+++++++++++++++++++++++ Date: Sun, 03 Feb 2002 12:26:34 -0600 To: "Rich Lentz" , elecraft at mailman.qth.net From: Tom Hammond =?iso-8859-1?Q?N=D8SS?= Subject: Re: [Elecraft] Calibration of BFO Rich: >I am having trouble setting up the center frequency of the narrow bandpass >filters (FL3 and FL4). I have gone through the "Calibration Functions" >section of the manual and each of the various and many guides on how to use >"Spectrogram," however, it seems like one correction offsets the other. I am >fairly sure how to adjust the bandpass width, but get confused on adjusting >the position of the bandpass and the BFO. The band pass is at the desired >width but when I try to get it at the desired center frequency, the received >frequency is shifted. For example, when I try to setup a 1.00 kHz bandwith >centered at 1.00 kHz for FL4 on USB (20 Meters) for PSK, the received signal >shifts by more than 50 Hz. If left this way, a QSO ends up "walking" up the >band with each exchange. A shift in the received frequency, generally indicates that the BFO for the selectesd filter has been set to the opposite side of the filter center frequency. To determine filter center, select a moderate-to-narrow filter bandwidth, then crank the BFO from one end of its tuning range to the other (0 thru 254, or the reverse). At some point you will find that the tone of the background noise is at it's lowest point. THIS indicates the center of the filter passband. WRITE this frequency down. Now, knowing the center of your filter passband, ensure that all of your LSB and CWn BFO settings are LOWER than filter center, and that all of your USB and CW Rev BFO settings are ABOVE filter center. >I am trying to get the minimal change of the received frequency with the >desired bandpass centered on the desired frequency. Could you please explain >how to adjust: 1. The center frequency, 2. The BFO, 3. The Band-width, and >4. Which order to adjust these. Is there any interaction with these? Should >I run CALL PLL between any of these adjustments? Should the CW, USB or LSB >be set first? There is NO(!) significant interaction, if any at all. Filter center should determined first. So you'll know which side of the filter the BFO should be placed for the selected mode. Generally, you'll select the filter number, then the bandwidth, and then position the BFO to center the signal on top of the sidetone marker, or BETWEEN the passband markers if using SSB. There is NO required 'order' for setting CWn CW Rev, LSB, USB... they're all independent of each other and do NOT interact with each other. In general, you should not see a frequency of more than 20 Hz between modes, if everything is properly adjusted. Hope this helps, Tom Hammond N0SS +++++++++++++++++++++ Reply-To: "Don Wilhelm" From: "Don Wilhelm" To: "Rich Lentz" , "Elecraft Reflector" , "Gary L Surrency" Subject: Re: [Elecraft] Calibration of BFO Date: Sun, 3 Feb 2002 19:25:20 -0500 Rich, It sounds to me like you are using a single signal rather than a wideband noise source ('dead band' noise will do) when using Spectgram. When using a single carrier to attempt to set things up the result is just like you describe based on my previous experience. The problem is that the single signal may not actually be set in the exact center of the filter passband, and when you change the BFO, the display of the received frequency will shift by the amount you moved the BFO - then you retune the desired signal - and then do it all over again, and again and again .... Try using a noise generator or band noise (with no signals present) along with Spectogram and I believe you will have better luck. You may have to re-do it several times to take out a very small frequency shift (10 to 30 Hz), but you should get within 20 or 30 Hz on the first try. Yes, set the SSB BFO first - that is the BFO that is used for transmit. Ideally this BFO should be set to provide the best transmitted SSB and may depend on your particular voice characteristics. Making a tape recording with another receiver is the best way to do that evaluation. 