Newsgroups: comp.robotics
Path: brunix!news.Brown.EDU!noc.near.net!howland.reston.ans.net!spool.mu.edu!sdd.hp.com!decwrl!decwrl!csus.edu!netcom.com!jfox
From: jfox@netcom.com (Jeff Fox)
Subject: Re: HELP: Simple A/D converter?
Message-ID: <jfoxCD0H77.8A@netcom.com>
Summary: Cheap A/D
Sender: jfox@netcom.com (Jeff Fox)
Organization: Netcom Online Communications Services (408-241-9760 login: guest)
References: <CD09zt.MK2@news2.cis.umn.edu>
Date: Wed, 8 Sep 1993 01:23:31 GMT
Lines: 54

In article <CD09zt.MK2@news2.cis.umn.edu> horton@molbio.cbs.umn.edu
 (Robert Horton) writes:
>
>Dear Hardware Hackin' Gurus of Netland;
>
>I need to build/buy/get a board or (preferably) serial or (maybe even) paralel
>device that will allow me to monitor a thermocouple with my PC. The only
>things my computer-savvy associates here have been able to find in catalogs
>cost about $300 (for the d/a converter - I'd still need the thermocouple
>on top of that). This seems *way* expensive to me, and it's also overkill, 
>these boards will monitor multiple signals, and I only need one. Some day I'll
>have my very own government grant, and I'll be able to buy these kinds of 
>things with YOUR tax dollars, but for now I wanna make this thing with
>MY OWN MONEY, so I'm very sensitive to cost :)  The good part is that if I
>find a cheap and easy way to do this, I'll tell all of my scientist buddies,
>and these guys DO spend A LOT of your tax dollars. 
>
>The temperatures I need to monitor are in the range of 25 to 100oC, and the
>measurement must be to within 0.5oC (more precise is probably better). I'd
>like to be able to record the temperature every half second or so. The main
>requirements beyond this are that the device be simple enough that your
>average biochemist can build it him/herself, and cheap enough that he/she
>would be willing to risk a few (?) dollars of precious Reagent Money
>(which is much more valuable than Equipment Money) on parts for a gizmo
>they have to make themselves.
>
>Any suggestions, hints, comments, clues, tips, or ideas would be most welcome.
>
>Sincerely, 
>Robert M. Horton, Ph.D.
>
>Bob Horton            /\ "Crash programs fail because of the theory that
>U. of Minnesota, CBS  || with nine women pregnant you get a baby a month" 
>1479 Gortner Ave.    /||\   -Werner von Braun.  Disclaimer:"Bob who?"
>St. Paul, MN 55108    ^^   horton@molbio.cbs.umn.edu/(612) 624-3790

Dear Bob,
The cheapest A/D I know of is constructed using 1 bit of I/O from your
parallel port, or your game port.  Set it high and charge a capacitor,
then stop charging and let the capacitor discharge through your sensor.
Simply time the discharge from 1 to 0 to determine the resistance of
the sensor.  I dont know how linear your thermocouple is, but you will 
have to calibrate your system and adjust for non linear measured values
from the sensor anyway even with an expensive A/D.   The range of 75
degress in half degrees is only 150, so 8 bits of resolution may be
enough.
Anyway the cost is only a few cents.  I think there is a very wide range
for A/D.  $5 for a A/D with serial output from Radio Shack.  .. all the
way up to very expensive.  But you are talking VERY slow sample rate,
and very low resolution, so almost anything will do.  There is no need
to spend more taxpayes $ than you must, (I hope.) :-)

Jeff Fox
jfox@netcom.com
