2 May
2012
2 May
'12
11:57 a.m.
I doubt that was why they did it, though. I'm
sure it had much more
to do with simplifying device electronics so they could be implemented
on very cheap microcontrollers at low power consumption, rather than
Actually, thats' 2 more resons in favour of RS232 for me...
Firstly, USB _requires- a microcontroller really. The protocol is too
complciated to build from hardwired logic. Asynchronous RS232 can by done
using a microcontrolelr (either ther a built-in UART or bit-banged) _or
from simple logic chips if you so wish_. I guess I
don't like having to
have more computer power in the I/O chip than in the rest
of the machine :-)
Secondly, the fact that RS32 does not use TTL (or 3.3V logic) levels
means that most microcontrolelrs/FPGAs/etc (if not all) could not drive
it direcrlly. For a proper implementation of an RS232 interface you had
to have some sort of buffer, like a MAX232 chip. And that would bear the
brunt of any zaps on the interfce connector [1]. Such a chip is a lot easier
to get and replace than a programmed microcontrolker for a device that
you've never seen before.
I guess I don;'t want reliability... I'd rather hace somethign that fails
mroe often when the parts that fail are easy to get and replace than
soemthign that fails evey 5 years where I have major problems getting the IC>
[1] Although not always. The serial port in my HP9817 was defective when
I got the machine and I found that not only were the buffers damaged, so
was the 8250 serial chipe (yes, an 8250 in a 68K machine. Odd...). I am
not sure what had happened to the machine.
anything to do with the direction; after all, it's
half-duplex anyway.
There's only one pair, and it goes both ways.
Bear in mind that Ethernet has the same issues (we just call "null
modem" cables "crossover" cables instead), and people got similarly
Not the ethernet I use.... All cables are srraight and have one screen
and one cetnre core.
-tony