[ntp:questions] Very large jitter and offsets on GPS ref clock after upgrade to "p5"

unruh unruh at invalid.ca
Mon Jan 9 21:09:58 UTC 2012


On 2012-01-09, E-Mail Sent to this address will be added to the BlackLists <Null at BlackList.Anitech-Systems.invalid> wrote:
> David Woolley wrote:
>> Mike S wrote:
>>> unruh wrote:
>>>>
>>>> Make sure you properly terminate the cable with 100 ohms.
>>>>  Otherwise you will get reflections back and forth along
>>>>   the cable, and the cable will slowly fill with charge,
>>>>   broadening the pulse by a huge amount.
>>>>  It may also be that the signal is decreasing due to
>>>>   resistance in the line.
>>>
>>> The proper termination for EIA-232 is 3000-7000 ohms.
>>
>> EIA-232 deliberately mis-terminates in order to make the line
>> capacitive, and slow down the rise times.
>> It also has a 50 foot limit, to limit the capacitance and rise time.
>
> The RS-232 standard (currently TIA/EIA-232-F)
>  {normalized with ITU-T V.24, V.28, and ISO/IEC 2110}
>
><BlockQuote> Electrical Specification:
> The maximum cable length originally was defined in RS-232-C
>  as 15 meters;  However, this has been revised in EIA-232-D
>  and TIA/EIA-232-E and is now specified more correctly as a
>  maximum capacitive load of 2500 pF.  This equates to about
>  15 to 20 meters of line length, depending on cable capacitance..
></BlockQuote>
>
> I'd have to dig up the specs, but Cat5e is usually around
>  mid 50's pf per meter (@ some frequency).
>
> With +/- 12VDC RS232 drivers, I've done 800' and 1000' runs
>  of Cat5e @ 19.2Kbps without issues.
>
>  {On the other hand, on cables designed for RS232 decades before Cat5e,
>    I've had issues at one tenth that far, (~100pf to 200pf / meter)
>    at 19.2Kbps.

With proper termination, the capacitance of the cable should be a
non-issue. The pulse simple tranvels down the cable and disappears into
the resistor. That of course assumes that the driver can supply the
current needed for that voltage into 100 ohms for whatever time is
needed. (ie for 12V, it would be 120mA continuous, or for 5V, 50mA.) 

The specs sound like they are not for transmission cable, but that the
signals are slow enough that the cable can be treated as simply a
capacitor that gets charged up, more and more slowly. This is typical of
an unterminated cable (High input impedance on the receiving end) in
which case signals keep bouncing back and forth along the cable,
interfering with each other and messing up the signal.


>



More information about the questions mailing list