19 Aug
2012
19 Aug
'12
12:23 a.m.
On 08/19/2012 02:04 AM, Mouse wrote:
Sure. Again, I hadn't realized we were speaking about a specific
limited-market device.
Yes, I'm sorry, I probably should've been more clear about that.
Perhaps I was too abbreviated. To put it more precisely, "someone
else's idea of `unclean' that's liberal enough to trip on just using a
USB<->serial chip on the UART pins of a SoC instead of going through
the whole USB VID dance can go hang, in the sense that I am not going
to care about it".
Understood.
Yes, that part IS evil. Suits got involved. :-(
FWIW, my current design (the FTDI was used in the previous two) has a
three-pin 0.1" header with TTL-level async serial for the CLI I wrote
into the firmware. ;)
-Dave
--
Dave McGuire, AK4HZ
New Kensington, PA
Like RS232, it is, strictly speaking. Also like
RS232, the standard is
widely ignored in at least some respects (I've got almost as many
male-to-female extender cables with host-port male on one end and
host-port female on the other as I do devices; such cables should not
exist at all). Also like RS232, the hardware is only part of the
problem; the software layers above are usually at least as important.
Agreed mostly...The USB standard isn't "creatively interpreted"
anywhere near as badly as RS232. It's not even in the same league in
that department. That's the only thing I don't like about RS232.
But then, 2/3/7 will get you talking to 98% of everything out there,
and for the stuff that needs more, one of those little DB25->LEDs->DB25
boxes makes short work of nearly everything else. I do very much like
RS232 in general; I'm quite pleased that it didn't "die off" as
everyone
(mostly suits with a vested interest in USB) predicted it would. It
only left the desktop market.
If you _must_ go USB,
I must. It is a
requirement from the customer, [...] "Unclean" can go hang - especially if [...]
"Unclean" will not "go hang" here. I am incredibly anal with
my
designs. Period. USB is not
evil.
Not most of it. Some aspects of it come pretty close, though, perhaps
most notably the vendor IDs; that paradigm could hardly have been
designed better even by deliberate intent to lock out everyone but
for-profit companies (and, maybe, a few of the richest hobbyists). The biggest problem I have with it is that it requires
a full-fledged
CPU or a comparable amount of custom silicon to speak at all - it is
not amenable to "throw together a breadboard lashup from discretes"
implementations. Of course, this is a purely personal point of view;
there's only one other person here I'm reasonably sure will agree with
it, that being tony.
Well, the FTDI chips (assuming you're talking about the remote end)
have that integrated, all nicely bundled up with very well-documented
I/O and behaviors. They really are a pleasure to use.
One other
important point here...there is insufficient space
available for a DB25 or DE9 connector on the last two boards I
designed.
That's a problem only if you think serial necessarily means one of
those two. Indeed, there's no technical reason you couldn't use
whatever connector you'd use for USB but put RS232 signals on its pins.
(It'd be a problem waiting to happen in various respects, but those are
reasons you might not want to do it, not reasons you couldn't if you
wanted to. I'd prefer some other connector of similar size...though,
again, that's speaking from a general-purpose-device viewpoint.)