73, Don Wilhelm - Wake Forest, NC W3FPR home page: http://www.qsl.net/w3fpr/ QRP-L # 485 K2 SN 0020 mailto: w3fpr at arrl.net ++++++++++++++++++ From: "Rich Lentz" To: "Don Wilhelm" , "Elecraft Reflector" , "Gary L Surrency" Subject: RE: [Elecraft] Calibration of BFO Date: Sun, 3 Feb 2002 20:38:41 -0600 -----Original Message----- From: elecraft-admin at mailman.qth.net [mailto:elecraft-admin at mailman.qth.net]On Behalf Of Don Wilhelm Sent: Sunday, February 03, 2002 6:25 PM To: Rich Lentz; Elecraft Reflector; Gary L Surrency Subject: Re: [Elecraft] Calibration of BFO | Rich, | It sounds to me like you are using a single signal rather than a wideband | noise source ('dead band' noise will do) when using Spectgram. | Try using a noise generator or band noise (with no signals present) along | with Spectogram and I believe you will have better luck. I set everything up as described in the KSB2 Manual. I then optimized the bandpass with spectrogram. The Settings did not change much with the Spectrogram optimization but are now about 0.30 to 0.50 different than those provided in the manual. This was needed to get the bandpass so that it started at about 500 HZ and had the desired BW (2.20, 2.0, 1.9, and 1.0 KHz on LSB/USB and 1.5, 0.70, 0.40, and 0.10 on CW and CWR) for the upper end point. LSB, USB, CW, CWR use OP1 for FL1 and USB/LSB use OP1 for FL2, just like the manual recommends. The frequency displayed in CALL Fill on LSB and CW are all the same. And this provides me with absolutely no shift of the received signal when changing filters (XFIL1 - 4). On the lower bands. However, on USB and CWR the received frequency shifts between 10 to 50 Hz. (After reading Tom N0SS post I played with it for about an hour and got it down to about 10 Hz for most filters.) This is probably not a big deal on CW as long as I select the filter before I call CQ or before I answer one. and probably not that big a deal even if I didn't. The problem is mostly with PSK when I am using a Soundcard TU. My transmitted signal will be offset by whatever the difference is because of this offset (the transmitted signal only uses FL1) and this causes the qso to walk up the band when I try and use a filter. Should I give up on this and just use LSB for PSK? Don't have the problem with RTTY as I use the normal convention for RTTY and don't use the "reverse" function of the software to correct for using USB. | Yes, set the SSB BFO first - that is the BFO that is used for transmit. I always start with XFL1 and they are within the set ability tolerance of the K2 for the manual settings. ++++++++++++++++++ Date: Mon, 04 Feb 2002 07:28:04 -0600 To: "Rich Lentz" , elecraft at mailman.qth.net From: Tom Hammond =?iso-8859-1?Q?N=D8SS?= Subject: RE: [Elecraft] Calibration of BFO Rich: >However, on USB and CWR the received frequency shifts between 10 to 50 Hz. >(After reading Tom N0SS post I played with it for about an hour and got it >down to about 10 Hz for most filters.) This is probably not a big deal on >CW as long as I select the filter before I call CQ or before I answer one. >and probably not that big a deal even if I didn't. The problem is mostly >with PSK when I am using a Soundcard TU. My transmitted signal will be >offset by whatever the difference is because of this offset (the transmitted >signal only uses FL1) and this causes the qso to walk up the band when I try >and use a filter. > >Should I give up on this and just use LSB for PSK? Don't have the problem >with RTTY as I use the normal convention for RTTY and don't use the >"reverse" function of the software to correct for using USB. Consider NOT using the OP1 filter for PSK31. You can set the other three filters FL2 thru FL4 to wide mid and narrow bandwidths as well... possibly, you can find settings which don't shift using strictly the CW filters. Note that it is NOT at all unusual to have a 10Hz - 20Hz frequency shift when switching between filters. If you managed to get it down to 0Hz shift, it was almost a fluke. Remember, PSK31 was not considered when the K2 was designed... at least, I don't htink it was considered... maybe not even available at that time. 73, Tom Hammond N0SS ++++++++++++++++++ Date: Mon, 04 Feb 2002 09:17:56 -0500 To: Tom Hammond =?iso-8859-1?Q?N=D8SS?= From: Dave Gingrich K9DC Subject: RE: [Elecraft] Calibration of BFO Cc: "Rich Lentz" , elecraft at mailman.qth.net At 07:28 2/4/2002 -0600, Tom Hammond N=D8SS wrote: >Consider NOT using the OP1 filter for PSK31. You can set the other three=20 >filters FL2 thru FL4 to wide mid and narrow bandwidths as well...