[time-nuts] New timing receivers?

Tom Van Baak tvb at LeapSecond.com
Mon Apr 28 05:45:10 UTC 2014

Hi Paul,

The whole issue of timing "accuracy" specs for GPS receivers is murky. Is it peak-to-peak, or TDEV, or RMS? How many sigma? Is it averaged over a minute, an hour, or a day? With or without sawtooth correction? In 2D or 3D or zero-D mode? At what latitude? Is it referenced to GPS or USNO(UTC) or UTC?

I don't know if there's a solid definition of a timing receiver. It would include some of:

 - has 1PPS output
 - provides sawtooth correction
 - allows optional zero-D mode
 - can perform self-survey
 - provides pseudo-range or RINEX data
 - manufacturer specified and validated timing latency (may vary by f/w revision)
 - takes external 5 or 10 or 20 MHz reference standard input

The other spec issue is that there's is a huge difference between "absolute" timing accuracy and "relative" timing accuracy. Almost all of us time-nuts use timing receivers for their relative accuracy, not absolute accuracy.

I am currently re-measuring many of the GPS boards I have here to make a complete performance matrix. That may shed light on the specs you mention. This includes various M12's and uBlox's as well as SiRF 3 and 4, Skylab, and MT3329. Some of these cheap 1PPS receivers are amazingly good.

Attached is the ADEV/MDEV plot for a Adafruit "Ultimate GPS" (MTK3339) board. Yes, the jitter/sawtooth really is +/- 5 ns. That's small enough, or at least small enough compared to the tens of ns of wander over hours, that the lack of sawtooth correction is not a problem. This runs in 3D mode (no survey needed, ok for mobile use); upstairs, indoors, onboard patch antenna.


----- Original Message ----- 
From: "Paul" <tic-toc at bodosom.net>
To: "Discussion of precise time and frequency measurement" <time-nuts at febo.com>
Sent: Sunday, April 27, 2014 8:45 PM
Subject: Re: [time-nuts] New timing receivers?

> On Sun, Apr 27, 2014 at 2:14 AM, Poul-Henning Kamp <phk at phk.freebsd.dk>wrote
>> Position Hold is what makes a GPS receiver "timing", the sawtooth
>> correction
>> is icing on the cake.
> I've been curious about this for a while:
> SkyNav says the SKG16 (based on the MT3329) has a timing accuracy of 60ns
> rms.  Sure says 20ns.  TVB measured 20ns with 60ns p-p sawtooth.
> GlobalTop says the FGPMMOPA6H (based on the MT3339) used in the previously
> mentioned AdaFruit break-out has a PPS with 10ns jitter(?).
> u-blox says 30ns or 15ns with sawtooth correction.  In the 2011 edition of
> Tom&Rick (the slide has gone missing in the 2013 edition*) they show the
> LEA-6T at 20ns uncorrected and ~5ns corrected.
> Trimble says <15ns rms for the  SMT.
> So I wonder if the MTK units are doing fixed position timing, if these
> numbers have not relationship to each other or if it's not that important.
> *btw, that's where they say "Even if (/when?) the Motorola/iLotus M12’s
> become unavailable, the uBlox LEA6T can step in as a replacement"
-------------- next part --------------
A non-text attachment was scrubbed...
Name: sigma-2.gif
Type: image/gif
Size: 58275 bytes
Desc: not available
URL: <http://febo.com/pipermail/time-nuts_lists.febo.com/attachments/20140427/4a8b4fdd/attachment.gif>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: jitter10.gif
Type: image/gif
Size: 77519 bytes
Desc: not available
URL: <http://febo.com/pipermail/time-nuts_lists.febo.com/attachments/20140427/4a8b4fdd/attachment-0001.gif>

More information about the Time-nuts_lists.febo.com mailing list