Hi Steve,

Well... let's say it's possible, but it is by no means easy.

First off, this calls for a special timing receiver, like the Navman Jupiter-T or the Motorola M12+. These are intended for stationary applications and have been designed to provide the most stable and accurate timing reference possible. They can also be programmed to compensate for delays introduced by the antenna, cables, etc.

Secondly, you should not use the NMEA output which, as Craig said, is not reliable enough. In stead, you must use the binary protocol, combined with the 1PPS output. In the Navman module, a binary message is output before the PPS pulse, so that it essentially becomes a time mark. I haven't checked how the Motorola does this, but I expect it's very similar.

So now you have a 1PPS signal that's within +/- 50ns of UTC (probably even better), but now the question is: How are you going to maintain this accuracy, given the 'granularity' of your controller? I mean, even using a Cygnal device at 25MHz, an instruction cycle is 40ns; how are you going to use the time information and not introduce errors far larger than the accuracy of the time mark?

Best regards,
Rob.