Discussion:
[Elecraft] KX3: thermal frequency drift
Edward R Cole
2014-04-11 11:43:07 UTC
Permalink
I decided to test the frequency stability of the KX3 on 6m under
key-down conditions one would see using JT65. Note : I have not
performed the Temperature Compensation Procedure so not expecting
frequency to hold as close as after doing the procedure.

I checked frequency five times with quick transmissions that would
not result in heating:
50.125000 was the dial frequency measured with 5w output measured on
a Bird 43 meter.
50.124981 -29 Hz error
50.124950 -50 Hz error
50.124960 -40 Hz error
50.124957 -43 Hz error
50.124957 -43 Hz error

Next I ran ten sequences where I keyed down for 50-seconds and Rx for
70-seconds to duplicate JT65 transmitting sequences: I got drift downward of
50, 49, 34, 25, 29, 32, 19, 12, 9, and 12 Hz
Cold frequency at the start of each sequence varied from 979 to 960 to 994 Hz.

It did appear that the radio warmed up over the 20-minutes test and
high/low frequency deviation narrowed to about 10-Hz downward
drift. At the end the radio case was quite warm and the HI TEMP
warning came on! From this I would not recommend running JT65 for
extended time with the stock KX3.

Next I installed the nice external heat sink being made by Fred
Meier, VE7FMN. I installed it in place of the original
heat sink plate which took about 20-minutes to install (remove four
screws that hold the heat plate and install the new heat sink using
supplied allen screw bolts. I decided to purchase the plain aluminum
heat sink vs the black power-coat version. I used no heat-sink
compound trusting just the metal surface conduction and two 4-40
screws that secure the transmit transistor tabs.

Testing at 5w I observed 50.124962 +/- 1 Hz over three 50-second
cycles. The heat sink only felt mildly warm (barely
detectable). Note: this is without performing the Temperature
Calibration Procedure! To say I am impressed it an
understatement. I will be interested to see if the initial 38-Hz
cold-temp frequency error is minimized by the Temp Compensation
procedure. Note: I tested my frequency counter accuracy against my
10-MHz OCXO reference and saw less than a Hz error.

How this will affect temperature related drift using the new 2m
module remains to be determined as the Tx device on the 2m board
probably does not use the existing heat sink interface. I'll post the
raw data on my website later on.

73, Ed - KL7UW
http://www.kl7uw.com
"Kits made by KL7UW"
Dubus Mag business:
dubususa at gmail.com
Edward R Cole
2014-04-11 20:39:12 UTC
Permalink
It was suggested I do the Temp Compensation Procedure. Unfortunately
I do not have a stable 50-MHz signal source. However I did run the
Reference Frequency Calibration using my XG3. Call this "Part Two":

I sampled the RF output of the XG3 with my counter and also input
that to the KX3 for the Calib. This required running at 0 dBm to
inject strong enough signal for the counter (approx -3 dBm). I
turned on the ATT for the KX3 to keep input from overdriving. Got
about S9 signal level.

I got it adjusted to within about 2-Hz. The XG3 drifts approx +/- 2
Hz so difficult to set the Ref*Frequency using it. But that
eliminated most of the 38 Hz offset I saw last night. Running the Tx
at 5w afterward drift was in the range of 0 to +7 Hz for a 50-second
key down run. The drift is not steady in one direction so average
error is probably in the neighborhood of +5 Hz with 5-Hz uncertainty
in the measurement. Crude but gives a ballpark idea what to expect.

Today's measurements were all with new heat sink installed. Temp
Comp. probably would lower these numbers a bit.

I might try running my K3 as signal source for Temp Comp. if it holds
more stable then the XG3 at 50-MHz. I am using the EXREF so it may be better?

73, Ed - KL7UW
http://www.kl7uw.com
"Kits made by KL7UW"
Dubus Mag business:
dubususa at gmail.com
Matt VK2RQ
2014-04-11 21:58:42 UTC
Permalink
You are probably better off zero-beating to WWV. You should easily be able to get within 500mHz or less.

73,
Matt VK2RQ
I sampled the RF output of the XG3 with my counter and also input that to the KX3 for the Calib. This required running at 0 dBm to inject strong enough signal for the counter (approx -3 dBm). I turned on the ATT for the KX3 to keep input from overdriving. Got about S9 signal level.
I got it adjusted to within about 2-Hz. The XG3 drifts approx +/- 2 Hz so difficult to set the Ref*Frequency using it. But that eliminated most of the 38 Hz offset I saw last night. Running the Tx at 5w afterward drift was in the range of 0 to +7 Hz for a 50-second key down run. The drift is not steady in one direction so average error is probably in the neighborhood of +5 Hz with 5-Hz uncertainty in the measurement. Crude but gives a ballpark idea what to expect.
Today's measurements were all with new heat sink installed. Temp Comp. probably would lower these numbers a bit.
I might try running my K3 as signal source for Temp Comp. if it holds more stable then the XG3 at 50-MHz. I am using the EXREF so it may be better?
73, Ed - KL7UW
http://www.kl7uw.com
"Kits made by KL7UW"
dubususa at gmail.com
______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:Elecraft at mailman.qth.net
This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html
Message delivered to matt.vk2rq at gmail.com
Oliver Dröse
2014-04-11 22:03:04 UTC
Permalink
As you need to do the extend temp procedure at +/- 50 MHz that's no
option, Matt.

