Schematic of a GM ALDL -> PC converter

[Darrell Norquay]

At 01:39 PM 12/3/96 -0500, SMD wrote:

>Used to be that -5 to +5 (or +-7) was invalid for RS232 inputs back when
>everyboy went from -12 to +12.  (The invalid area was a hysteresis band so
>you HAD to cross both boundries to toggle the bit)  Now with all the 3.3V
>chips around, they use a charge pump to get almost +6.6V and -6.6 (usually
>+6.2 and -6.0 in my exp) and the hysteresis band is reduced to something
>like +-2V.  I believe it is still bipolar, so if you only go from 0V to 5V
>the bit may get 'stuck' in one direction.  Hope this helps...  You might
>also have a serial interface that doesn't truly follow the RS232 spec and
>toggles around 0.7-2.7V though.  

The true RS232C standard specifies +/- 15 volt signals.  This was sortof 
relaxed to +/- 12 volts when personal computers started becoming popular.
The signal was supposed to have a "dead band" between +/- 3 volts, which 
allowed for lots of signal loss over long wires.  There is actually another 
standard, (I forget the RS number), which is a modified RS232 that only
swings +/- 5V.  This was (and is) used a lot on laptops, etc.  It was still
called an RS232 port, however, (even though it's not) because if your cables
weren't too long, it would still talk to a true RS232C input.  I don't think
this will work with a single-ended 0-5v signal, since it must swing
significantly negative to get it to change from space to mark.  

Any of the "backwards" converters using a MAX232 presented here recently
should work nicely, though.  Since "mark" is negative, and "space" is
positive, an additional inverter may be required to get the proper
signal into the serial port of a PC.



regards
dn
dnorquay at iul-ccs.com