🚀 go-pugleaf

In article <15OCT199216130132@rover.uchicago.edu> frank@rover.uchicago.edu writes:
>In article <93243@netnews.upenn.edu>, george@mech.seas.upenn.edu writes...
>> 
>>'feels like' a pretty trivial thing to make, with the box and
>>connection ports being the most costly components.
>
>	1/ You really need a micro to do the handshaking, parity calculations
>	   handling number of data bits, etc.

In general, yes.

>	2/ You need a prom to go with the micro. (Forget about the programmer
>	   and debug/programming costs for now.)

Well, unless you're going to produce 100000 it's cheaper that way.  For
large quantites, mask ROMs right on your CPU are the way to go.

>	3/ You need several RS232<>TTL converters.

One Max-232 should do fine.

>	4/ You need a UART for the serial-parallel part.

Built in most standard microcontrollers these day.

>	5/ You need 5v to +/-12 v converter chip.

Built into the Max 232.

>	6/ You need a baud rate generator, (for your serial/parallel part.)

Again in the microcontroller.

>	7/ You need a circuit board.
>	8/ You need 1 DB25 and 	1 centronics socket
>	9/ You need misc wire-wrap sockets, or circuit board capabilities.

Yes.

>	Add it all up, and (quantity one,) you'll probably pay around $100
>	for just the parts.
>
>	And THEN YOU STILL HAVE TO BUILD IT.

If you pick your parts effiently, it'll probably cost you somewhere around
$50.  Now, if you're a _company_ you also have employees to pay and a building
to pay for.  Finally, you want to make a profit... so $100 is a fairly
reasonable price for such a device.  I've seen a few that are cheaper,
but not by much.

>	As an aside, I've been doing stuff like this for years, and am still
>	amazed that:

I certainly hope that you're aware of what I've said above.  Otherwise, you're
making inefficient (price wise, and power wise as well) converters.  (At least
given the capabilities of most $100 converters -- not ridiculously fast
baud rates, and not having to worry about more than 2 lines each way).

>	A/ People say "it only takes a few parts,' and when they actuall cost
>	   up ALL THE PARTS, they slap their foreheads and say "I never thought

Yes, people don't think of costs like the circuit board, the box to put it
in, all the resistors and capacitors, etc.

>	B/ Nobody ever counts the cost of their own labor. My own rule of
>	   thumb is that semi-custom, kit, hand-made etc the labor cost is
>	   at least as much as the parts, and probably is closer to 2times
>	   for projects at the $100 level.

Quite true.  Of course, many people have lots of time of their hands anyway...

>	As an aside, why didn't you spend a little more and get the right
>	kind of printer? ;-)

Yeah, adding a serial port to a printer when the printer is being designed
is a lot cheaper than making a general purpose converter later!  (Since
you already have you CPU, PROM, power supplies, casing, etc., etc., etc.!)

>   Frank R. Borger - Physicist    __       Internet:  Frank@rover.uchicago.edu
>Michael Reese - Univ. of Chicago |___      Phone : 312-791-8075 fax : 567-7455
>  Center for Radiation Therapy   | |_) _   I guess I must have had New England
>                                   | \|_)  ancestors away back and inherited
>    "Birthplace of Softball"          |_)  some of their staunch and rugged
>					   fear of the police. - O. Henry

						---Joel Kolstad