Elecraft-K3@groups.io | K3 sudden frequency jump

Bob Benedict, KD8CGH

#32945

I'm participating in the HamSCI Eclipse festival
https://hamsci.org/june-2020-eclipse-festival-frequency-measurement
which involves monitoring WWV using fldigi in frequency analysis mode. My K3 is sitting alone in the clear on a workbench powered by an Anstron 20 amp linear supply with a 17 amp hour lead acid battery backup (diode cut over type). I noticed this morning that although set to the same dial frequency, the received frequency as reported by fldigi jumped by about 10 Hz at some time last night. It then steadied at 10 Hz low. The K3 was in a closed room and not touched by anyone. Any thoughts about what happened and how to prevent this?

snippet of fldigi frequency analysis output - note change at UTC 11:48:26 and then some missing points

UTC	Freq	Freq Err	Vpk
11:48:19	10000000.143	0.143	0.0503
11:48:20	10000000.153	0.153	0.037964
11:48:21	10000000.162	0.162	0.04021
11:48:22	10000000.091	0.091	0.081219
11:48:23	10000000.011	0.011	0.137612
11:48:24	10000000.065	0.065	0.168657
11:48:25	9999999.122	-0.878	0.158187
11:48:26	9999996.705	-3.295	0.117768
11:48:27	9999993.314	-6.686	0.066463
11:48:28	9999989.708	-10.292	0.039115
11:49:17	9999989.89	-10.11	0.008997
11:49:55	9999989.675	-10.325	0.018503
11:49:56	9999990.454	-9.546	0.014056
11:49:57	9999990.923	-9.077	0.011183
11:49:58	9999990.84	-9.16	0.010979

Wes Stewart

#32946

You don't say what reference oscillator you have in your K3 or which synthesizer is installed. I'm not a fan of fldigi so I am not comfortable commenting on its performance, but if I understand correctly, it is recording the beat note of the detected audio in SSB mode. The beat note is then subject to changes in 1) the incoming frequency, 2) the RX LO, 3) the RX "BFO" and 4) the soundcard sampling frequency.

Parameters 2 and 3 are derived from the K3 reference oscillator but (I think) not in a 1:1 relationship and SSB demodulation is a DSP process and not a real a real BFO driven mixer as in an analog RX.

My takeaway is that there are too many variables to point a finger at any one of them.

Wes N7WS

Show quoted text

On Sunday, June 21, 2020, 7:09:49 AM MST, Bob Benedict, KD8CGH <rkayakr@...> wrote:

I'm participating in the HamSCI Eclipse festival
https://hamsci.org/june-2020-eclipse-festival-frequency-measurement
which involves monitoring WWV using fldigi in frequency analysis mode. My K3 is sitting alone in the clear on a workbench powered by an Anstron 20 amp linear supply with a 17 amp hour lead acid battery backup (diode cut over type). I noticed this morning that although set to the same dial frequency, the received frequency as reported by fldigi jumped by about 10 Hz at some time last night. It then steadied at 10 Hz low. The K3 was in a closed room and not touched by anyone. Any thoughts about what happened and how to prevent this?

snippet of fldigi frequency analysis output - note change at UTC 11:48:26 and then some missing points

UTC	Freq	Freq Err	Vpk
11:48:19	10000000.143	0.143	0.0503
11:48:20	10000000.153	0.153	0.037964
11:48:21	10000000.162	0.162	0.04021
11:48:22	10000000.091	0.091	0.081219
11:48:23	10000000.011	0.011	0.137612
11:48:24	10000000.065	0.065	0.168657
11:48:25	9999999.122	-0.878	0.158187
11:48:26	9999996.705	-3.295	0.117768
11:48:27	9999993.314	-6.686	0.066463
11:48:28	9999989.708	-10.292	0.039115
11:49:17	9999989.89	-10.11	0.008997
11:49:55	9999989.675	-10.325	0.018503
11:49:56	9999990.454	-9.546	0.014056
11:49:57	9999990.923	-9.077	0.011183
11:49:58	9999990.84	-9.16	0.010979

Ray Maxfield

#32947

Wes…… I agree. As a Standards Engineer, I don’t See A Standard that Is 10 X ( normally Used) to Evaluate …..

The 4 Area’s you have Defined.

Ray Maxfield WA6VAB

Sent from Mail for Windows 10

Show quoted text

From: Wes Stewart via groups.io
Sent: Sunday, June 21, 2020 10:16 AM
To: elecraft-k3@groups.io; Elecraft-K3@groups.io
Subject: Re: [Elecraft-K3] K3 sudden frequency jump

You don't say what reference oscillator you have in your K3 or which synthesizer is installed. I'm not a fan of fldigi so I am not comfortable commenting on its performance, but if I understand correctly, it is recording the beat note of the detected audio in SSB mode. The beat note is then subject to changes in 1) the incoming frequency, 2) the RX LO, 3) the RX "BFO" and 4) the soundcard sampling frequency.

Parameters 2 and 3 are derived from the K3 reference oscillator but (I think) not in a 1:1 relationship and SSB demodulation is a DSP process and not a real a real BFO driven mixer as in an analog RX.

