32.768 KHz crystal failure

On Mon, 29 Jun 2009 16:26:37 +0100, "N_Cook" put finger to keyboard and composed:

I've never observed that type of failure. I've seen crystals that resonate at an overtone rather than their rated frequency. Others were shattered by physical impact (you can hear the shards rattling inside), and others just wouldn't start oscillating without some coaxing, either by heating or freezing. I've also shattered a crystal when I cut up a circuit board with a band saw.

The strangest failure was in an IBM PC whose graphics chip was outputting H & V sync frequencies that were much higher than normal, but were not an integer multiple of the correct frequency. I traced the fault to a crystal oscillator. It was suggested to me that maybe a quartz dag that was present during manufacturing had later fallen off. This would reduce the mass of the crystal, resulting in an increase in its resonant frequency.

- Franc Zabkar

--
Please remove one \'i\' from my address when replying by email.

pins

have

Certainly not the simple slab with simple silvering on each side. As the open ends of the tuning fork look ground, on a 30x ,then perhaps literally a tuning fork , actively tuned to frequency, and the odd trace patterning is for modulation of capacitance. I don't suppose a 13 ohm build up of Ag over

100 micron gap would be observable, as stable at 13R via soldering/movement doubtful it was some mechanical breakage/movement.

-- Diverse Devices, Southampton, England electronic hints and repair briefs , schematics/manuals list on

have

My 30x microscope+camera can only capture about 1mm diameter

overal quartz sliver is 6.5x1.5x0.3mm White is the silvering with some illumination flare, some connected to one pin and some connected to the other pin, similar but not the same on the other side. G is the area of quartz ground down end and face, a bit , compared to the lower finger of the "tuning fork". Area near the notch , dimension between + and + is 0.75mm , half the width. So minimum gap between silverings about 60 micron. Width of notch about 0.25mm There is also silvering along some of the 0.3mm edges, so very complicated silvering geometries

-- Diverse Devices, Southampton, England electronic hints and repair briefs , schematics/manuals list on

It is a tuning fork reference. Bulova type, transistor switched to oscillate at 360 hz.

Non-sequitur. The Bulova Accutron, dating back more than 45 years, used a rather large tuning fork. I still have my father's.

I have one made in 1965, solid gold.

So it's not a tuning fork but rather a crystal.

That cut is referred to as a 'Tuning fork crystal', because of the long, thin shape.

A 32768 hz crystal is used by some MPU ICs with a built in PLL multiplier. The MC68340, of the MC68020 family is one example.

Improper loading can cause it to run at the wrong frequency, and a

10:1 scope probe will detune it. I had to use a special Tektronics FET input scope probe to troubleshoot a design problem on a custom embedded controller.

These crystals are fragile, and need to be mounted so that they can't vibrate. Most, if not all datasheets recommend laying them flat to the PC board and soldering the top end of the cap to a pad on the PC board, but a lot of designers fail to do this.

--
You can\'t have a sense of humor, if you have no sense!

Very familiar with xtal controlled phase locked loop circuits for decades, just never heard that description.