@Ed:
The K3 will work fine as source for the ext. temp. procedure. Just turn
it down to lowest power or use the transverter out port.

73, Olli - DH8BQA

Contest, DX & radio projects: http://www.dh8bqa.de
Post by Matt VK2RQ
You are probably better off zero-beating to WWV. You should easily be able to get within 500mHz or less.
73,
Matt VK2RQ
I sampled the RF output of the XG3 with my counter and also input that to the KX3 for the Calib. This required running at 0 dBm to inject strong enough signal for the counter (approx -3 dBm). I turned on the ATT for the KX3 to keep input from overdriving. Got about S9 signal level.
I got it adjusted to within about 2-Hz. The XG3 drifts approx +/- 2 Hz so difficult to set the Ref*Frequency using it. But that eliminated most of the 38 Hz offset I saw last night. Running the Tx at 5w afterward drift was in the range of 0 to +7 Hz for a 50-second key down run. The drift is not steady in one direction so average error is probably in the neighborhood of +5 Hz with 5-Hz uncertainty in the measurement. Crude but gives a ballpark idea what to expect.
Today's measurements were all with new heat sink installed. Temp Comp. probably would lower these numbers a bit.
I might try running my K3 as signal source for Temp Comp. if it holds more stable then the XG3 at 50-MHz. I am using the EXREF so it may be better?
73, Ed - KL7UW
http://www.kl7uw.com
"Kits made by KL7UW"
dubususa at gmail.com
______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:Elecraft at mailman.qth.net
This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html
Message delivered to matt.vk2rq at gmail.com
______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:Elecraft at mailman.qth.net
This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html
Message delivered to droese at necg.de
Matt VK2RQ
2014-04-11 22:20:28 UTC
Permalink
You are right, you can't use WWV for the temperature compensation data collection procedure itself. However, we are talking about the reference calibration that you must perform after the temperature compensation data collection -- for that, you need a precisely known frequency reference. Ed below said he was using his XG3 for that, but it was drifting around 2Hz or so. I was saying that you can do better than that by zero-beating with WWV, even with Doppler shift error. Note that when I say "zero beat", I don't mean tuning the carrier down to "0Hz" (that topic has been beaten to death, so I won't repeat it here :-) ).

73,
Matt VK2RQ
As you need to do the extend temp procedure at +/- 50 MHz that's no option, Matt.
The K3 will work fine as source for the ext. temp. procedure. Just turn it down to lowest power or use the transverter out port.
73, Olli - DH8BQA
Contest, DX & radio projects: http://www.dh8bqa.de
Post by Matt VK2RQ
You are probably better off zero-beating to WWV. You should easily be able to get within 500mHz or less.
73,
Matt VK2RQ
I sampled the RF output of the XG3 with my counter and also input that to the KX3 for the Calib. This required running at 0 dBm to inject strong enough signal for the counter (approx -3 dBm). I turned on the ATT for the KX3 to keep input from overdriving. Got about S9 signal level.
I got it adjusted to within about 2-Hz. The XG3 drifts approx +/- 2 Hz so difficult to set the Ref*Frequency using it. But that eliminated most of the 38 Hz offset I saw last night. Running the Tx at 5w afterward drift was in the range of 0 to +7 Hz for a 50-second key down run. The drift is not steady in one direction so average error is probably in the neighborhood of +5 Hz with 5-Hz uncertainty in the measurement. Crude but gives a ballpark idea what to expect.
Today's measurements were all with new heat sink installed. Temp Comp. probably would lower these numbers a bit.
I might try running my K3 as signal source for Temp Comp. if it holds more stable then the XG3 at 50-MHz. I am using the EXREF so it may be better?
73, Ed - KL7UW
http://www.kl7uw.com
"Kits made by KL7UW"
dubususa at gmail.com
______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:Elecraft at mailman.qth.net
This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html
Message delivered to matt.vk2rq at gmail.com
______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:Elecraft at mailman.qth.net
This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html
Message delivered to droese at necg.de
______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:Elecraft at mailman.qth.net
This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html
Message delivered to matt.vk2rq at gmail.com
Eric Swartz - WA6HHQ, Elecraft
2014-04-12 00:35:01 UTC
Permalink
The XG3 is perfect for this calibration. Freq is not critical as ling as its
close to 50. Our XG50 is actually a little below 50 Mhz. (I think it used the
ref oscillator module from the K3.)