=20 >possibly, you can find settings which don't shift using strictly the CW=20 >filters. Only problem is OP1 is always used on transmit so you need to minmize the=20 shift between OP1 and the chosen receive filter, even if you never use OP1= =20 to receive. >Note that it is NOT at all unusual to have a 10Hz - 20Hz frequency shift=20 >when switching between filters. If you managed to get it down to 0Hz=20 >shift, it was almost a fluke. I have found that you can get the shift reduced to less than 1 Hz by just=20 repeating the procedure numerous times. Sometimes by just adjusting one=20 filter at a time. Often I just go in to the procedure and just the BFO by= =20 one 'notch' and back again and exit, so it will recalculate the offset. I= =20 have been able to get it down to the point where there is no audible pitch= =20 change (musicians ear) between the filter settings, and just a slight phase= =20 twist on the PSK tuning eye. It WILL drive you nuts, but it can be much better than 10 - 20 Hz. Dave Gingrich, K9DC - Indianapolis, Indiana USA K2 #2211, K1 #931, QRP-L #2376, ARS #1109, FPQRP #389, IRLP #473, k9dc.ampr.org, CCIE #6748 +++++++++++++++++++ Reply-To: From: "Ron D' Eau Claire" To: Subject: RE: [Elecraft] Calibration of BFO Date: Mon, 4 Feb 2002 08:39:15 -0800 Dave, K9DC, wrote: > Actually, I believe the adjustment increment of the BFO is in 20 Hz > steps. The VFO steps are 10 Hz. My comment about shifts less than 1 Hz is > referring to frequency shifts when stepping through the four filter > bandwidths. It can be less than 1 Hz. It also will drive you > nuts getting > there. Sometimes just entering the filter alignment mode, making a change > then returning back to the same setting and exiting will get it right on, > sometimes it makes it worse. I've not discoverd the key to getting it > right except to keep repeating the procedure... and then don't mess with > it once you get it right. As I recall a discussion a couple of years ago, the 12-bit Digital to Analog Converters (DAC's) used in the K2 permit a 10 Hz resolution. That is, they can't come up with a control voltage that will set the oscillator any closer than that. If you get lucky, the DAC will pick a value that hits closer, but it's purely a matter of luck. The system can't tell the difference. And if you do anything at all to make it store the value again, it may easily store a new value that translates into a frequency error of up to 10 Hz. I've always assumed that the reverse would be true when operating the rig. Whenever the digital value representing a frequency was recalled from memory, it would be translated into an analog voltage by the DAC, but only with an accuracy that would place it within 10 Hz of the original spot. So it would not be possible to store a value that would always be more accurate than 10 Hz. I've always noted a 10 Hz step in at least one filter. Sometimes when setting up the filters they sounded and looked 'exactly on' but when I operated the rig later the 10 Hz frequency error would be back. Never more than that, but it's always been there no matter how often I did the alignment. Ron AC7AC K2 # 1289 +++++++++++++++++ From: "Randy Moore" To: "Dave Gingrich K9DC" Cc: Subject: Re: [Elecraft] Calibration of BFO Date: Mon, 4 Feb 2002 08:42:44 -0600 Dave, when you get the frequency shift between filters down to 1 Hz, does it stay there? I have noticed changes with time that are greater than 10 Hz on my K2. 73, Randy, KS4L ----- Original Message ----- From: "Dave Gingrich K9DC" ...I have found that you can get the shift reduced to less than 1 Hz by just repeating the procedure numerous times. +++++++++++++++++++ Date: Mon, 4 Feb 2002 10:14:20 -0500 (EST) From: Dave Gingrich K9DC To: Randy Moore Cc: elecraft at mailman.qth.net Subject: Re: [Elecraft] Calibration of BFO On Mon, 4 Feb 2002, Randy Moore wrote: > Dave, when you get the frequency shift between filters down to 1 Hz, does it > stay there? I have noticed changes with time that are greater than 10 Hz on > my K2. If you are referring to BFO calibration, no I've noticed no change over time. Only when I change something, I usually screw it all up and have to start over... Remember - Patience is a virtue. :) ================================================ Dave Gingrich, K9DC - Indianapolis, Indiana USA ++++++++++++++++++++ Date: Mon, 04 Feb 2002 10:25:18 -0500 From: "David A. Belsley" To: Dave Gingrich K9DC , =?ISO-8859-1?Q?Tom_Hammond_N=D8SS?