K3S XREF GPSDO recommendation needed

[Grant Youngman-2]

GPSDO “Holdover” mode (loss of GPS sync) runs typically better than 1x10 (to the) -9 ppm per day.  These devices use the last control values and historical data to manage the OCXO (or rubidium, or cesium standard) during periods without GPS sync.  They don’t just default to some nominal frequency or run at random.  Outside of the occasional times I’ve had to disconnect the GPS antenna to move things around, I’ve never observed any of the GPSDOs I’ve used go into holdover mode on their own.  I used to have logs monitoring the old HP standards when I had them, but no longer do that — since it was like monitoring the location of a period on a page :-)

The K3 will retain the last Ref Cal frequency if you unplug the GPSDO altogether.  Which nominally, should be the Ref Cal setting you would get if you went through a manual cal and set it at the current rig temperature.

Which is kind of the point.  Without the external sync, if you cal the radio after it’s been warming up for 30 minutes, and then start running RTTY or FT8 or some other mode that’s going to heat the radio up, the cal will go to h*ll in a hand basket — if a few Hz or 10s of Hz is a hand basket.  Well not in any practical sense, necessarily,  but it will change.  The GPSDO “nudges” it back into place and is not dependent on the temp of the internal K3 standard.

It’s an compulsive disorder, and not very much of a practical thing, in any case :-)

Grant NQ5T
K3 #2091, KX3 #8342

>
> If GPS lock is lost on the Bodnar device it reverts to a TCXO that is likely no better than the one in the K3, assuming the high stability option.  (Note: I haven't pursued all of the relevant specs, this is just gut feel.)
>

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html
Message delivered to [hidden email]

[Grant Youngman-2]

I just went back and read the Bodnar specs again, and it does in fact use a TCXO and NOT an OCXO which is used in most commercial units.  My bad.  Should have realized that given its small size.  It’s holdover characteristics won’t be as good as a unit with an ovenized oscillator.

Nonetheless if holdover periods are few and far between, it may not be much of an issue.  In any case, I’ll stick with my Trimble as long as it doesn’t smoke the room up :-)

Grant NQ5T
K3 #2091, KX3 #8342

> On Jun 5, 2018, at 6:57 PM, Grant Youngman <[hidden email]> wrote:
>
> GPSDO “Holdover” mode (loss of GPS sync) runs typically better than 1x10 (to the) -9 ppm per day.  These devices use the last control values and historical data to manage the OCXO (or rubidium, or cesium standard) during periods without GPS sync.  They don’t just default to some nominal frequency or run at random.  Outside of the occasional times I’ve had to disconnect the GPS antenna to move things around, I’ve never observed any of the GPSDOs I’ve used go into holdover mode on their own.  I used to have logs monitoring the old HP standards when I had them, but no longer do that — since it was like monitoring the location of a period on a page :-)
>
> The K3 will retain the last Ref Cal frequency if you unplug the GPSDO altogether.  Which nominally, should be the Ref Cal setting you would get if you went through a manual cal and set it at the current rig temperature.
>
> Which is kind of the point.  Without the external sync, if you cal the radio after it’s been warming up for 30 minutes, and then start running RTTY or FT8 or some other mode that’s going to heat the radio up, the cal will go to h*ll in a hand basket — if a few Hz or 10s of Hz is a hand basket.  Well not in any practical sense, necessarily,  but it will change.  The GPSDO “nudges” it back into place and is not dependent on the temp of the internal K3 standard.
>
> It’s an compulsive disorder, and not very much of a practical thing, in any case :-)
>
> Grant NQ5T
> K3 #2091, KX3 #8342
>
>>
>> If GPS lock is lost on the Bodnar device it reverts to a TCXO that is likely no better than the one in the K3, assuming the high stability option.  (Note: I haven't pursued all of the relevant specs, this is just gut feel.)
>>
>

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html
Message delivered to [hidden email]

[Edward R Cole]

Wes,

You got it pretty close.  I was kind of stumped by Clay's
question.  Disciplining I take to mean controlling from an external
source - huh? why?

If you want GPS frequency accuracy then buy a GPSDO for a couple
hundred dollars, but as far as affecting K3 frequency accuracy it
will not be significantly more accurate than using an OCXO (Oven
Controlled Xtal Oscillator).  The Morion spec for short term
stability is +/- 5 E-12.  Long term stability is+/- 5 E-5.  The OCXO
drifts about 1-Hz per 6-months from my actual measurements.  So, yes
it needs to be disciplined occasionally (twice a year) manually by
adjusting the freq. adjust pot comparing its output to an accurate
10-MHz source (I use my LPRO Rubidium as my reference).   The short
term stability of the Rb is +/- 5 E-11 but that is "good enough".

BTW the K3 frequency is established by the TCXO and the external
reference merely performs a periodic check. Any freq error causes a
correction instruction to the TCXO to bring it back onto
frequency.  The TCXO-3 is rated at 0.5 ppm which is +/-5 E-7.  Adding
the EXREF limits drift to 0.1ppm or E-7.

