+++++++++++++++++++++++++++++++++++++++++++++++++++ Date: Mon, 5 Mar 2001 11:21:27 EST From: LHLOUSEK at aol.com Subject: [Elecraft] K2 Filter Drift I was helping a friend align his K2s filters with Spectrogram when I noticed that the center frequency of the 100 Hz filter drifted upward about 75 to 100 Hz on warm up. I verified this on my K2 and Ron AC7AC independently verified on his. I had centered the 100 Hz filter on the sidetone. The next day when I turned on the units (still hooked to Spectrogram) I saw that the 100Hz filter was centered about 100 Hz low on both K2s and was attenuating signals at the side tone frequency by about 10db. I continued to monitor the 100 Hz filters with Spectrogram and after about 30 minutes they became centered again. I heated my K2 a little under a lamp and the pass band drift up about another 50~75 Hz. The VFO accuracy on WWV at 10MHz did not change more than 10 Hz over this period and with the heating. This is certainly not very significant on the wider settings but it can be on the 100Hz filter. In the past, when operating, I had noticed that at times the 100Hz filter was attenuating the signal quite a bit and the band pass frequency seemed low. To compensate I would use the RIT to move the signal down if frequency. I never made the correlation with warm up until now. A drifting VCO or BFO would change the VFO frequency calibration and since that did not happen on the three K2s tested I would suspect that it is actually the filter that is drifting and not the oscillators. Luigi W7DZN +++++++++++++++++++++++++++++++++++++++++++++++++++++ Date: Mon, 05 Mar 2001 09:49:13 -0800 From: Wayne Burdick Subject: Re: [Elecraft] K2 Filter Drift Luigi, W7DZN wrote: > > I was helping a friend align his K2s filters with Spectrogram when I noticed > that the center frequency of the 100 Hz filter drifted upward about 75 to 100 > Hz on warm up. ... > A drifting VCO or BFO would change the VFO frequency calibration and since > that did not happen on the three K2s tested I would suspect that it is > actually the filter that is drifting and not the oscillators. Hi Luigi, The passband of the CW filter is varactor-tuned, and at 100 Hz, the varactor bias voltage is very low (probably < 2 V). The temperature stability of hyperabrupt junction varactors is poor, and will be most noticeable at low bias voltages, so some change in passband during warmup is to be expected. Also, both the PLL reference oscillator and BFO are also varactor-tuned, so they too will contribute a small amount of warm-up drift. Short of adding a number of temperature-compensation components or a small oven, I don't think there's anything that we could do about this. Hopefully it isn't too much of an inconvenience. 73, Wayne N6KR +++++++++++++++++++++++++++++++++++++++++++++++++++++ Date: Mon, 05 Mar 2001 10:44:35 -0700 From: Larry East Subject: Re: [Elecraft] K2 Filter Drift At 11:21 AM 3/5/01 -0500, you wrote: >I was helping a friend align his K2s filters with Spectrogram when I noticed >that the center frequency of the 100 Hz filter drifted upward about 75 to 100 >Hz on warm up. Yes, the BFO drifts -- especially so now that a toroid is used for the inductor in the BFO circuit. The drift was less with the original shielded choke inductor. With the original inductor, it drifts down about 50-60Hz on warmup. I pointed that out several months ago... Oh well. The bottom line is that you need to make sure that the radio has been running for at least 15 minutes before starting filter alignment. It also means that if the radio is subjected to low temperatures during in-the-field operation, filter widths less than 200Hz or so may become unusable -- and your receive offset will no longer match your sidetone. 