AVR-Chat

David Kelly wrote:
> The specified range is 3 to 30, positive and negative, with -3 to +3  
> being the "keep out zone".
>
> +-5 and +-12 used to be very common voltages to have inside a  
> computer as Once Upon A Time common DRAM memory required negative  
> voltage bias to operate.
>
> The 1488 driver dropped about 2 volts so when driven with +-12 its  
> output would swing +-10.
>
> +-25 was much harder to drive a clean signal any interesting distance  
> so the RS-232 spec is limited to 19,200 and 50'. Lower voltages work  
> longer distances and higher data rates which prompted the move to LVD  
> drivers.

Well, the reason lower voltages work longer distances is usually because 
of the lower threshold voltages specified, not any magic property of low 
voltages. Every signal takes a finite amount of time to "slew" through 
from one voltage to another, the more voltage, the longer it takes to 
get there.  Ask a driver to slew from -12V to +12V and it takes about 5 
times longer than an equivalent driver would take to go from 0V to +5V. 
BTW, RS-232 drivers have a built-in series drive resistance of 300 ohm 
(per the standard) which definitely prevents it from driving long lines 
with any real punch or ability to slew rapidly.

The reality is that not only the old 1489 receivers but also the newer 
MAX232 genre all have switching thresholds which are non-RS-232. The 
1489's input is a grounded emitter bipolar npn with a 4K series base 
resistor and a 10K base pulldown. There is also a reverse diode from 
base to emitter but this tends to be a by-product of integrating 
circuits on a common semiconductor substrate (hence the term IC). What 
does this all mean? The turn-on threshold of such a configuration is 
above 0.6V or so and needs a bit higher to saturate the transistor so 
that it can turn on fully. Go below this voltage and the transistor will 
be turned-off. Simple as that, no rocket science in this. So unless you 
are driving "relatively" long lines at "relatively" higher baud rates 
(say 20 metres plus at above 19.2K) then there is absolutely no problem 
with driving a common RS-232 receiver with 0 to +5V (even +3.3V dare I say).

Not standard you say? There is nothing standard about RS-232, even in 
it's heyday of the 60's and perhaps 70's (they didn't have much else 
then). We may have to make a professional product RS-232 compliant more 
out of legacy then out of compatibility, but that does not make it any 
better IMHO than a low-voltage interface. If I ever want distance, 
speed, noise-immunity etc then RS-232 is definitely not the choice 
anyway. I only use it for simple hook-ups to equipment such as PCs that 
usually span no more than a few metres. If you don't think that 0 to +5V 
is suitable for this distance then you had better design RS-232 drivers 
into your PS/2 keyboard and mouse and so on and so on, good luck.

Anyone who can quote an example of where 0 to +5V wouldn't work with 
some equipment is probably working in a deep dark basement at the 
Smithsonian Institute where the paper-tape reader still requires all 25 
blessed-blinking-RS232-%^*#! pins hooked-up as well :)

Anyway, why argue, use what works.

my2cents
*Peter*