Projects, articles and links to information that may interest Radio Amateurs and Electronics enthusiasts





Trimble Thunderbolt GPS Monitor/ Controller

A stand-alone OLED display module to monitor and control the Trimble Thunderbolt GPS with 10MHz reference oscillator.

  I have several pieces of equipment in my shack that I rely on being at least reasonably accurate in their measurements. Items of test gear such as frequency counters and my Aeroflex IFR 1200S need to be accurate to be able to give any meaningful readings. Also there is the possibility of getting into microwave and this would lend itself nicely as a reference for that application also. I decided to obtain a 10MHz reference of some description and after searching many a sale site, decided to buy a Trimble Thunderbolt GPS DO (or GSP based Disciplined Oscillator). These have been very popular due to their low price and high availability. Software for the PC is easily obtainable but having some way of easily and continuously monitoring the unit would be a definite advantage. After reviewing several other displays and monitors for the Trimble by other developers, I decided to "roll my own". After all, I have written code before and I needed a new project. So I set about learning the protocol used by Trimble; in this particular GPS - called TSIP (or Trimble Standard Interface Protocol). I had already developed another project with a display and RS232 interface, so the hardware was pretty much already designed. I set up this module similar to others available but added a couple of my own features for good luck. As well as monitoring the time and date, oscillator accuracy, alarms, location and supply voltage, the GPS control buttons form part of the one unit, as is fan control because I mounted my GPS inside a 2 RU cabinet which caused everything inside to build up what I considered was too much heat. The Trimble has its own inbuilt temperature sensor which I used to determine if things are getting too hot and a cabinet fan which normally runs in quiet mode is switched to full speed if required. There is also an audible alarm for those times that you don't notice there is a problem with the GPS. I originally designed this module for myself and another Amateur who also bought the same GPS and so I did not prepare it for general distribution. The original PCB was a modified version of what I had used on a previous project but as of August 2015, I have designed a whole new board. The controlling micro is an AVR but is not Arduino. The user guide for version 2.1 can be downloaded here. Several other photos can be viewed on this page where you will see how I set the GPS, power supply and a divider in the 2 RU cabinet. The 4 way divider allows me to use separate outputs for test equipment or microwave frequency referencing. This divider was supplied by "Down East Microwave" in Florida.

To set up the Trimble Thunderbolt ready to take the above described display, software called "TboltMon" can be used. The settings required are described within the user guide for the display. This is where the program can be downloaded.

Fast forward to 2019 when the next GPS week number rollover occurs in April. (Note that this is not a problem for the 10MHz reference - only the date and time display)

The Trimble Thunderbolt like perhaps many other GPS devices was not prepared for the next Epoch rollover of week numbers. Each 19 years roughly since 1980 when GPS systems came on line, a counter called the week number increments to 1023 and then starts again at 0. The problem with this is that when week 0 comes around, the GPS will not output the correct date because it was not designed to go past the 1023 count that really is starting again from 0. Newer GPS deevices have taken this into account and will pass through the rollover without a fuss but liken this to the Y2K confusion when lots of managers had to be sure that their systems would not fail on day 1 of 2000. Since so much technology relies on GPS satellite timing now, many companies are discarding their older GPS devices in favour or one that is guaranteed to be stable after April 6, 2019.

With the explanation aside, the exercise began to rewrite the monitor software to allow for this event and although the solution sounds simple, the calculations require some thinking. The date which is ouput from the Thunderbolt is generated within the firmware of the GPS itself and is not a result of data sent by the satellites. Therefore if the firmware cannot be upgraded (Trimble consider it better economy for consumers to purchase new rather than upgrade old devices and understandably so) the date must be calculated outside the GPS. Since the device also outputs the week number and number of seconds into the week known as time of week, then our own calculations can be done using Julian Days and Julian Dates. So after a rehash of the date and time using these two bits of data, the display will now show accurate date and time from January 2019 until around 2038 when hopefully I wont have to worry about it again. The new calculations takes leap seconds into account and the display will now only show UTC time. Therefore there no longer will be a "U" or a "G" showing in the top right of the display and the GPS/UTC button will not perform any function unless in program mode.