[GPS_Standard] GPSstd_PLL testing: V0.13 with an improved State Machine 2

[SP2IQW]

Hi Bob and All,

I'm carefully observing your software work on improving Berts's GPSDO.
Few years ago I have build VE2ZAZ GPSDO with some hardware rework. To 
eliminate jitter caused by noisy and not so stable 5V PS line I have 
added uA723 based 5V voltage regulator supplying only CMOS buffer 
(non-inverting) on the PWM output.

As per GPS, your choice to use MTK-chipset means current "state of art" 
within simple general purpose receivers. All MTK3329, MTK3339, MTK3333 
(this one is GPS/Glonass) are claimed to has ±10 ns RMS PPS accuracy 
which is very promising result.
For new GPSDO design even with such good Mediatek chipsets available I 
would prefer dedicated modules from Trimble as Resolution SMT™ GG 
Multi-GNSS Timing Module: 
http://www.trimble.com/timing/pdf/022542-039A_Resolution_SMT_GG_DS_0412_US_LR.pdf 

They are build especially for timing purposes with TRAIM functionality 
(see datasheet).


My one cent about 2N2222 inverter used on PPS input between Adafruit 
receiver and 5V PIC input.
All precise timing are referred to the rising edge on the PPS output 
signal. So if unchanged in GPS or PIC software, the microcontroller will 
look for relatively slow rising edge on its input (rise time = multiply 
of input capacitance and pullup resistor). Events on the falling edge of 
PPS signals are not defined such well as raisng edge so uncontrolled 
jitter may occur between succesive falling PPS pulses. I would suggest 
to use two 74HCT series inverters in series with extra 470ohm 5V pullup 
resistor on the PIC input istead of 2N2222 inverting stage. HCT logic 
works well with 3.3V CMOS outputs.


Best 73,
Michal
sp2iqw



W dniu 2013-11-13 21:55, Bob Stewart pisze:
> Here I am running a test on what is essentially an upgraded Version 0.13 with a slightly improved State Machine 2.  Jim had asked me to run a long test of the one with just State Machine 1 that I posted previously, but this will have to suffice for now.  The performance should be more or less the same as V0.8 since the Adafruit receiver doesn't have the wild phase jumps on the 1PPS signal that come from the UT+ that I was trying to tame.  I've decided to run this until at least tomorrow morning.   So I'll update it then.
>
> http://www.evoria.net/AE6RV/GPSstd_PLL/Plots/SM1SM2.30hrs.png
>
> Just as a reminder, this is with using an Adafruit Ultimate Breakout GPS Receiver with the antenna near the peak of the attic.  There is a PN2222A serving as an amp/inverter in the 1PPS line, as the Adafruit doesn't have enough drive to provide a reliable TTL signal to the 18F2220.
>
> For test equipment, I am using a Prologix Ethernet GPIB adapter and an HP 5334B to provide the phase timing.  The 10MHz signal from Bert's board is serving as the timebase for the 5334B.  The 1PPS signal goes to the A Channel and the 10MHz output from Bert's board (different port from timebase) is going to the B Channel.  The mode is TI A->B, which measures the time interval from the rising edge of the pulse on Channel A (1PPS) to the rising edge of the next pulse on Channel B (10MHz from Bert's board).  The monitoring software is home-grown.
>
>
> Bob - AE6RV
> ______________________________________________________________
> GPS_Standard mailing list
> Home: http://mailman.qth.net/mailman/listinfo/gps_standard
> Help: http://mailman.qth.net/mmfaq.htm
> Post: mailto:GPS_Standard at mailman.qth.net
>
> This list hosted by: http://www.qsl.net
> Please help support this email list: http://www.qsl.net/donate.html
>