(Comp.sys.handhelds)
Item: 3945 by _jwtrav at hpcvbbs.cv.hp.com
Author: [John Wettroth]
  Subj: RS232 vs. IR Battery Life
  Date: Sat Sep 07 1991

     I have recently done some calculations of how much power the IR
Transceiver draws versus the RS-232 Transceiver in the 48.  I was surprised to
find that the IR draws about 1/2 the power of the RS-232 in most cases and much
much less when the traffic is light!  My interest is mainly in medium speed
data logging links.  Consider the following scenario: The HP-48 sends a command
to a "voltmeter" (or some other instrument) 10 times per second to request the
present voltage.  Let's assume this command is 2 bytes long and that the
"voltmeter" responds with a 4 byte response.
 
     For the RS-232 case (any baud rate)- the power is essentially independent
of whether the transmitter is sending or not.  It is strictly dominated by the
IO being "open" (on).  A standard RS-232 load is 3000 ohms, so that the current
is 5v/3000 = 1.76 ma average.  It is actually slightly more than this because
the RS-232 level shifter draws some nominal ammount of current.
 
     The IR case is a bit more complex because all the current is drawn in
brief pulses.  There is virtually no quiescent current in the transmitter or
reciever- the power drawn depends solely on comm activity.  The transmitter
sends narrow pulses (52 us) for 0's and nothing for 1's.  The peak currents
with fresh batteries are 75 ma, however the average currents are very low.  In
order to send the worst case command (2 bytes of all zero's) 18 pulses of 52 us
of 75 ma would be sent or .070 ma-sec of charge.  The reciever would draw about
.5 ma 36 times (4 bytes of 9-52 us pulses) to recieve the worst message for a
charge of .00094 ma-sec.  The total is .0701 ma-sec per transaction, with 10
transactions per second this would be .7 ma average, less than half of that for
the RS-232 case.
 
     Note that if only one transaction per second were required, the average
current for the IR would be .07 ma while the RS-232 case would remain the same.
This is significant.  The "break even point" where the two circuits draw
appoximately the same current is about 45 characters/second being transmitted.
The RS-232 and IR recievers draw about the same ammount of power in all cases.
 
     **** Moral - if you are are doing interfacing at modest rates; (almost all
that is possible with an HP 48) you are better off with IR.  It also has the
advantage of isolating the HP 48 from the interface.  This rather surprising
result is brought to you by Travtech, manufacturers of (what else) an IR Link
Data Acquisition and Control module called the ADCM-48.
 
John Wettroth
TRAVTECH, San Diego
(619)-566-7861