Re: [gpsd-users] Support for PPS on 10Hz Devices

[Andrew Yong]

Heya Gary,

> On 16 Sep, 2013, at 12:17 pm, "Gary E. Miller" <address@hidden> wrote:
> 
> Yo Andrew!
> 
> On Mon, 16 Sep 2013 08:53:26 +0800
> Andrew Yong <address@hidden> wrote:
> 
>> Hi Anthony,
>> 
>>> On 16 Sep, 2013, at 4:23 am, Anthony Stirk <address@hidden> wrote:
>>> 
>>> I might be out of turn here but isn't the PPS output still 1Hz
>>> regardless of the NMEA output frequency ?
>> 
>> Yep, the PPS still comes in at 1Hz, but for some reason GPSD doesn't
>> pick it up.
> 
> How do you know it is coming if you can't see it?  What RS-232
> pin is it coming in on?  Have you tried high debugging levels in 
> gpsd to see hat gpsd is seeing?

I ran it at 10 and grepped for PPS|NTP. Nothing of note is output at 10Hz. Same 
module at 1Hz NMEA has PPS detected. 

I have it on a USB FTDI chip’s DCD pin. I see DCD blinking at 1Hz using 
RealTerm on Windows, regardless of 1/5/10Hz NMEA updates. 

Pretty sure it's picking up PPS correctly at 1Hz as my PPS LED and GPSD debug 
output coincide perfectly. 

> 
>> I recall seeing a patch to the code base to specifically support the
>> Garmin 5Hz LVC.
> 
> Yes, applied a long time ago and it ws working last time someone checked.

I'll take a better look at the hack later today. If I recall correctly, a 
comment in the code mentioned that it was a dirty hack. :P

> 
>> Perhaps GPSD is overwhelmed with the 10Hz output, or just is
>> expecting 1Hz NMEA?
> 
> NMEA is all over the map.  Even on a 5HZ GPS some sentences are only
> output every 5 seconds, or less.  None of it applies to PPS.  All PPS
> cares is that there is a good fix and the pulse is seen.
> 
>>> I did noticed when using the Adafruit module it seems to fire PPS
>>> on both the rising and falling, i.e I was getting 2 pulse per
>>> second. Switched it for a ublox which seems to work better (by
>>> better I mean 1 pulse per second).
> 
> gpsd can handle a wide variety of 'PPS' inputs.  There are no reports
> of anything out there it does not handle.
> 
>> I can customize the PPS pulse duration; it's set to 100ms by default.
>> Should I get that decreased or increased for better compatibility
>> with GPSD's PPS detection?
> 
> gpsd can handle everything from a pulse so small the UART does not see
> both edges to a 2HZ square wave where each edge is on the second.  100
> milli Sec is good because there is no ambiguity on which edge, but 1
> milli Sec to 200 milli Sec would be equally good.  Just let it be. 

That's good to know. At 1Hz it picks both edges and always chooses the rising 
edge. So I'm going to say it works as intended. 

However it rejects the PPS because of a "no GPS seconds" error. I've requested 
a firmware for the MTK3339 with $GPZDA sentences in case that fixes it. 

Anyway, I'm getting around 5-10ms jitter at 10Hz, no PPS. 
http://www.pool.ntp.org/scores/2001:d98:4800:1::1

> 
> RGDS
> GARY
> ---------------------------------------------------------------------------
> Gary E. Miller Rellim 109 NW Wilmington Ave., Suite E, Bend, OR 97701
>    address@hidden  Tel:+1(541)382-8588

smime.p7s
Description: S/MIME cryptographic signature