At 28-MHz that implies a 28 E+6/E-7 or 2.8 Hz.  The TCXO drift is
limited to +/- 2.8 Hz.  The OCXO is 10,000 times more stable and thus
more accurate.

The reason you do not get the OCXO level of stability is that the
EXREF system does not phase-lock the TCXO.  It periodically corrects
the error but the TCXO drifts normally between freg corrections which
happen often enough to reduce the amount of drift from 5.6 Hz to 2.8
Hz at 28-MHz.

The OCXO returns the TCXO to between 28,000,000.0005 Hz to
27,999,999,9995 Hz but then the TCXO drifts 2.8 Hz by the next
frequency check.  I think 0.0005 Hz is close enough.  So maybe the
GPSDO is +/- 5 E-13.  That would correct the K3 to 0.000005 Hz, but
it would still drift 2.8Hz by the next frequency check.

My actual frequency measurements show 0.5 to 1.5 Hz inaccuracy at
28-MHz and 1.5 to 2.5 Hz inaccuracy at 50-MHz so my TCXO-3 is doing
better than 0.5ppm.

Now if you want to know why Elecraft chose periodic freq correction
over true phase-locking, the answer is low LO phase noise.  Much
harder to get low phase noise in a PLL than in a TCXO.  And now you
know the rest of the story.

73, Ed - KL7UW

From: Wes Stewart <[hidden email]>
To: [hidden email]
Subject: Re: [Elecraft] K3S XREF GPSDO recommendation needed

I'm not answering for Ed, and I don't have the Ref Lock doodad for my K3.

That said, a decent OCXO, like a Morion, will have, after warm up and a little
aging, stability more than adequate to discipline a K3.? The K3 isn't phase
locked to an external reference anyway, it's just "nudged" on frequency
periodically.

If GPS lock is lost on the Bodnar device it reverts to a TCXO that is
likely no
better than the one in the K3, assuming the high stability option.? (Note: I
haven't pursued all of the relevant specs, this is just gut feel.)

I have the Bodnar unit because my sweetie asked me what I wanted for
Christmas a
few years ago and I had just read about it.? I have no compelling need, but I
did use it with a modified SDR-IQ to try my hand at the last ARRL FMT.? I also
have a Morion, off eBay, that I thought I should have.? I have yet to apply
power to it :-)? I suppose I was remembering the days when I would have killed
to have something that accurate.

Likewise after reading timenuts I thought I should get a TAPR GPS
time standard
kit.? I have absolutely no need for that either but Lady Heather
looks nice on a
spare computer screen ticking away at one second increments with nanosecond
accuracy :-).

Wes? N7WS


  ?On 6/5/2018 12:55 PM, Clay Autery wrote:

 > Just curious...? Why would you choose NOT to discipline the OXCO?


73, Ed - KL7UW
   http://www.kl7uw.com
Dubus-NA Business mail:
   [hidden email]

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html
Message delivered to [hidden email]

[Elecraft mailing list]

I don't think that is correct. Nothing touches the TCXO, the VCTRL pin
is not used. The EXREF module just reads the actual ref frequency and
that information is used to set up the synths. As the TCXO drifts the
synths just use the new measured ref frequency.

Regards,

Mike VP8NO

 > BTW the K3 frequency is established by the TCXO and the external
 > reference merely performs a periodic check. Any freq error causes a
 > correction instruction to the TCXO to bring it back onto frequency.  The
 > TCXO-3 is rated at 0.5 ppm which is +/-5 E-7.  Adding the EXREF limits
 > drift to 0.1ppm or E-7.

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html
Message delivered to [hidden email]

[Clay Autery-2]

Thank you for the response, Ed.  Why discipline the OXCO, or even the DOXCO?

Because you can.

Because the ability is there to progressively steer the correction
instead of stomp on it twice a year.

Lots of reasons, not the least of which is not having to open a chassis
and/or drill a strategically placed hole in a chassis in order to allow
you to manually make the correctly.

And there are those of us who use the frequency reference for things
other than just our radios. <smile>

BUT I understand why you do it the way you do, now.  Which was the
intent of the original question.  Thanks!

And I do appreciate the information that the K3 is not phase locked and
why the choice was made.
Does anyone have any idea how often that "check and nudge" procedure is
done or better the logic behind how it is determined when to make a
correction?  I assume it is implemented in firmware.

Have a great day! <smile>

______________________
Clay Autery, KY5G
[phone]

On 06/06/18 02:45, Edward R Cole wrote:

