The K2 will perform CAL FIL when you select it from the MENU. But unlike CAL PLL, nothing happens until you make a change to a parameter and then that parameter is re-written into the BFO look-up table. If you are setting up the filters and BFOs for the first time, you will likely alter each value during the process. The BFO frequency and its corresponding DAC input code is written into the table for easy access when the K2 wants to show the operating frequency in the display.
Every time you change the reference oscillator setting and run CAL PLL to correct for a frequency display error, you must also run CAL FIL and change the BFO settings to write new values into the table. Actually you don’t have to modify the BFO values, but you must change them even if they are correct. When they are already correct, you cycle through all the filters and BFOs, moving each BFO frequency up one notch and then down again. This will cause the K2 to sense that the BFO has been changed and rewrite the table entry – using the current reference oscillator frequency for the clock.
The K2 can have up to 24 BFO settings depending on the number of filters (either 3 or 4) you have set in the menu parameters and the number of modes (CW, SSB and RTTY) that you are using. Each one must be changed if you need to set new values in the table in order to achieve good dial calibration. There is one set of circumstances where this will not be necessary. If you had set the reference oscillator accurately just before the BFOs were set the last time, and this time you set the reference oscillator to the exact same frequency as the last time, the current entries will still be valid.
That is all you need to do when you run CAL FIL and make no real changes in the relationship between the filter passband and its associated two BFO frequencies.
If you were not satisfied with the way your filters were sounding and/or you observe a significant change in pitch on a received signal when changing from one filter to another, it may be time to re-do the entire filter alignment process.
There are good documents on the Elecraft website http://www.elecraft.com/ describing the process of aligning the filters. Look under the BUILDERS RESOURCES section, navigate to the ‘Application Notes’, and then select ‘K2 Filter Adjustment and Mod Articles’. Or, you can use a direct link - http://www.elecraft.com/Apps/new_fil_docs/filter_apps.htm. View or download any documents that you need. There are 2 documents in .pdf format (one for CW filters and the other for SSB) created by Tom Hammond (N0SS) that I recommend highly. You may also download the latest version of Spectrogram software from that page. Spectrogram 5.17 (the last freeware version) is entirely adequate for aligning the K2 and can be downloaded from Tom Hammond's website www.n0ss.net.
Tom has provided us with a great deal of documentation about Spectrogram. One example with configuration information and additional download sources is included in Tom’s reflector posting dated June 16, 2003. You can find others in his documentation and in the Elecraft reflector archives.
You will need some form of wideband noise input for the K2 – a wideband noise generator like the Elecraft N-GEN is the most convenient, but if you don’t have one you can put a short antenna wire alongside a fluorescent lamp (with the lamp lighted), or you can even use ‘dead band’ noise. If you chose to use the ‘dead band’ noise, do be certain that there are no signals that the K2 can hear since a single frequency signal will disrupt the patterns that are detected and may produce spectrum displays that are confusing.
The other thing needed is a way to display the audio spectrum coming from the K2. The Spectrogram program is recommended simply because most K2 owners use it. Other programs are available that can produce a display of the audio spectrum. I prefer using the line display (similar to that shown on an RF spectrum analyzer) rather than the waterfall display. A waterfall display will work fine for the narrower CW filters, but for the wider filters (filter widths more than 1.8 times your sidetone pitch), the filter passband corner frequencies are more important than the center of the passband, and I find that the waterfall display does not display sufficient definition.
You will also need a cable with an 1/8th inch (3.5mm) stereo plug on each end – long enough to reach from the K2 to the soundcard of your computer. You only need to connect to the tip and the sleeve, but it will work fine if you have a two channel cable, both tip and ring connected. You can wire your own or pick one up at your local audio store. I have even spotted them at the local Wal-Mart store as headphone extension cables but you will need to add a short adapter with a plug on each end.
Information about setting the parameters for Spectrogram is in the documents by Tom Hammond (those instruction are for a previous version of Spectrogram, but are still valid – the placement of the check boxes may have changed). John Grebenkemper posted updated setup information on the Elecraft reflector – if you need the update, it should be available in the archives.
Tom Hammond recently provided me with screenshots of the setup window for 3 versions of Spectrogram. The links below will show you the setup screen in a separate window.
Short form instructions:
Just a note of caution about filters that are wider than about 1.8 times your sidetone pitch. If you are to preserve single signal reception, that is no signal heard on the opposite side of zero beat (‘real zero’ audio frequency), you must skew the passband a bit off center. A spectrum analyzer cannot show this condition since the FFT of zero frequency is mathematically undefined. In addition to the marker, look for the amplitude of the passband to decrease on the left hand slope.
CW filters done? – on to the SSB filters.
Last filter finished? – all done with SSB filters
If you have the RTTY mode set on, you will have another set of filters to align. For normal use with PSK31 or other computer soundcard digital modes, you may want to set these the same as you did for the SSB filters. If you are using a real TNC, you may want to narrow up the filters and position the passband so the demodulated audio will pass the tone frequencies that you use. The narrower filters will increase your signal to noise ratio and allow you to copy weaker signals than possible with a wider passband. For real RTTY with a TNC, I might set the passband to about 600 Hz then adjust the BFO for a low frequency corner at about 1650 Hz and the high corner at 2300 Hz so the passband comfortably contains the tones of interest. Be aware that you may run out of BFO range with these high frequency filter centers. You can shift (or extend) the range a bit by changing the BFO capacitors, but then you will have to re-do all the filter alignments afterward.