Return to the Main Procedure page


Determining that the reference oscillator is set to 4000.000 kHz can be as easy or as difficult as the equipment available to you will allow.  Your patience, and your available time are also factors for assessing the difficulty level.  In general, the faster methods require more accurate equipment.  Review the equipment you have available and make an honest appraisal of its accuracy, then select the method most applicable to your situation. 


If you want additional information about precision, accuracy, and calibration of equipment, click here for a brief tutorial


WARNING: Do not connect a probe to the reference oscillator in an attempt to measure its frequency directly.  Doing so will 'pull' the reference oscillator and result in a frequency change after removing the probe.


When building your K2, you did connect the internal counter probe to C22.  This was only a test to confirm the counter was working.  As noted in the K2 manual, this is not an indication of how well the 4MHz oscillator is calibrated.


If you must couple to the signal from the reference oscillator, do so with an insulated inductive loop placed in the vicinity of the microprocessor (right above the crystal can), and be certain that the frequency does not change as you move the loop further and further away.  If you notice a change in frequency, you can be certain that the process had pulled the oscillator.


Adjusting the reference oscillator frequency is easy, you just have to change C22 on the control board.  Getting C22 to the correct position is the hard part!!!


Various methods of setting the reference oscillator have been published and several owners have mentioned more in their postings to the Elecraft reflector.  Any of these methods will work.  Some require no equipment other than the K2 itself, while others require the availability of well-calibrated equipment.  I have categorized the various methods into two major groups and then I will break down these groups into several techniques that have been useful to some of the K2 users.  You can select one, or several, that are applicable to your situation and equipment.


·         Count something with a frequency counter (either external or the K2 internal counter)

·         Zero Beat a signal using a receiver


All of the methods/techniques I will refer to are known to work.  Those I have not attempted myself have received reliable testimony in the reports posted to the Elecraft reflector.  If you have the required equipment you may want to try several methods yourself to see which works best (easiest, fastest, whatever defines "best") for you.  Some require only a functional K2 and the ability to receive a specific signal, so even if you have no equipment available, all is not lost.


In general, the techniques that ‘Count Something’ can be more accurate and are the fastest and easiest, but they do require that you have access to accurate, recently calibrated equipment.   The accuracy of the result depends on the accuracy of the external equipment.  Equipment accurate to 0.01ppm or better can be expected to produce more accurate results that the ‘zero beat’ methods.  Even if the accuracy is the same for both categories, I usually find the methods using external equipment (counter or frequency standard) can be done more quickly.


I have used both methods in tandem when the available equipment accuracy is not trustworthy.  Initially I proceed as though I were using very accurate equipment, then zero beat with WWV to see how far off I actually am, and repeat using the same equipment but offset the readings on the equipment the second time around (do not readjust your equipment calibration mid-stream unless you plan to start over from the beginning).  I have done this several times, and can usually achieve good results in no more than 2 passes.  You may have to run an additional pass or two until you develop a feel for your particular equipment and the technique chosen.  With really good equipment and practiced techniques, you can be assured of good results the first pass of a K2 calibration job.


If this is your first time using a particular technique, as you proceed with your observations, it is a good practice to take notes, including equipment setup information, and relevant dial readings.  That way, if you do not achieve your target accuracy the first time around, you can compare your notes to the instructions for the technique you are using and see if there was any discrepancy.  Things usually get better and easier with a bit of practice and patience.


There may be other postings in the past that I have not included here, but those I have referenced cover the techniques that I am aware of.  I will update with additional information if it is made available to me.


Return to the Main Procedure page


METHODS for ‘Count Something’ – there are 2 methods

·         Use an external frequency counter and compare with the K2 internal counter

o        This method is outlined in the K2 manual in a textual description; I have repeated it here as an outline of the steps and added some small enhancements.

o        The theory behind this method is that the internal frequency counter in the K2 will only provides the correct reading if the reference oscillator is correct – when the K2 shows the right numbers, the oscillator will be correct.

o        A well calibrated counter is required.

·         The results depend on your counter accuracy

o        Also see the Elecraft reflector posting by John Grebenkemper KI6WX dated April 27, 2003.

·         Click here to review John’s posting.

·         Note the cautions about letting the K2 warm up adequately and calibrating with the top cover in place.

·         Rich Lentz KE0X has also provided a list of "Things to remember" when setting the K2 calibration.


1.      Select the 10 meter band.

2.      Set the K2 to CAL FCTR (enter the menu).

3.      Connect the K2 internal probe to TP1.

4.      Connect the external frequency counter to TP1.

·         You will have both counters connected to TP1.

5.      Adjust C22 until the K2 displays the same frequency as the external counter.

·         The K2 display does not show the Hz digit.  This means that your possible error can be +/- 5 Hz.  I know of no way around it with this method.