My takeaway is that there are too many variables to point a finger at any one of them.

Wes N7WS

On Sunday, June 21, 2020, 7:09:49 AM MST, Bob Benedict, KD8CGH <rkayakr@...> wrote:

I'm participating in the HamSCI Eclipse festival
https://hamsci.org/june-2020-eclipse-festival-frequency-measurement
which involves monitoring WWV using fldigi in frequency analysis mode. My K3 is sitting alone in the clear on a workbench powered by an Anstron 20 amp linear supply with a 17 amp hour lead acid battery backup (diode cut over type). I noticed this morning that although set to the same dial frequency, the received frequency as reported by fldigi jumped by about 10 Hz at some time last night. It then steadied at 10 Hz low. The K3 was in a closed room and not touched by anyone. Any thoughts about what happened and how to prevent this?

snippet of fldigi frequency analysis output - note change at UTC 11:48:26 and then some missing points

UTC	Freq	Freq Err	Vpk
11:48:19	10000000.143	0.143	0.0503
11:48:20	10000000.153	0.153	0.037964
11:48:21	10000000.162	0.162	0.04021
11:48:22	10000000.091	0.091	0.081219
11:48:23	10000000.011	0.011	0.137612
11:48:24	10000000.065	0.065	0.168657
11:48:25	9999999.122	-0.878	0.158187
11:48:26	9999996.705	-3.295	0.117768
11:48:27	9999993.314	-6.686	0.066463
11:48:28	9999989.708	-10.292	0.039115
11:49:17	9999989.89	-10.11	0.008997
11:49:55	9999989.675	-10.325	0.018503
11:49:56	9999990.454	-9.546	0.014056
11:49:57	9999990.923	-9.077	0.011183
11:49:58	9999990.84	-9.16	0.010979

Colin Smithers

#32948

A shot in the dark: did ambient temperature move and cause a change in a temperature compensated oscillator?

73, Colin G4CWH

Show quoted text

On 21 Jun 2020, at 19:12, Ray Maxfield via groups.io <wa6vab@...> wrote:

Wes…… I agree. As a Standards Engineer, I don’t See A Standard that Is 10 X ( normally Used) to Evaluate …..

The 4 Area’s you have Defined.

Ray Maxfield WA6VAB

Sent from Mail for Windows 10

From: Wes Stewart via groups.io
Sent: Sunday, June 21, 2020 10:16 AM
To: elecraft-k3@groups.io; Elecraft-K3@groups.io
Subject: Re: [Elecraft-K3] K3 sudden frequency jump

You don't say what reference oscillator you have in your K3 or which synthesizer is installed. I'm not a fan of fldigi so I am not comfortable commenting on its performance, but if I understand correctly, it is recording the beat note of the detected audio in SSB mode. The beat note is then subject to changes in 1) the incoming frequency, 2) the RX LO, 3) the RX "BFO" and 4) the soundcard sampling frequency.

Parameters 2 and 3 are derived from the K3 reference oscillator but (I think) not in a 1:1 relationship and SSB demodulation is a DSP process and not a real a real BFO driven mixer as in an analog RX.

My takeaway is that there are too many variables to point a finger at any one of them.

Wes N7WS

On Sunday, June 21, 2020, 7:09:49 AM MST, Bob Benedict, KD8CGH <rkayakr@...> wrote:

I'm participating in the HamSCI Eclipse festival
https://hamsci.org/june-2020-eclipse-festival-frequency-measurement
which involves monitoring WWV using fldigi in frequency analysis mode. My K3 is sitting alone in the clear on a workbench powered by an Anstron 20 amp linear supply with a 17 amp hour lead acid battery backup (diode cut over type). I noticed this morning that although set to the same dial frequency, the received frequency as reported by fldigi jumped by about 10 Hz at some time last night. It then steadied at 10 Hz low. The K3 was in a closed room and not touched by anyone. Any thoughts about what happened and how to prevent this?

snippet of fldigi frequency analysis output - note change at UTC 11:48:26 and then some missing points

UTC

Freq

Freq Err

Vpk

11:48:19

10000000.143

0.143

0.0503

11:48:20

10000000.153

0.153

0.037964

11:48:21

10000000.162

0.162

0.04021

11:48:22

10000000.091

0.091

0.081219

11:48:23

10000000.011

0.011

0.137612

11:48:24

10000000.065

0.065

0.168657

11:48:25

9999999.122

-0.878

0.158187

11:48:26

9999996.705

-3.295

0.117768

11:48:27

9999993.314

-6.686

0.066463

11:48:28

9999989.708

-10.292

0.039115

11:49:17

9999989.89

-10.11

0.008997

11:49:55

9999989.675

-10.325

0.018503

11:49:56

9999990.454

-9.546

0.014056

11:49:57

9999990.923

-9.077

0.011183

11:49:58

9999990.84

-9.16

0.010979

Bob Benedict, KD8CGH

#32949

Thanks for thinking about it.

I'm running a K3 with a K3IOB board and no external reference, so its the stock synth and the sound card in the K3IOB.

temperature shift - the rig is indoors and after 36 hours the shift was over 9 seconds and then remained stable for another 12 hours (so far).

