TAC-LITE status:                                  22 June 1996

A number of people have asked about the TAC-LITE activity to support
the TAPR group purchase of Garmin GPS-20 receivers:

All the documentation and info on my "Totally Accurate Clock" (TAC)
project resides on my aleph ftp server at URL

    ftp://aleph.gsfc.nasa.gov/GPS/totally.accurate.clock/
                            
(1) HARDWARE: My plan is that we will soon have a new "TAC" support circuit
  board that will handle the Garmin or Motorola receiver boards. The new
  board is not available yet -- I'm still working on it. For folks who
  want to get something going right now, there are two simple alternatives:

 - On aleph fetch the file  emulate.zip  which shows the simple cable
   connections to emulate a TAC by adding only one wire to the RS232
   cable to the computer. The added wire puts the CMOS-level 1PPS
   signal onto the DCD RS232 pin.

 - On aleph, fetch the file  tac-lite-prototype.zip. In this file you
   will find a PostScript drawing of the TAC-Lite prototype I built a
   few months ago that uses two IC's (7805 and 74AC04) to make a "real"
   TAC using the Garmin GPS-20. A few words of explanation on the
   schematic are in order:
     - The 1PPS signal is converted into "real" RS232 voltages. The
       negative supply for the RS232 signal is derived from the host
       computer by stealing voltages from the RTS, RXD and DTR lines
       with a simple diode (D1+D2+D3)/capacitor (C1) circuit. The PNP
       transistor Q1 acts as a level converter/inverter to drive the
       RS232 DCD line between -8v and +5v.
     - U1 (74AC04) acts as a low-impedance 1PPS driver for a user output
       and display LED. The LED is also lightly driven from +5v so
       that it is dim for 90% of the time and flashes bright 10% of
       the time when the receiver sends its 1PPS signal.
     - I have found that the microprocessor in the GPS-20 doesn't
       always start properly when fed from +5.0v. The 7805 regulator
       (U2) is "lifted" by R8=27 ohms so that the internal supply is
       about 5.3v. Be sure to insulate the U2 mounting tab so that
       R8 isn't shorted out. A TO220 "bolt-down" 7805 will run pretty
       warm. Be sure to have an adequate heat sink/radiator. My "TAC-Lite"
       has abour 2 square inches of copper foil to achieve this.
     - The Garmin receiver is slow in locking up after power is applied.
       To buffer against power dropouts, my TAC-Lite prototype includes
       a 9v NiCd battery to keep it running for ~5 minutes. The battery
       can be charged at 2 different rates. The "slow" trickle charge
       is about 3 ma (thru a 1k resistor and 2 diodes) and the "fast"
       rate is about 15 ma (180 ohm resistor).

  FYI -- my TAC-LITE prototype (as documented in the PostScript Schematic)
  is housed in a small calculator-sized instrument case about 4.3" x
  2.6" x 1.4" in size -- I had the case laying around and have no idea
  where I got it! This case was big enough for the GPS-20 plus the
  kludge board containing the added circuitry plus the 9v battery. 


(2) SOFTWARE: The latest version of my SHOWTIME software now supports
  the Garmin fully. On aleph you will find the latest SHOWTIME.EXE in
  the file showtm33.zip. If you have never installed SHOWTIME, change
  to the ~/distribution directory and fetch the latest distribution copy
  (about 1Meg in size). Then unzip it to a floppy with the -d option
  to create a mirror image of my distribution diskette and run setup.

  NOTE: When the Garmin receiver is powered up, the 1PPS output doesn't start
  until at least one GPS satellite is detected and then its downlink data
  is decoded. When this happens, the 1PPS LED on the TAC-LIT starts to flash.
  My SHOWTIME software requires the presence of the 1PPS signal on the
  COM1 (or 2) DCD line, and the software will not come to life until the
  receiver has this initial lock.
       
73 de Tom, W3IWI