= Cc: Rich Lentz , elecraft at mailman.qth.net Subject: RE: [Elecraft] Calibration of BFO Dave: My experience differs somewhat with yours. I find that if I attempt to=20 adjust a single filter, the shift can (and usually does) become larger.=20 So, if I want to adjust a single filter, I adjust that filter and then I=20 reset all the filters, starting with number 1, in succession (including the = one just adjusted). By resetting the filter I mean: I note its setting,=20 move the setting several numbers down and then back again to the original=20 and shift to the next filter (I do this for both the c and cbar filters=20 when doing this on cw). The least shifting occurs for me after I have set=20 all filters in this manner in rapid succession. This procedure works well=20 even if you have not reset any filter but simply notice after a period of=20 time that there seems to be increased shifting. Simply go and reset every=20 filter back to its original value. When I do this, I notice that the shift = is either gone altogether (rarely) or has only one 10Hz jump somewhere=20 along the line (typical). best wishes, dave belsley, w1euy ++++++++++++++++ Date: Mon, 04 Feb 2002 10:30:35 -0500 From: "David A. Belsley" To: Randy Moore , Dave Gingrich K9DC Cc: elecraft at mailman.qth.net Subject: Re: [Elecraft] Calibration of BFO Randy: I'm not sure what the meaning is of shifts less than 1 Hz -- other than 0Hz. It is my understanding that the BFO granularity is 10Hz, so the shift would be some multiple (including zero) of 10Hz. And yes, Randy, there can be changes with time. See my previous post for the method that has been reasonably successful for me in reestabilishing a good relation among the filters. best wishes, dave belsley, w1euy +++++++++++++++++++ Date: Mon, 4 Feb 2002 10:53:19 -0500 (EST) From: Dave Gingrich K9DC To: "David A. Belsley" Cc: Randy Moore , Subject: Re: [Elecraft] Calibration of BFO On Mon, 4 Feb 2002, David A. Belsley wrote: > Randy: > I'm not sure what the meaning is of shifts less than 1 Hz -- other than > 0Hz. It is my understanding that the BFO granularity is 10Hz, so the shift > would be some multiple (including zero) of 10Hz. Actually, I believe the adjustment increment of the BFO is in 20 Hz steps. The VFO steps are 10 Hz. My comment about shifts less than 1 Hz is referring to frequency shifts when stepping through the four filter bandwidths. It can be less than 1 Hz. It also will drive you nuts getting there. Sometimes just entering the filter alignment mode, making a change then returning back to the same setting and exiting will get it right on, sometimes it makes it worse. I've not discoverd the key to getting it right except to keep repeating the procedure... and then don't mess with it once you get it right. Once right, the shift between filters is stable... Which is not the same as overall frequency stability of the rig. The K2 can drift a little especially with temp changes, but it is generally much less 50 Hz so it has never been a problem for me. ================================================ Dave Gingrich, K9DC - Indianapolis, Indiana USA ++++++++++++++++++ Date: Mon, 04 Feb 2002 11:56:28 -0500 From: "David A. Belsley" To: rondec at easystreet.com, elecraft at mailman.qth.net Subject: RE: [Elecraft] Calibration of BFO Ron: This is my experience exactly. dave --On Monday, February 4, 2002 8:39 AM -0800 Ron D' Eau Claire wrote: > Dave, K9DC, wrote: > >> Actually, I believe the adjustment increment of the BFO is in 20 Hz >> steps. The VFO steps are 10 Hz. My comment about shifts less than 1 Hz is >> referring to frequency shifts when stepping through the four filter >> bandwidths. It can be less than 1 Hz. It also will drive you >> nuts getting >> there. -- snip -- > > As I recall a discussion a couple of years ago, the 12-bit Digital to > Analog Converters (DAC's) used in the K2 permit a 10 Hz resolution. That > is, they can't come up with a control voltage that will set the > oscillator any closer than that. > -- snip -- > > Ron AC7AC > K2 # 1289 ++++++++++++++++ Date: Mon, 04 Feb 2002 11:24:56 -0600 To: Dave Gingrich K9DC From: Tom Hammond =?iso-8859-1?Q?N=D8SS?