72/73, Larry W1HUE/7 ++++++++++++++++++++++++++++++++++++++++++++++++++ Date: Mon, 5 Mar 2001 13:04:32 EST From: LHLOUSEK at aol.com Subject: Re: [Elecraft] K2 Filter Drift Hi Wayne, << I don't think there's anything that we could do about this. Hopefully it isn't too much of an inconvenience. >> It's not a significant inconvenience, just something to be aware of. The filter works quite well but a note in the manual might help prevent the uninitiated from wondering where the signal went when he switched to the 100Hz filter. Luigi W7DZN ++++++++++++++++++++++++++++++++++++++++++++++++++ Date: Mon, 05 Mar 2001 10:36:58 -0800 From: Wayne Burdick Subject: Re: [Elecraft] K2 Filter Drift Larry, The temperature coefficient for the new BFO toroid is much lower than that of the original shielded choke (about 300 ppm/degree vs. 1000-1500 ppm). So any noticeable BFO drift is probably related to the varactor bias. The varactor diode TC dwarfs that of any other components in the circuit, and varies all over the map with applied bias voltage. 73, Wayne Larry East wrote: > Yes, the BFO drifts -- especially so now that a toroid is used for the > inductor in the BFO circuit. The drift was less with the original shielded > choke inductor. With the original inductor, it drifts down about 50-60Hz on > warmup. +++++++++++++++++++++++++++++++++++++++++++++++++++ Date: Mon, 05 Mar 2001 10:42:45 -0800 From: Wayne Burdick Subject: Re: [Elecraft] K2 Filter Drift LHLOUSEK at aol.com wrote: > The filter works quite well but a note in the manual might help prevent the > uninitiated from wondering where the signal went when he switched to the > 100Hz filter. > > Luigi W7DZN We'll definitely add a note to the manual about this--thanks. Wayne +++++++++++++++++++++++++++++++++++++++++++++++++ Date: Mon, 05 Mar 2001 12:10:12 -0700 From: Larry East Subject: Re: [Elecraft] K2 Filter Drift At 10:36 AM 3/5/01 -0800, Wayne Burdick wrote: >Larry, > >The temperature coefficient for the new BFO toroid is much lower than >that of the original shielded choke (about 300 ppm/degree vs. 1000-1500 >ppm). So any noticeable BFO drift is probably related to the varactor >bias. The varactor diode TC dwarfs that of any other components in the >circuit, and varies all over the map with applied bias voltage. I found that the drift of the original choke was opposite to that of the varactor so that they almost canceled. The drift of the toroid and varactor are in the same direction so they add, resulting in more drift. At least that's the case in my radio. I know that the toroid inductor results in better carrier suppression on SSB. With the original inductor mounted slightly off the PCB and with proper tweaking of the SSB filter, I get 39.5 +/- 0.5 dB carrier suppression on 40M and greater than 40 dB on all other bands (relative to 10W output). I consider that completely acceptable. With the toroid inductor, I got about 2 dB greater carrier suppression on 40M. Since I rarely operate SSB on 40M, I decided to opt for the configuration that resulted in the least amount of drift. I spent several hours trying to reduce the drift with NPO caps, but obtained no significant improvement. Anyway, 50-60Hz drift ain't all that bad. As you pointed out, the varactor temperature coefficient is voltage dependent, thus my results may not be typical. 73, Larry W1HUE/7 K2 #491 +++++++++++++++++++++++++++++++++++++++++++++++++++++ Date: Mon, 5 Mar 2001 11:36:41 -0800 From: "Ron D'Eau Claire" Subject: RE: [Elecraft] K2 Filter Drift LHLOUSEK at aol.com wrote: > The filter works quite well but a note in the manual might help prevent the > uninitiated from wondering where the signal went when he switched to the > 100Hz filter. > > Luigi W7DZN We'll definitely add a note to the manual about this--thanks. Wayne ======== It's sufficient 'drift' in my K2 sitting in an evenly heated room (the indoor hamshack) that the 100 Hz bandpass "doesn't work" for about 15 minutes after the rig is turned on. Signals are well down on one side of the filter response curve. That's a period of time when I'm usually "spooking the band" and with wider filter bandwidths, so I hadn't