I've had the K3 for many years and I don't believe I've ever seen this in operating, or in calibrating after a long warm up or in the 48 test run last weekend.

Colin Smithers

#32950

If there is digitally implemented compensation, it could have decided that the temperature had just crossed a boundary. Perhaps it remains a long shot.

Show quoted text

From: Elecraft-K3@groups.io [mailto:Elecraft-K3@groups.io] On Behalf Of Bob Benedict, KD8CGH via groups.io
Sent: 21 June 2020 20:48
To: Elecraft-K3@groups.io
Subject: Re: [Elecraft-K3] K3 sudden frequency jump

Thanks for thinking about it.

I'm running a K3 with a K3IOB board and no external reference, so its the stock synth and the sound card in the K3IOB.

temperature shift - the rig is indoors and after 36 hours the shift was over 9 seconds and then remained stable for another 12 hours (so far).

I've had the K3 for many years and I don't believe I've ever seen this in operating, or in calibrating after a long warm up or in the 48 test run last weekend.

Bob Benedict, KD8CGH

#32952

Colin

   Thanks. I've seen discussions on temperature compensation for the KX3. I'll dig around for info on the K3.

--

73
   Bob   KD8CGH

Ray Maxfield

#32953

You might want to use WWV for a Standard, to Listen as a Signal Source.

WA6VAB

Show quoted text

From: Bob Benedict, KD8CGH
Sent: Monday, June 22, 2020 4:29 AM
To: Elecraft-K3@groups.io
Subject: Re: [Elecraft-K3] K3 sudden frequency jump

Colin

   Thanks. I've seen discussions on temperature compensation for the KX3. I'll dig around for info on the K3.

--

73
   Bob   KD8CGH

Bob Benedict, KD8CGH

#32954

Ray
In my original post I mentioned that I was monitoring WWV at 10 MHz.
--

73
Bob KD8CGH

Wes Stewart

#32955

Somewhere in the literature for the K3S I saw mention of digital correction of the TCXO. I posed a question about it and never got a reply. But the TCXO has only three pins so it can't have any.

Wes N7WS

Show quoted text

On Mon, Jun 22, 2020 at 04:28 AM, Bob Benedict, KD8CGH wrote:

Colin
   Thanks. I've seen discussions on temperature compensation for the KX3. I'll dig around for info on the K3.

--

73
   Bob   KD8CGH

Ray Maxfield

#32956

Ok, Sorry ..did not see that…… 73 wa6vab

Sent from Mail for Windows 10

Show quoted text

From: Bob Benedict, KD8CGH
Sent: Monday, June 22, 2020 12:43 PM
To: Elecraft-K3@groups.io
Subject: Re: [Elecraft-K3] K3 sudden frequency jump

Ray
In my original post I mentioned that I was monitoring WWV at 10 MHz.
--

73
Bob KD8CGH

Bob Benedict, KD8CGH

#32958

After the HamSCI Eclipse festival I turned my K3 off and let it cool down. After turning it back on and letting it warm up I saw that it was still 10 Hz low when monitoring WWV at 10 MHz. I had to recalibrate and it has remained calibrated since then.
So, it appears that something changed in my K3 during use that required a recalibration. I have serial #70 so its about 12 years old. Perhaps this is component aging.

73
Bob KD8CGH

Kent Trimble, K9ZTV

#32959

Or it may have been slightly off before the event but not noticed.

K3 serial number 21 has been used by our club's CW crew for the past twelve Field Days. I do the REFCAL routine before and after and about twice a year.

Components do age, as we all know, but I have always marveled at how well the K3 maintains its calibration. A slight touch-up every now and then is nothing to worry about.

By the way, for previous Field Days I would disable the WARC bands and any band for which the rig would not have an antenna connected (i.e., 160 and 6 meters). I would also put a dummy load on the second antenna jack because the ATU TUN button is immediately above the ANT button and sleepy operators (SSB on a friend's K3) have been known to push the wrong button. Ten minutes and no QSOs later, they realized they had pumped 100 watts into no antenna, thus earning a ticket back to Aptos.

For the newly licensed, always let your rig warm-up prior to doing any calibration procedure. The manual says "at least 15 minutes" but I personally wait 30-60 minutes before tinkering.

73,

Kent K9ZTV

On 6/24/2020 8:34 AM, someone wrote:

After the HamSCI Eclipse festival I turned my K3 off and let it cool down. After turning it back on and letting it warm up I saw that it was still 10 Hz low when monitoring WWV at 10 MHz. I had to recalibrate and it has remained calibrated since then. So, it appears that something changed in my K3 during use that required a recalibration. I have serial #70 so its about 12 years old. Perhaps this is component aging.

Virus-free. www.avg.com

Bob Benedict, KD8CGH

#32960

Kent
If you look at the original post you can see that the frequency suddenly jumped in the middle of the three day HamSCI eclipse event after being steady for more than 24 hours. I recalibrated a few days before the event so I had a fresh cal.
Agreed on the warm up - the manual warm up before cal is way to short. I let my K3 warm up at least two hours before a careful cal.

73
Bob KD8CGH