[ntp:questions] Oddities in termination of cable from gps18.

David Woolley david at ex.djwhome.demon.invalid
Fri Feb 24 00:34:05 UTC 2012


Chris Albertson wrote:

>  I just looked up the most common RS232 chip, the MAX232.   The
>  receiver section of the MAX232 sees anything over 3.5 volts as "high"
>  and  anything under 0.4 volts as low

The common chips from the seventies, 1489 and 1489A, had adjustable 
thresholds, but the default threshold region started at about +0.8V. 
The 1489 hysteresis region extended to about 1 volt, and the 1489A to 
just under 2 volts.  (The lowest specified negative going switching 
point for both was 0.75V and the highest specified positive going one 
was 1.5V for the 1489 and  2.25V for the 1489A.)

They have a lot of hysteresis; they are basically Schmidt triggers. 
There is no input rate of rise constraint, so I suspect that they had a 
very small meta-stable region.

As I noted elsewhere, the default transition region is deliberately all 
above zero volts.

> 
>  The driver section sends at least a +/-5 volts, +/-7 is typical.
> 
>  I think the problem is that with TTL you typically have a high value
>  pull-up resister  then we find someone trying to "terminate" the line
> with a 200R or less resistor.   All the "terminator" does is form
>  voltage divider with the pull up resistor.

Typically you have CMOS masquerading as TTL, these days!

Incidentally, for non-line driver, real, TTL devices, I seem to remember 
the maximum recommended line length was about 10 inches.

> 
>  Other RS232 chips might be different but the MAX232 is common

The MAX232 should only be common in single rail environments.  It used 
to be a lot more expensive than the 1488/9 chips, which need dual rails 
for the driver, although the receiver uses a single, 5V, rail.



More information about the questions mailing list