noticed it until Luigi reported it. It takes my K2 about 30 minutes or a little more of warm-up for the bandpass to become properly centered on the sidetone frequency at 100 Hz bandwidth, but the filter is quite useable after 10 or 15 minutes. Thanks for the explanation, Wayne. It had me stumped too! Ron AC7AC K2 # 1289 +++++++++++++++++++++++++++++++++++++++++++++++ Date: Sun, 27 May 2001 11:09:18 -0700 From: Louis Hlousek Subject: Re: [Elecraft] Noisy K2. Hi Ken. Have you used Spectrogram to align your filters? It really helps get the filter centered on the side tone. Even when the filter is aligned on it's peak I find that when set to 100Hz the signal gets knocked down more than 3 db. Also, when set to 100 Hz the filter is actually more like 150 Hz wide at -6 dB, 200 Hz wide at -10dB and about 275Hz wide at -20dB. Another thing I found is that the passband of the K2 drifts upward for about 20 minutes when first turned on. It drifts about 75 Hz on the two K2s I've tested and this has been confirmed by others. It is enough so that if you center the filter on the side tone when warmed up the signal will be attenuated about -10 to -15 dB when the unit is first turned on cold. BTW. My comment about the FT1000MP reflector was meant to be humorous. "The K2 is generating a lot of noise because it has a quiet receiver." Get it? Lou W7DZN ++++++++++++++++++++++++++++++++++++++++ Date: Sun, 27 May 2001 16:15:33 -0600 From: Larry East Subject: Re: [Elecraft] Noisy K2. At 11:09 AM 5/27/01 -0700, Louis Hlousek wrote: >Even when the filter is aligned on it's peak I find that when set to >100Hz the signal gets knocked down more than 3 db. A good deal more -- typically 6-10 dB. There is a mod to reduce the attenuation at narrow filter settings at the expense of flatness of wider filter settings. Tried to install it last night but screwed it up ... guess my ability to concentrate after 10PM is degrading. >Another thing I found is that the passband of the K2 drifts upward for >about 20 minutes when first turned on. It drifts about 75 Hz on the two >K2s I've tested and this has been confirmed by others. It is enough so >that if you center the filter on the side tone when warmed up the signal >will be attenuated about -10 to -15 dB when the unit is first turned on >cold. Actually, it is the BFO that drifts. Very good point, however -- wait AT LEAST 30 MINUTES AFTER POWER-UP BEFORE MESSING WITH FILTER SETTINGS! :-) L. ++++++++++++++++++++++++++++ Date: Sun, 27 May 2001 15:07:05 -0700 From: Louis Hlousek Subject: Re: [Elecraft] Noisy K2. <<>> That was said the first time I brought up the filter drift issue but I pointed out that on at least three K2s the frequency calibration as determined by tuning to the WWV carrier in CW mode at 10MHz didn't drift during warm-up. The only way for this to happen would be for the VCO to drift the same amount but in the opposite direction as the BFO to cancel the BFO drift. That indeed could be happening, I guess, but the end result is that the frequency calibration does not drift but the filter position does. Lou W7DZN +++++++++++++++++++++++++++++++++++ Date: Sun, 27 May 2001 19:04:43 -0600 From: Larry East Subject: Re: [Elecraft] Noisy K2. At 03:07 PM 5/27/01 -0700, Louis Hlousek wrote: ><<>> > >That was said the first time I brought up the filter drift issue but I >pointed out that on at least three K2s the frequency calibration as >determined by tuning to the WWV carrier in CW mode at 10MHz didn't drift >during warm-up. The only way for this to happen would be for the VCO to >drift the same amount but in the opposite direction as the BFO to cancel >the BFO drift. That indeed could be happening, I guess, but the end >result is that the frequency calibration does not drift but the filter >position does. That is indeed the case, as you can verify with a stable EXTERNAL frequency counter (the internal counter also drifts during warmup). L. +++++++++++++++++++++++++++++++++++