[Tfug] Tiny "serial console" --- Re: tfug Digest, Vol 127, Issue 13
ssl at email.arizona.edu
Mon Feb 17 20:19:07 MST 2014
Hmm, Maybe something like this for $50?
Also it looks like there are Dell D430's around for < $70 or so, 12"
On the other hand, if all I wanted was a terminal interface to a serial
connection, This would be more fun. I think we have one somewhere stored
with a DB-9 to DB 25 cable, and a keyboard, you would be set forever,
and no worries about whether your serial port was real or not.
Course I guess its not so tiny. :(
> Message: 1
> Date: Sun, 16 Feb 2014 16:30:25 -0700
> From: Bexley Hall <bexley401 at yahoo.com>
> To: Tucson Free Unix Group <tfug at tfug.org>
> Subject: Re: [Tfug] Tiny "serial console"
> Message-ID: <53014A11.8060704 at yahoo.com>
> Content-Type: text/plain; charset=ISO-8859-1; format=flowed
> Hi Harry,
> [PM me when you know what your "free" schedule looks like]
> On 2/16/2014 11:46 AM, Harry McGregor wrote:
>> The big problem is a lot of USB "rs232" adapters don't use proper voltage.
> Well, that's been the case with "real" serial ports for quite a while.
> The "5 volt only" transmitters with on-board charge pumps were already
> stressing the terms of the *original* EIA232 specification. But,
> manufacturers have continued to cut corners whenever they can
> rationalize (often, "to themselves"!) such a change.
> [I.e., moving to a 9 pin port already rationalized away much of the
> original intent of the RS232!]
>> They try and get away with +/- 5 volt instead of +/- 12 volt, and some
>> really crappy ones try and do +5 volt and 0 volt, instead of going negative.
> Or, won't drive a low enough line impedance (especially for longer
>> You can find some USB serial adapters that have a DC/DC converter in
>> them that do the full +/- 12 volt
> But they still have timing problems. I.e., many external "protocol
> converters" rely on characteristics of a "genuine" serial port to
> implement features (via software) that might otherwise require
> additional (though trivial!) hardware.
> E.g., when talking to a party-line (485) bus, often the modem control
> signals are used to control the "direction" of the bus transceiver
> in the (external to the serial port) adapter. I.e., "It is now time
> for *me* to use the bus so I want to enable the transmitter in the
> adapter by asserting <some_modem_control_signal>".
> When you put another protocol in the middle (i.e., USB) between the
> processor and the actual serial port hardware, you add uncertainty
> and timing skews/variations to this that the software (in the PC)
> isn't aware of -- it *thinks* it is talking to a real UART!
> Or, the driver fails to completely virtualize the UART so software
> that goes looking for a particular "UART" doesn't *find* it!
> [A fair bit of work went into making the "enhanced/deep-buffered"
> UARTs compatible with their "dumb" ancestors. That care does not
> seem to be present in the USB<->serial (or parallel) converters.]
>> If you want a relatively small laptop with a true RS232 port, look at a
>> toughbook CF-19, you can find some on ebay in the<$200 range.
> That's at least as large as the current "laptop+port expander"
> solution I'm using.
> I was really hoping for something along the lines of one of those
> little *tiny* (like hand-sized) computers that were en vogue a
> while back (useful for little more than email!). This isn't
> the sort of application where I need a nice, big, comfortable screen
> and keyboard. Rather, something that is used *just* long enough
> to figure out why something doesn't boot. Or, to adjust some
> configuration parameter, etc.
> I'll see what I can find and maybe cobble a true serial port into
> one "after the fact" (I only need to make *one*, after all! :> )
More information about the tfug