AF Marker Generator

I have four 2048KHz quartz crystals, probably originally in some old 32 channel PCM telecomms equipment, and was thinking how I might use them, or at least one of them. I also have the same number of CMOS logic binary divider ICs, types HCF4060BEY and CD4060BE. 2048 = 2e11; dividing it by 2e12 ( 4096 ) will produce a 0.5KHz square-wave from the divider chip.
In my circuit, the crystal is used in a 2048KHz oscillator to clock the divider chip and the 0.5KHz square-wave is taken from the /4096 output on pin 1. A resistive potential divider reduces the level of the square-wave to 1.05V on the output phono socket.
As a square-wave is the sum of harmonically related sine waves, in the frequency domain there will the fundamental 0.5KHz signal and its odd numbered harmonics; 3rd, 5th, 7th, 9th etc. For my purposes the most useful signals are the fundamental, 3rd and 5th harmonics; 0.5KHz, 1.5KHz and 2.5KHz respectively, as these lie within the AF band 300Hz-2.7KHz typically used in radio communication, and can be used as accurate frequency calibration markers.
I have connected the marker generator to the sound card in one of my PCs in order to display the audio spectrum. With the cursor arrow placed as accurately as I can manage on any one of the signals, e.g. the 3rd harmonic near the band centre, and reading off the frequency there is a frequency error of +7Hz due to the inaccuracy, compared with the nominal, of the sampling rate of the sound card input.
After that slight digression I had better get back to PIC programming, or maybe I'll make a start on putting together the pre-amplifier, ( see June 16th ).