Eric
elecraft.com
It was suggested I do the Temp Compensation Procedure. Unfortunately I do not
have a stable 50-MHz signal source. However I did run the Reference Frequency
I sampled the RF output of the XG3 with my counter and also input that to the
KX3 for the Calib. This required running at 0 dBm to inject strong enough
signal for the counter (approx -3 dBm). I turned on the ATT for the KX3 to
keep input from overdriving. Got about S9 signal level.
I got it adjusted to within about 2-Hz. The XG3 drifts approx +/- 2 Hz so
difficult to set the Ref*Frequency using it. But that eliminated most of the
38 Hz offset I saw last night. Running the Tx at 5w afterward drift was in
the range of 0 to +7 Hz for a 50-second key down run. The drift is not steady
in one direction so average error is probably in the neighborhood of +5 Hz
with 5-Hz uncertainty in the measurement. Crude but gives a ballpark idea
what to expect.
Today's measurements were all with new heat sink installed. Temp Comp.
probably would lower these numbers a bit.
I might try running my K3 as signal source for Temp Comp. if it holds more
stable then the XG3 at 50-MHz. I am using the EXREF so it may be better?
73, Ed - KL7UW
http://www.kl7uw.com
"Kits made by KL7UW"
dubususa at gmail.com
______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:Elecraft at mailman.qth.net
This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html
Message delivered to elist_copy at elecraft.com
David GM4JJJ
2014-04-12 09:41:14 UTC
Permalink
From:?Eric Swartz - WA6HHQ, Elecraft eric at elecraft.com
Reply:?Eric Swartz - WA6HHQ, Elecraft eric at elecraft.com
Date:?12 April 2014 at 01:35:33
To:?Edward R Cole kl7uw at acsalaska.net, elecraft at mailman.qth.net elecraft at mailman.qth.net
Subject:? Re: [Elecraft] KX3: thermal frequency drift

The XG3 is perfect for this calibration. Freq is not critical as ling as its?
close to 50. Our XG50 is actually a little below 50 Mhz. (I think it used the?
ref oscillator module from the K3.)?

Eric?
elecraft.com?


Eric,

Curiously the Elecraft procedurehttp://www.elecraft.com/manual/KX3%20Custom%20VFO%20TC%20rev%20A9.pdf says?

Note that the Elecraft XG3 lacks the necessary stability and is not recommended. ?



73



David GM4JJJ
Wayne Burdick
2014-04-12 15:25:41 UTC
Permalink
The XG50 will provide very consistent results. It's based on a K3 TCXO.

But a completely warmed-up, stabilized XG3 might be OK if it's kept well away from the hair-dryer that's used to heat up the rig during the procedure.

Wayne
N6KR
From: Eric Swartz - WA6HHQ, Elecraft eric at elecraft.com
Reply: Eric Swartz - WA6HHQ, Elecraft eric at elecraft.com
Date: 12 April 2014 at 01:35:33
To: Edward R Cole kl7uw at acsalaska.net, elecraft at mailman.qth.net elecraft at mailman.qth.net
Subject: Re: [Elecraft] KX3: thermal frequency drift
The XG3 is perfect for this calibration. Freq is not critical as ling as its
close to 50. Our XG50 is actually a little below 50 Mhz. (I think it used the
ref oscillator module from the K3.)
Eric
elecraft.com
Eric,
Curiously the Elecraft procedurehttp://www.elecraft.com/manual/KX3%20Custom%20VFO%20TC%20rev%20A9.pdf says
Note that the Elecraft XG3 lacks the necessary stability and is not recommended.
73
David GM4JJJ
______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:Elecraft at mailman.qth.net
This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html
Message delivered to n6kr at elecraft.com
Bruce Beford
2014-04-11 20:55:53 UTC
Permalink
Post by Edward R Cole
It was suggested I do the Temp Compensation Procedure. Unfortunately
I do not have a stable 50-MHz signal source.
Ed,
I did mine using the 5th harmonic from my Trimble Thunderbolt GPSDO. I think
I used about 60dB of attentuation between the oscillator and the rig, I
don't recall exactly. If you have a precision 10 MHz reference (and I
believe you do), you could use that similarly.

73,
Bruce N1RX
Phil Genera
2014-04-11 21:06:00 UTC
Permalink
Post by Bruce Beford
Post by Edward R Cole
It was suggested I do the Temp Compensation Procedure. Unfortunately
I do not have a stable 50-MHz signal source.
Ed,
I did mine using the 5th harmonic from my Trimble Thunderbolt GPSDO. I think
I used about 60dB of attentuation between the oscillator and the rig, I
don't recall exactly. If you have a precision 10 MHz reference (and I
believe you do), you could use that similarly.
73,
Bruce N1RX
I did the same without any attenuation, and both kx3s survived :).
--
Phil
kj6pon
Loading...