·         Ideally your external counter should display the ones digit so you can average the 10s digit (is it closer to xxxx.x3 or xxxx.x4?).

6.      The reference oscillator is set – go on to step 2 of the main procedure.


·         Use an external Frequency Standard

·         Bob Lewis (AA4PB) contributed this method in his reflector posting of July 28, 2003.

·         Read the text by clicking here.

·         Bob’s technique used a 10MHz ‘standard’ but may also be used with any known accurate frequency standard that you have available.

·         The frequency must not exceed the frequency the K2 internal counter maximum.

·         Unspecified maximum, but we do know it will count the VFO, which goes up to 24 MHz.

·         Your standard should be stable and its frequency accurately known.


1.       Put the K2 in frequency counter mode (CAL FCTR).

2.       Apply the frequency standard output to the K2 counter input.

a.       Don’t forget that the ground wire must be connected as well as the probe tip.

3.       Adjust C22 until the K2 dial reads the frequency of your standard.


·         Observe the cautions about warming up the K2 and with the top cover on – as above.

·         Also see the last paragraph in Michael Masleid’s (AB9GV) posting dated July 28, 2003 for an alternative using a stable oscillator but not necessarily accurate.


METHODS for ‘Zero Beat a signal using a receiver’:


·         Determine dial setting change needed by tuning to a calibrated signal source (WWV or other)

o        Requires no equipment other than the K2 itself

o        This method is shown in the K2 instruction manual

o        This is an iterative method – you may have to repeat it until the dial calibration is satisfactory to you.

o        If you are using WWV, be aware of the schedule and know the frequency of the tones being transmitted.


§         Be careful not to confuse the carrier and the tones when listening.

Additional notes and an alternate way to detect zero-beat was described in a reflector posting by Glen Worstell, KG0T dated March 19, 2003.

1.      Tune the K2 to the calibrated signal at zero-beat

a.       Use LSB (or USB) because these modes do not introduce a receive frequency offset.

2.      Determine the amount and direction of the dial calibration error.

3.      Connect the internal counter to TP1 and adjust C22 to compensate

a.       If the dial reading was too high, adjust C22 for a lower frequency.

                                                                                      i.      If too low, adjust C22 for a higher frequency

b.      Adjust C22 to change the frequency by approximately the amount of the dial error.

4.      Continue with step 2 of the main procedure (Run CAL PLL and CAL FIL).

a.      It may be necessary to perform this process several times – each time obtaining better results.  Repeat until you are satisfied.


·         Beat the reference oscillator 5th harmonic with 20 MHz WWV

o        WARNING: Do not connect directly to the reference oscillator

o        See referenced postings below for details:

§         George, W5YR, dated July 26, 2003

·         (A correction – the WWV tone should be 440 Hz instead of 400 Hz)

·         Uses an external receiver

·         Detects zero beat (with 1000 Hz offset) by monitoring with DigiPan

o        Other analyzer programs also can be used

§         Rich Lentz (KE0X) dated July 26, 2003

·         2 methods described here – one with K2 as receiver, the other with an external receiver

·         Detects zero beat (with 1000 Hz offset) by using MixW

§         Ron D’Eau Claire (AC7AC), dated July 26, 2003

·         Describes coupling to reference without pulling


·         My additional notes:

o        Zero beat detection is key to success with any of these methods.

§         Neither the human ear nor a receiver can hear very low frequencies.  Detection of ‘real zero beat’ means knowing when you are hearing zero frequency, and we just can’t hear that.

·         Offsetting the receiver to hear a tone (1000 Hz is easy to add/subtract) makes the zero beat more easily detectable.

·         While we cannot hear really low tones, we can quite readily hear the BEAT between two tones, even down
 to 1/10ths of a Hz (1 beat in 10 Seconds).  The sound is the same as when spotting a CW signal, and when there are two steady signals of about the same volume, it is even easier than spotting a signal.  Once you know what to listen for, this beat detection is quite easy.

§         An audio frequency meter or audio spectrum analyzer is quite helpful, and may even be a necessity.

·         I am not gifted with ‘perfect pitch’ as some musicians are – I need a tool!!!

·         Matching tones can be done (match with the K2 spot tone for example) but are not as accurate as a displayed audio frequency indication.

·         Spectrogram, MixW, DigiPan, and other soundcard audio ‘analyzers’ are available as shareware or free for the download – any of them will adequately serve this purpose.

o        When receiving WWV, know the format and the tones being transmitted (see link above)

§         Spectrogram, MixW, DigiPan, and others can display the frequency of a tone with either a set marker or by placing the cursor on that tone.




            Stephen Gibbs (GU3MBS) has provided two alternate methods that you may also want to try.  His methods are available at




Return to the Main Procedure page


Continue on to Step 3 - CAL FIL