[time-nuts] New Product Release - GPSDO Frequency Standard

Bob Stewart bob at evoria.net
Thu May 19 02:33:53 EDT 2016


HI Attila,

I use a fixed voltage (not fixed current) RC integrator to measure the phase between the 1PPS from the LEA to the 10MHz signal from the OCXO.  At the same time, I'm counting the pulses from the OCXO to make sure I'm locked to 10MHz, not 10MHz +/- something.  The dsPIC33FJ128MC802 has a 12-bit ADC which I use to measure the voltage on the integrator.  I use software to correct for the RC curve, as well as real-world aberrations caused by the various chips in my TIC chain.  I apply 2 corrections: one to flatten the RC curve, and the other is a sine-wave correction.  I hand picked a constant (kappa) to tune the RC correction and another constant for the two sine-wave corrections.  They're different for 0-180 and 180-360 degrees.

>From there, the computed phase is corrected for the sawtooth from the LEA, and a somewhat modified PID controller controls the value of the DAC.  The DAC is actually a 16-bit PWM in the dsPIC33 dithered to 20 bits.  From there it goes through a 125 gate, low pass filter, and op-amp to the Trimble 34310-T OCXO.  The OCXO is sourced from China, and I hand select them for good ones.  The rejection rate is pretty high.  I need to update the pictures in the brochure.  The production unit can be configured to accept 10Mhz and output EFC to an external OCXO.  My engineer (Dan Kemppainen) has a 10811D (I think) buried in 30 pounds of sand, with a well located antenna.  He say's it's very stable.

I am currently working on software for temperature compensation.  I am on version 0.121 of my software.  The firmware in the unit can be updated through a USB port using the Microchip AN1094 bootloader.  I also have some plans for aging corrections, so that holdover will be better.  At the moment, the DAC is just frozen during holdover.  Owners will be invited to contribute to the software if they wish to and have a better solution than what I used.  System constants, such as the PID gains and surveyed position, are maintained across firmware updates and power outages.

When I said the PPS output from the unit comes directly from the receiver, I meant just that.  But that does not imply that my software doesn't compensate for the sawtooth in its own measurements for the 10Mhz/5MHz/1Mhz outputs.  I've been asked to add a PICDiv to the next build.  I don't know whether or not I can slew that exactly to phase 0.  It will be an interesting experiment to see if I can.

Tom actually made those plots from a pre-production unit I sent him last year.  The refHM(J2) refers to his hydrogen maser.  The measurements were done on his Timepod.  I'm not sure which antenna he used.  I believe he said it was a precision surveying antenna.  

Hope that answers your questions.  I could send you a schematic if you like.

Bob
--------------------------------------------
On Thu, 5/19/16, Attila Kinali <attila at kinali.ch> wrote:

 Subject: Re: [time-nuts] New Product Release - GPSDO Frequency Standard
 To: "Discussion of precise time and frequency measurement" <time-nuts at febo.com>
 Date: Thursday, May 19, 2016, 12:14 AM
 
 Hoi Bob
 
 On Thu, 19 May 2016 01:11:32
 +0000 (UTC)
 Bob Stewart <bob at evoria.net>
 wrote:
 
  
 > It's time I got off the dime and did a
 product release of my GPSDO.
 > My mission
 in creating this unit was to provide a stable and
 accurate
 > frequency reference with
 10Mhz, 5Mhz, and 1Mhz outputs.  It does have
 > a PPS output, but that is directly from
 the Ublox receiver, and is not
 >
 corrected for sawtooth error.  More information can be
 found in the
 > brochure and quick start
 guide in the link below.
 
 Can you give a bit more information on how your
 GPSDO works on the inside?
 What OCXO do you
 use?
 What technique do you use to get the
 phase information from the LEA?
 How do you
 steer the OCXO?
 Do you use any active
 temperature compensation?
 What were your
 references for the ADEV and phase noise plots?
 
 
        
     Attila Kinali
 -- 
 Reading can seriously damage your ignorance.
         -- unknown
 _______________________________________________
 time-nuts mailing list -- time-nuts at febo.com
 To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
 and follow the instructions there.


More information about the time-nuts mailing list