= Subject: RE: [Elecraft] Calibration of BFO Cc: "Rich Lentz" , elecraft at mailman.qth.net Dave: Yep.. you're RIGHT there... I sometimes tend to forget the 'details'...=20 thanks for bringing me back to earth. Tom At 08:17 AM 2/4/02, Dave Gingrich K9DC wrote: >At 07:28 2/4/2002 -0600, Tom Hammond N=D8SS wrote: > >>Consider NOT using the OP1 filter for PSK31. You can set the other three= =20 >>filters FL2 thru FL4 to wide mid and narrow bandwidths as well...=20 >>possibly, you can find settings which don't shift using strictly the CW=20 >>filters. --- snip -- +++++++++++++++++++++ Date: Tue, 05 Feb 2002 07:40:31 -0600 To: Elecraft Mailing List From: Tom Hammond =?iso-8859-1?Q?N=D8SS?= Subject: [Elecraft] Calibration of BFO - FWD from KI6WX Folks: The following message was posted by John Grebenkemper, KI6WX, but for some reason it was being rejected by John's e-mail server. So, I post it on his behalf. - Tom N0SS John Grebenkemper, KI6WX, writes: ----- Original Message ----- From: "Ron D' Eau Claire" To: Sent: Monday, February 04, 2002 8:39 AM Subject: RE: [Elecraft] Calibration of BFO > > As I recall a discussion a couple of years ago, the 12-bit Digital to > Analog Converters (DAC's) used in the K2 permit a 10 Hz resolution. > That is, they can't come up with a control voltage that will set the > oscillator any closer than that. The PLL reference that controls the VFO frequency is tuned by a 12 bit DAC, which has about 3000 steps that are useful. For the worst case tuning on 10 meters, this will yield a frequency resolution of about 8 Hz. It is much better on the lower bands, getting down to about 2 Hz on 80 meters. The BFO tunes with an 8 bit DAC and will have a tuning step size between 23 Hz to 32 Hz in the frequency range of interest. The curve is pretty similar whether you are using the solenoidal L33 or the toriodal L33. There are two cases of interest when running the radio in SSB mode. The first is to use the same filter for both transmit and receive. In this case, both oscillators are left unchanged and the radio should transmit on exactly the same frequency on which it was receiving unless there is some effect that pulls one of the VCXO's. I have not tried this experiment, so I don't know if there are any effects that would pull a VCXO frequency between transmit and receive. In the second case, a different filter is used to receive. Unless you use the exact same BFO frequency, this will result in a frequency difference in transmission. From trying to understand what the firmware does, I believe that Wayne adjusts both the BFO and VFO frequencies to get as close as he can to the same transmission frequency, but there will always be some error. Furthermore, the firmware assumes that the frequencies are the same as when you last did a CAL PLL and CAL BFO. These will almost certainly have drifted if those measurements have not been done very recently. The best way to avoid offset frequencies is to use the same filter for transmit and receive, or to set the BFO for a different receive filter to the same BFO frequency as the transmit filter. If you do the latter, be sure to reset both the transmit and receive BFO's at the same time so the firmware knows they are at the same frequency. -John KI6WX +++++++++++++++++++ Reply-To: From: "Ron D' Eau Claire" To: "Elecraft Mailing List" Subject: RE: [Elecraft] Calibration of BFO - FWD from KI6WX Date: Tue, 5 Feb 2002 07:25:45 -0800 John, KI6WX, wrote: > The PLL reference that controls the VFO frequency is tuned by a > 12 bit DAC, > which has about 3000 steps that are useful. For the worst case > tuning on 10 > meters, this will yield a frequency resolution of about 8 Hz. It is much > better on the lower bands, getting down to about 2 Hz on 80 > meters. I looked up what Elecraft had to say about the resolution. According to the K2 Manual, Theory of Operation, "Signal Flow" section (Page 102 in my Rev. C, April 19, 2000 manual) it says, "... 