[time-nuts] TimeLab

Tom Van Baak tvb at LeapSecond.com
Sun Oct 9 18:50:25 EDT 2016


> I'd guess it would work well for narrow pulses from GPSDOs but might be 
> complicated if you are looking at a typical low cost GPS unit.

Hi Hal,

No, not narrow pulses. Do not use the trailing edge of a 1PPS. This is more about 1 Hz from a stable frequency standard, not 1PPS from a noisy GPS receiver.

For this scheme to work you need a 1 Hz square wave. The idea is that you have two different stop edges to choose from, 0.5 second apart. Or, at least the edges have to wider than the deadtime of the TI counter, as well as the h/w and s/w delays in changing trigger edge over GPIB. 900 ms and 100 ms might work; but 0.9999 and 0.0001 is not helpful. 500 ms is a simple, safe bet.

It shouldn't be hard to get the rising edge and the trailing edge of a 1 Hz square wave equal to a couple of ns. This is especially true if the source of your 1PPS signal is a 10 MHz oscillator plus 10^7 square wave divider (like a TADD or picDIV).

If you dig deeper it could be that depending on your divider that the rising edge jitter is slight different than the falling edge jitter. Easy to measure and live with. It could also be that there is a small phase bias between the rising edge and the falling edge; either due to the driver, the cables, or the counter itself. This too can be measured. And unlike NTP, this asymmetrical bias could be subtracted from the raw data for that extra ns or two of accuracy.

Maybe I'm not explaining this 1 Hz either-edge method well. Let's wait until Magnus replies and then maybe I'll have to show you raw data to help explain the idea.

/tvb

----- Original Message ----- 
From: "Hal Murray" <hmurray at megapathdsl.net>
To: "Discussion of precise time and frequency measurement" <time-nuts at febo.com>
Cc: <hmurray at megapathdsl.net>
Sent: Sunday, October 09, 2016 3:00 PM
Subject: Re: [time-nuts] TimeLab


> 
> tvb at LeapSecond.com said:
>> 1) My *work-around* is to adjust the REF 1PPS by tens of microseconds, or
>> even 500 ms. That avoids running into sign changes and skipped samples when
>> a TIC gets near zero. This works really well for stable clocks where 500 ms
>> drift is next to impossible. 
> 
> How well does using the other edge of the pulse work?
> 
> I'd guess it would work well for narrow pulses from GPSDOs but might be 
> complicated if you are looking at a typical low cost GPS unit.
> 
> I have a scope with one probe on the PPS from a TBolt and another on the PPS 
> from a Sparkfun Venus board.  The start of the Venus PPS jumps around by 
> 50-200 ns p-p relative to the TBolt, but the trailing edge is solid relative 
> to the leading edge - at least at the ns level.  The infinite persistence 
> band is about 1 ns wide.  That's about the same as the leading edge, and 
> pushing things on a Rigol DS1102E.  After a longer wait, it's now 2 ns wide.  
> I'll guess that's the crystal changing temperature.  I don't see a crystal on 
> the board so I assume it's inside the big chip.  Or maybe the crystal in my 
> scope warmed up.
> 
> The Sparkfun/Venus PPS is 4 ms wide.
> 
> 
> --------
> 
> Another option is to use a scope.  :)
> 
> I haven't done it, but I'm pretty sure I could grab the data and rearm the 
> scope within a second.  It's not too hard to scan the data and find a level 
> crossing point.  I'll write some sample code if anybody is interested.
> 
> 
> 
> -- 
> These are my opinions.  I hate spam.
> 
> 
> 
> _______________________________________________
> 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