[Elecraft] 10 MHz external reference - GPSDO

[Edward R Cole]

Guys<

TCXO's use temperature compensating components to try to keep 
frequency stable.  They work fairly well if you are at lower 
frequencies or not multiplying an LO 18 times.  Better is the OCXO 
(which uses a heater inside an insulated "oven" to better keep 
temperature effects isolated).

The EXREF utilizes the better stability of 10-MHz sources to correct 
the drift of the TCXO.  I does that at a fixed interval in between 
which both TCXO-1 and TCXO-3 will drift.  Per Elecraft's 
specifications on page 8 the TCXO-3 is 5 times more stable than the 
TCXO-1.  I have no idea over what time interval the specs are cited, 
but most assuredly its more than 4 seconds.

I cannot compare the TXCO-1 as I only have the TCXO-3.  Both were 
provided long before the EXREF was designed.
I just made a measurement of my K3/10 running 12w for about half a 
minute 28.200.000 displayed at 28.200.001 at the start.  My REF CAL = 
49.380.089 before making the 28-MHz test.  After several minute long 
tests the frequency has not budged off 28.200.001.  REF CAL now is 
49.380.095 so the TCXO-3 has moved +6-Hz over about half an hour.  I 
have been running all afternoon so the K3 was at stable working temp 
before I made a measurement.

Unless you are into frequency measuring or run weak signal modes 
where hearing a signal is quite difficult, being "on frequency" is 
very handy.  If running CW you probably have bandwidth cranked down 
to 100-Hz.  Digital weak-signal modes tolerate even less error.  On 
2m-eme using JT65, drifting more than +/- 20 Hz will result in the 
digital signal failing to decode.

Adding the EXREF improves your K3 by five times so perhaps the TCXO-1 
is satisfactory for your operating modes.
My measurements were really accuracy tests and not stability 
tests.  I did do a 30-minute test at 2m JT65 over a 30-minute period 
where I would transmit one minuted continuous and then receive one 
minute.  MY K3+2m transverter only drifted upward 7-Hz.  It was off 
frequency by 6-Hz at cold start.  K3EXREF with TCXO-3 was used.  The 
transverter uses a heated 116-MHz overtone xtal oscillator inside an 
insulated can.

Does any of this help?
Think I beat this to death - OK?

73, Ed - KL7UW
http://www.kl7uw.com/K3EXREF.htm

Message: 12
Date: Sun, 13 Mar 2016 09:08:10 -0400
From: Don Wilhelm <w3fpr at embarqmail.com>
To: Oliver Dr?se <droese at necg.de>, elecraft at mailman.qth.net
Subject: Re: [Elecraft] 10 MHz external reference - GPSDO
Message-ID: <56E5663A.5080601 at embarqmail.com>
Content-Type: text/plain; charset=windows-1252; format=flowed

I think everyone is ignoring something here.  What causes the frequency
drift in the first place?
It is temperature changes - the designation TCXO signifies Temperature
Compensated Xtal Oscillator.  It is already compensated for temperature.
If the temperature of the TCXO stays the same, the frequency should be
stable.

There will be some frequency change during warmup after power on, and
there will be some frequency change during transmit because the
temperature inside the box will increase, and will cool during periods
of receive.
The difference between the TCXO-1 and the TCXO-3 is one of how much.
The TCXO-3 drift will be smaller than the TCXO-1 over the same
temperature change.

The drift over the 4 or 5 second correction from the external reference
will result in a small correction with either TCXO, but it will be
smaller with the TCXO-3.
Additionally, the frequency drift will be quite small even without the
external reference unless there is a rapid and drastic change in
temperature at the case of the TCXO.  These discussions should include
the temperature change during the measurement periods. to complete the
picture.

73,
Don W3FPR

On 3/13/2016 5:27 AM, Oliver Dr?se wrote:
 > Thanks Ed.
 >
 > I don't have a problem understanding how the TCXO's work. Maybe I was
 > not clear enough. ;-) Taking your example from below, having only the
 > TCXO-1 it is surely *not* drifting 140 Hz at 28 MHz within the 4 or 5
 > seconds of the EXREF update cycle (same as the TCXO-3 will not drift
 > by 28 Hz in that time), otherwise the K3 would be a *very* crappy
 > radio. ;-) Think you are missing the time domain here. ;-)
 >
 > So my question still stands: What's the real value of the TCXO-3 over
 > the TCXO-1 when using the K3EXREF?



73, Ed - KL7UW
http://www.kl7uw.com
     "Kits made by KL7UW"
Dubus Mag business:
     dubususa at gmail.com