> Wes,
>
> You got it pretty close.  I was kind of stumped by Clay's question. 
> Disciplining I take to mean controlling from an external source - huh?
> why?
>
> If you want GPS frequency accuracy then buy a GPSDO for a couple
> hundred dollars, but as far as affecting K3 frequency accuracy it will
> not be significantly more accurate than using an OCXO (Oven Controlled
> Xtal Oscillator).  The Morion spec for short term stability is +/- 5
> E-12.  Long term stability is+/- 5 E-5.  The OCXO drifts about 1-Hz
> per 6-months from my actual measurements. So, yes it needs to be
> disciplined occasionally (twice a year) manually by adjusting the
> freq. adjust pot comparing its output to an accurate 10-MHz source (I
> use my LPRO Rubidium as my reference).   The short term stability of
> the Rb is +/- 5 E-11 but that is "good enough".
>
> BTW the K3 frequency is established by the TCXO and the external
> reference merely performs a periodic check. Any freq error causes a
> correction instruction to the TCXO to bring it back onto frequency. 
> The TCXO-3 is rated at 0.5 ppm which is +/-5 E-7. Adding the EXREF
> limits drift to 0.1ppm or E-7.
>
> At 28-MHz that implies a 28 E+6/E-7 or 2.8 Hz.  The TCXO drift is
> limited to +/- 2.8 Hz.  The OCXO is 10,000 times more stable and thus
> more accurate.
>
> The reason you do not get the OCXO level of stability is that the
> EXREF system does not phase-lock the TCXO.  It periodically corrects
> the error but the TCXO drifts normally between freg corrections which
> happen often enough to reduce the amount of drift from 5.6 Hz to 2.8
> Hz at 28-MHz.
>
> The OCXO returns the TCXO to between 28,000,000.0005 Hz to
> 27,999,999,9995 Hz but then the TCXO drifts 2.8 Hz by the next
> frequency check.  I think 0.0005 Hz is close enough.  So maybe the
> GPSDO is +/- 5 E-13.  That would correct the K3 to 0.000005 Hz, but it
> would still drift 2.8Hz by the next frequency check.
>
> My actual frequency measurements show 0.5 to 1.5 Hz inaccuracy at
> 28-MHz and 1.5 to 2.5 Hz inaccuracy at 50-MHz so my TCXO-3 is doing
> better than 0.5ppm.
>
> Now if you want to know why Elecraft chose periodic freq correction
> over true phase-locking, the answer is low LO phase noise.  Much
> harder to get low phase noise in a PLL than in a TCXO.  And now you
> know the rest of the story.
>
> 73, Ed - KL7UW
>
> From: Wes Stewart <[hidden email]>
> To: [hidden email]
> Subject: Re: [Elecraft] K3S XREF GPSDO recommendation needed
>
> I'm not answering for Ed, and I don't have the Ref Lock doodad for my K3.
>
> That said, a decent OCXO, like a Morion, will have, after warm up and
> a little
> aging, stability more than adequate to discipline a K3.? The K3 isn't
> phase
> locked to an external reference anyway, it's just "nudged" on frequency
> periodically.
>
> If GPS lock is lost on the Bodnar device it reverts to a TCXO that is
> likely no
> better than the one in the K3, assuming the high stability option.?
> (Note: I
> haven't pursued all of the relevant specs, this is just gut feel.)
>
> I have the Bodnar unit because my sweetie asked me what I wanted for
> Christmas a
> few years ago and I had just read about it.? I have no compelling
> need, but I
> did use it with a modified SDR-IQ to try my hand at the last ARRL
> FMT.? I also
> have a Morion, off eBay, that I thought I should have.? I have yet to
> apply
> power to it :-)? I suppose I was remembering the days when I would
> have killed
> to have something that accurate.
>
> Likewise after reading timenuts I thought I should get a TAPR GPS time
> standard
> kit.? I have absolutely no need for that either but Lady Heather looks
> nice on a
> spare computer screen ticking away at one second increments with
> nanosecond
> accuracy :-).
>
> Wes? N7WS
>
>
>  ?On 6/5/2018 12:55 PM, Clay Autery wrote:
>
> > Just curious...? Why would you choose NOT to discipline the OXCO?
>
>
> 73, Ed - KL7UW
>   http://www.kl7uw.com
> Dubus-NA Business mail:
>   [hidden email]
> ______________________________________________________________
> Elecraft mailing list
> Home: http://mailman.qth.net/mailman/listinfo/elecraft
> Help: http://mailman.qth.net/mmfaq.htm
> Post: mailto:[hidden email]
>
> This list hosted by: http://www.qsl.net
> Please help support this email list: http://www.qsl.net/donate.html
> Message delivered to [hidden email]

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html
Message delivered to [hidden email]

[Elecraft mailing list]

This is lifted from an Elecraft document:

"ELIMINATING FREQUENCY JUMPS ON 2M WHEN USING THE K3EXREF AND K144XV PLL
LOCK MODULE"

<snip>
In order to lock to an external reference, the K3EXREF module counts the
frequency of the 49.380-MHz reference oscillator for about one second,
repeating about every four seconds. During this count time, the
circuitry draws about twice as much current as when it's static (it goes
from 7 mA to 15 mA). When it's drawing more current, it slightly
decreases the supply voltage to the reference oscillator, causing a
small frequency shift observable at 2 meters and possibly even on 6 meters.
<snip>

The mod is simple to apply.

Regards,

Mike VP8NO


On 06/06/2018 16:17, Clay Autery wrote:

> Does anyone have any idea how often that "check and nudge" procedure is
> done or better the logic behind how it is determined when to make a
> correction?  I assume it is implemented in firmware.
______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html
Message delivered to [hidden email]