10 Hz increments are provided by a 12-bit DAC driving an 11 MHz VCXO (PLL reference oscillator)." Since this oscillator is operating over the same frequency range regardless of the band in use, so I don't see how the band selected affects the resolution. I would expect 10 Hz on any band. > The best way to avoid offset frequencies is to use the same filter for > transmit and receive, or to set the BFO for a different receive filter to > the same BFO frequency as the transmit filter. If you do the latter, be > sure to reset both the transmit and receive BFO's at the same time so the > firmware knows they are at the same frequency. I think we are talking "Apples and Oranges" here. I was talking about the fact that commonly there will be a 10 Hz step between two different i-f filters as one steps through them during receive - not about a difference in the receive passband center frequency and the transmit frequency. You make a good point about the stability of the oscillators also causing a shift in the BFO frequencies as one steps through the filters during receive. Ron AC7AC K2 # 1289 +++++++++++++++ Date: Tue, 5 Feb 2002 10:00:00 +0000 From: "Julian, G4ILO" Subject: Re: [Elecraft] Calibration of BFO To: "Elecraft List" I've noticed a shift after thinking I'd pretty much eliminated it. Perhaps I don't have the necessary patience, but I've come to the conclusion that eliminating the shift between filters on my K2 is impossible. If others have achieved it, it is probably just because on their K2 the DAC steps just happen to coincide with the exact amount of voltage needed to shift their BFO the right amount, or something. So some people will manage it, and others never will. Although if Ron's theory is right, it's a matter of luck for all of us. The CW filters were the worst, I found. The only way to eliminate a really noticeable pitch change between two of the filters was to choose a different center frequency, which presumably moved the BFO voltage/frequency relationship to a different part of the curve. I really hope the K3 will have true 1Hz resolution! :-) By the way, does everyone see this =20 =3D garbage in messages, or is it only in the digests? It does make some posts hard to read. -- Julian, G4ILO. (RSGB, ARRL, K2 #392) Homepage: http://www.qsl.net/g4ilo "Randy Moore" wrote: Dave, when you get the frequency shift between filters down to 1 Hz, does it stay there? I have noticed changes with time that are greater than 10 Hz on my K2. ++++++++++++++++ From: "Dan Barker" To: "Elecraft" Subject: [Elecraft] Calibration of BFO Date: Tue, 5 Feb 2002 11:11:39 -0500 Your mail reader should turn =3D20 to a space, and =3D3D to an equals = sign for you. Thaose are standard "Quoted Printable" escape strings. Dan / WG4S / K2 #2456 -----Original Message----- From: elecraft-admin at mailman.qth.net [mailto:elecraft-admin at mailman.qth.net]On Behalf Of Julian, G4ILO Sent: Tuesday, February 05, 2002 5:00 AM To: Elecraft List Subject: Re: [Elecraft] Calibration of BFO -- snip -- By the way, does everyone see this =3D20 =3D3D garbage in messages, or = is it only in the digests? It does make some posts hard to read. -- Julian, G4ILO. (RSGB, ARRL, K2 #392) +++++++++++++++++++++++ Date: Tue, 05 Feb 2002 23:57:42 -0800 From: "John Grebenkemper, KI6WX" Subject: Re: [Elecraft] Calibration of BFO - FWD from KI6WX To: Elecraft Mailing List ----- Original Message ----- From: "Ron D' Eau Claire" To: "Elecraft Mailing List" Sent: Tuesday, February 05, 2002 7:25 AM Subject: RE: [Elecraft] Calibration of BFO - FWD from KI6WX > John, KI6WX, wrote: > > > The PLL reference that controls the VFO frequency is tuned by a > > 12 bit DAC, > > which has about 3000 steps that are useful. For the worst case > > tuning on 10 > > meters, this will yield a frequency resolution of about 8 Hz. It is much > > better on the lower bands, getting down to about 2 Hz on 80 > > meters. > > I looked up what Elecraft had to say about the resolution. According to the > K2 Manual, Theory of Operation, "Signal Flow" section (Page 102 in my Rev. > C, April 19, 2000 manual) it says, "... 