?

All

As some of you know, I have what I call a "GPS Simulator" which is an ATmega88 I programmed to send a NMEA string with the $GPRMC sentence, and it also has a realtime clock which updates correctly in the $GPRMC, and produces a 1pps signal output. Using this I can model different settings, baud rates, and 1pps pulse widths.

From the comments on which GPS modules worked and which didn't, there seemed to be a consensus opinion that GPS modules having very short (microseconds) pulse widths on the 1pps signal, do not work properly in v2.02a. So I have set my simulator with a very short pulse width, less than 1us. Using this setting I can see the seconds jump sometimes in 2 second jumps. So I am able to reproduce the same problem some GPS modules show.?

Next I printed out my code and stared at it all week on the train to and from work. ?I came up with a strategy for fixing it. This evening I tried the strategy. It works! In my simulator at least! Everything is ticking away here nicely with 1 minute frame, message "E" in FSK/CW with speed 006, GPS on, and < 1us pulse width. Heartbeat is beating nicely and 125MHz calibration has occurred, as have time setting and Maidenhead locator.

I will now send a v2.03 chip (as it will be called) to Eddie G3ZJO (proud Trimble owner) and Andy G0FTD (UP502 GPS) and Tony G0BZB (EM406). If you three can try it and it works on all three then I think this will be excellent progress indeed, maybe finally the Ultimate2 GPS issues will be over and we'll have widespread compatibility with most GPS modules.?

Fingers crossed...

73 Hans G0UPL

?