Re: Those GPS 1pps problems. (Some bystander thoughts)
|
Hi Eddie, Andy
Thanks for the thoughts on this. This is the next challenge - to make the GPS Interface work with both very short, and very long pulses.
I don't know if I mentioned it before, But I have what I call a "GPS simulator" which I use for testing such things. It basically is an ATmega168 (the same chip) programmed to produce a $GPRMC sentence and code the time and location into it, and a 1pps signal. It also has a button to toggle the GPS on/off which can simulate the effect of disconnecting it or it losing satellite lock. I am able to adjust the pulse-width in this "simulator". So with sufficient effort, I ought to be able to model the Trimble and the Rayax 501 and produce code which works with both of them, even though I physically own neither.
I believe the Rayax 501 worked with both U1 and U2 v2.01 and v2.02. So it is as Andy says - GPS modules with the very short pulse times seem to be the challenge.?
73 Hans G0UPL
On Sun, Jun 16, 2013 at 10:33 AM, g3zjo <g3zjo@...> wrote:
?
On 16/06/2013 10:04, andyfoad@... wrote:
> For example I've seen figures of nanoseconds and microseconds quoted on
> the troublesome modules, yet my Reyax UP501 which works flawlessly has
> a whole 100ms width for the 1pps output.
>
> >From where I am standing it does seem that this pulse width is the
> key to it all.
Andy
Did the Rayax work on v2.01
The 1pps pulse width does seem to be tied up with the claimed accuracy
for for timing / frequency purposes. Hans obviously knows how to handle
long and short versions, a mid way compromise seems to be needed, if
v2.01 had not got it right for the Trimble I may have been thinking
about adding a Schmitt Trigger.
Hans was still exuding confidence on a solution last time we spoke on
the subject.
I didn't comment at the time because I haven't tried it, but your tip of
using an LED to detect the PPS signal I am fairly sure would not work on
the Trimble GPS, its too quick for the Scope so I don't think the eye
would catch it.
73 Eddie
|