10 Hz increments are provided by a > 12-bit DAC driving an 11 MHz VCXO (PLL reference oscillator)." > > Since this oscillator is operating over the same frequency range regardless > of the band in use, so I don't see how the band selected affects the > resolution. I would expect 10 Hz on any band. > The K2 manual only gives an approximation of the operation of the VFO circuit. The exact operation is considerably more complicated. I will run through exactly how you determine the VFO step size due to the DAC granularity by band; please refer to the K2's schematic to see the parts of the circuit to which I refer. The basic frequency determining element of the VFO is the PLL reference oscillator. This oscillator tunes roughly (exact frequencies may vary by 1-2 kHz) from 12081 kHz to 12103 kHz. Only the top 12.5 kHz of the PLL reference is used as a reference frequency for the VFO. This maximum tuning range of the PLL reference is determine by the lowest potential frequency of operation, which is 0 Hz (VFO=4.915 MHz; Wayne was thinking of future possibilities to make sure that the radio oscillator circuit could support any future bands below 160 meters). The top part of the tuning range is picked because this will minimize phase noise, improve frequency stability and decrease the step size. The upper 12.5 kHz of this range is calibrated in the CAL PLL setup in version 2 of the firmware. The PLL reference frequency is tuned by a 12 bit DAC (U5) on the RF board. Only the upper ~2800 steps of the DAC are used to tune this 12.5 kHz range. At the lowest frequency (12090 kHz), each step size corresponds to a 6.8 Hz step, while at the highest frequency of 12103 kHz, the step size is 2.9 Hz. The output of the PLL reference goes to the PLL Synthesizer which generates the VFO frequency in a classic phase lock loop with dividers on both the input reference and the output oscillator. The divisor ratios are only changed for every 5 kHz increment in the VFO output, and the PLL reference frequency is changed for the smaller step sizes. This is an ingenious circuit to get a small step size in a single loop PLL. When the radio is tuned to 1.8 MHz, the VFO is tuned to 6.715 MHz. This requires that the PLL reference tune over the upper 9 kHz, or a frequency range of 12094 kHz to 12103 kHz. At 12094 kHz, the PLL reference step size is about 5.7 Hz which translates into a worst case 2.3 Hz step in the VFO. When the PLL reference is at 12103 kHz, the step size will be 1.2 Hz. The firmware chooses the step that is closest the desired 10 Hz increment. When the radio is tuned to 3.5 MHz, the VFO is tuned to 8.415 MHz. This frequency requires that the PLL reference tune over a 7.2 kHz range. As the tuning range of the PLL reference narrows, the firmware requires that the lowest frequency of the PLL reference to not exceed 12095 kHz, so in this case the PLL reference will tune from 12095 kHz to 12102.2 kHz. The worst case VFO step size is now 3.8 Hz. Rather than repeating this for every band, I will now jump up to 10 meters. At 30 MHz, the VFO is tuned to 25.085 MHz and the PLL reference tunes from 12095 kHz to 12097.4 kHz. The worst case VFO step size is 11.4 Hz. It is important to remember that the radio firmware tunes the VFO in nominal 10 Hz steps. The firmware selects a PLL reference frequency and PLL Synthesizer divisor ratios to tune it to the nearest 10 Hz. Because of the granularity of the PLL reference DAC, you can not achieve precise 10 Hz steps. The radio will track closer to 10 Hz steps on the lower frequency bands because the VFO can be tuned in smaller increments by the firmware. At 30 MHz, you will get a number of 11 Hz steps followed by a 0 Hz step. At the lower end of 10 meters you will get a number of 9 Hz steps followed by a 20 Hz step. You may also get a odd step size every 5 kHz (generally between 0 to 20 Hz depending on how much your oscillators have drifted from your last CAL PLL). I think these are very minor penalties to pay for the much greater advantage of the single loop PLL circuit and its lack of off frequency spurs. -John KI6WX +++++++++++++++++ From: "Rich Lentz" To: =?iso-8859-1?Q?Tom_Hammond_N=D8SS?