Discussion specific to the DIP and TQFP packaged ZX devices like the ZX-40, ZX-44, ZX-32 and ZX-328 series. The differences between these devices is primarily the packaging and pinout so most issues will apply to all devices.
Don_Kirby wrote:[W]hat is the stated accuracy of the crystal supplied with the ZX-40 Interface Board?
The part, Digi-Key X175-ND, has a tolerance of +/-30ppm according to the datasheet. Over 19 hours, that would be a variation of about 2 seconds.
The crystal load capacitors supplied with the ZX Interface kit are 27pF. You might try some 22pF caps to see if that has an effect - it might make it run slightly faster.
If I'm going to be this picky about timing, I suppose I should do it right. To avoid the scope probes causing excess loading when measuring the clock frequency (and therefore changing the measured frequency), I need a buffer between the crystal and the probe. Any suggestions on a simple buffer circuit?
If you are in posession of a digital readout shortwave receiver, you might be able to measure the clock frequency with great precision minus significant coupling to the circuit. As a ham radio operator I've often put a small loop of wire across a coax and connected the other end to the receiver. My receiver has a precision of 10 Hz and can be checked against WWV at 5 and 10 MHz for accuracy. It's pretty easy to measure frequencies from 40 kHz to over 100 MHz with it as a sniffer. Moving the loop of wire within a few inches of the oscillating circuit usually provides plenty of signal. Plus, you can hear if you are pulling the oscillator by coupling too closely if the pitch changes in the receiver output. Receivers are sensitive to microvolt signals as opposed to frequency counters without special preamps that require signals thousands of times higher in voltage.
I thought I'd try measuring the frequency on a ZX-24 and check on the stability of the oscillator vs outside capacitance to the crystal. I measure 14.74518 MHz with a likely error of 10-20 Hz since I didn't do more than a casual calibration. The loop was inches away from the crystal, so significant coupling is negligible. Placing my finger on the crystal case increased the strength of the signal greatly and produced no noticeable changing of the oscillator frequency.
Touching the crystal pin closest to ZX-24 pin 1 with a short wire connected to my fingers results in 14.74491 MHz and the other crystal wire contact gives 14.74478 MHz. So, you can pull the crystal a few kilohertz with severe outside coupling. What's notable is that the oscillator remained stable and continued to run with such a large perturbation. Of course, I made sure I wasn't about to blast the circuit with a static discharge.
I'm sure temperature will slide it around a bit also.