, again, it was
> a long time ago and I don't remember any details.  (I wonder if the cable is
> still there.)
>
> Does anybody have a good URL on lossy transmission lines?  Is there any
> obvious reason why twisted pairs should be different from coax?
>
> -----------
>
> Does anybody know what the PPS driver in a TBolt is?  I assume it's a typical
> CMOS logic family.  Is it one section or several in parallel?  What
> chip/family?  ...
>
> It clamps reflections.  In any case, it's not linear.
>
> ----------
>
> I collected some cable and tried a few experiments.
>
> Theses are all nominally 100 feet long.  I didn't measure any of the lengths.
>
> Coax:
>   http://www.megapathdsl.net/~hmurray/time-nuts/coax/Coax-20ns.png
>   http://www.megapathdsl.net/~hmurray/time-nuts/coax/Coax-100ns.png
>   http://www.megapathdsl.net/~hmurray/time-nuts/coax/Coax-1us.png
>
> Twisted Pair:
>   http://www.megapathdsl.net/~hmurray/time-nuts/coax/TP-20ns.png
>   http://www.megapathdsl.net/~hmurray/time-nuts/coax/TP-100ns.png
>   http://www.megapathdsl.net/~hmurray/time-nuts/coax/TP-1us.png
>
> The Belden 8723 is 52 ohms.  (I probably used 50 ohms.)  The other twisted
> pairs are 100 ohms.
>
> I'll have to try harder to find some really-junky twisted pair.
>
> The scope is the standard Rigol 100 MHz.  I had to work a bit on the setup to
> get clean pictures.  An early attempt with several feet of clipleads and such
> added a lot of garbage.
>
> The difference between 3 ft of brand-X RG-58 and 1 ft of good RG-58 from the
> TBolt to the scope is easy to see.  The brand-X isn't 50 ohms.
>
>
>
> --
> These are my opinions.  I hate spam.
>
>
>
>
> _______________________________________________
> time-nuts mailing list -- time-nuts at febo.com
> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.



-- 

Chris Albertson
Redondo Beach, California