= , "Dave Gingrich K9DC" Cc: Subject: RE: [Elecraft] Calibration of BFO Date: Mon, 4 Feb 2002 19:41:08 -0600 Tom, Dave, Elecraft group; Thanks for the help. Like you said, it's a matter of trial and error. And sometimes all it was was making sure that each filter got reset in memory and not just the one that seemed out of whack with the rest. Took about another hour tonight, but I think I got it as close as is possible. On lower bands (below the reversal of BFO shift) got the following results reading the frequency off of the MIXW 2 display. (Seems to provide a finer resolution than Spectrogram.) FL1 FL2 FL3 FL4 USB REF -4 -6 -10 LSB REF 0.0 0.0 0.0 CW REF -5 -5 -5 CWR REF -7 -7 -7 USB, CW, and CWR all seem to provide absolutely no change in pitch (Using 750 Hz) Listening to a steady carrier on the air with RF gain and ATT set so that it is just audible and at normal listing level. Will probably play some more over the weekend (After I write down all the numbers.) and see if it possible to get closer. Messing around with PSK, It looks to me as there is no reason not to use LSB (on the lower bands), especially if you intend on dropping to a narrow filter. > >I have found that you can get the shift reduced to less than 1 Hz by just >repeating the procedure numerous times. Sometimes by just adjusting one >filter at a time. Often I just go in to the procedure and just the BFO by >one 'notch' and back again and exit, so it will recalculate the offset. I >have been able to get it down to the point where there is no audible pitch >change (musicians ear) between the filter settings, and just a slight >phase twist on the PSK tuning eye. > >It WILL drive you nuts, but it can be much better than 10 - 20 Hz. ++++++++++++++++++ From: DolfinDon at aol.com Date: Thu, 14 Feb 2002 09:01:36 EST Subject: Re: [Elecraft] K2 BFO Frequencies To: K4IA at aol.com, elecraft at mailman.qth.net In a message dated 2/14/2002 7:41:02 AM Central Standard Time, K4IA at aol.com writes: I think I am OK now although I have an issue because my tone changes upwards when going from F1/F2 to F3/F4. That I know isn't right but I am not quite sure what to do with it. ------------------- Hi There are probably other methods but here is the way I do it. Set a Spectrogram marker at your sidetone frequency. Center the bandpass for each filter on that marker for CWn and CWr. Now with the antenna disconnect find the internal birdie at around 7.000 MHz. Go out of menu to normal and adjust the VFO so the birdie on the Spectrogram display is on top of the marker. Now without touching the VFO go into cal filter and make a small adjustment to each filter so the birdie is on top of the marker for each filter in CWn and CWr. When all is set up properly you can zero beat a CW signal and switch between filters or CWn or CWr with no change in tone. Don Brown KD5NDB ++++++++++++++++++++++ From: K4IA at aol.com Date: Thu, 14 Feb 2002 19:14:12 EST Subject: Re: [Elecraft] K2 BFO Frequencies To: DolfinDon at aol.com Cc: elecraft at mailman.qth.net Don Fantastic idea and it got all my CW tones to be the same no matter what filter I use. This tip should be on the Elecraft web page. Radio K4IA Craig Buck Fredericksburg, Virginia USA QRP ARCI #2550 FISTS #6702 CC 788 Diamond #64 K1 #470 K2 #2460 ++++++++++++++++++++++ Date: Fri, 15 Feb 2002 05:39:29 -0500 To: DolfinDon at aol.com From: Charles Greene Subject: Re: [Elecraft] K2 BFO Frequencies Cc: K4IA at aol.com, elecraft at mailman.qth.net At 09:01 AM 2/14/2002 -0500, DolfinDon at aol.com wrote: Don, True, you can get the CW tones very close, but you still cannot get them closer than 5 to 8 Hz, which is still a problem for PSK31. >In a message dated 2/14/2002 7:41:02 AM Central Standard Time, K4IA at aol.com >writes: > > > I think I am OK now >although I have an issue because my tone changes upwards when going from >F1/F2 to F3/F4. That I know isn't right but I am not quite sure what to do >with it. > > >==================== > > >Hi > >There are probably other methods but here is the way I do it. Set a >Spectrogram marker at your sidetone frequency. Center the bandpass for each >filter on that marker for CWn and CWr. Now with the antenna disconnect find >the internal birdie at around 7.000 MHz. Go out of menu to normal and adjust >the VFO so the birdie on the Spectrogram display is on top of the marker. Now >without touching the VFO go into cal filter and make a small adjustment to >each filter so the birdie is on top of the marker for each filter in CWn and >CWr. When all is set up properly you can zero beat a CW signal and switch >between filters or CWn or CWr with no change in tone. > > >Don Brown >KD5NDB 73, Chas, W1CG ++++++++++++++