From: Winfield Hill
Subject: Re: Querry, Kick Starting Crystal Oscilator.
Date: 12 Nov 2002 10:48:57 -0800
Organization: Rowland Institute
References: <3DCA905A.59FD928B@mmm.com.DELETETHIS> <3DCC9A71.7DF3C3A9@mfi.net>
X-Newsreader: Direct Read News 4.10
Tim Shoppa firstname.lastname@example.org wrote...
> Winfield Hill email@example.com wrote...
>> Like most crystals, those have a Q of 60,000 to 90,000 which
>> means they'll take to 2 to 3 seconds to reach 50% of the full
>> oscillation energy. That's the way it goes; the energy cannot
>> simply be kick-started into the crystal.
> I agree with your calculations, Win, but in my experience even
> the simple and good CMOS 32.768 kHz oscillators are up and at
> full amplititude within a fifth of a second, and are stable and
> producing a nice 32.768 kHz square wave within 50 ms.
> The data point I just took at my desk: a Fairchild CD4060BC
> configured as a 32.768 kHz oscillator as in the Fairchild data
> sheet (page 3), and I measured the amplitude at pin 10 as I
> applied Vcc. The crystal is an Epson C-002RX32.768K-A (digikey
> part number SE3202-ND, I think this is as generic as a 32.768 kHz
> crystal gets, 35 cents each.) The amplitude at pin 10 is at 2V
> P-P within 50 ms and is at full amplitude (about 4.5V P-P) in
> under 200ms.
> I had previously tried using TI 74HC4060's in the oscillator
> and they didn't work stably *at all* (see my previous 32.768 kHz
> oscillator threads, everyone says that the inverter is buffered
> and cannot work well as an oscillator in the 74HC.)
Those are nice comments and data for us, Tim. I'm not surprised,
because the you observed on pin 10 is an amplified version of the
small voltage developing on the crystal's output, pin 11. The
gain of this amplifier stage in the 4060 is very high at 32kHz
(see figure 3.62 in AoE), perhaps over 100x, so the 2V level
you measured may have corresponded to under 20mV on the crystal
at that time. We certainly expect to get logic outputs long
before the crystal reaches full operation.
However, the issue is often not getting full logic swings, but
rather getting fast rise and fall time on these logic lines.
As the crystal voltage increases the logic-level rise and fall
times will decrease. Tim, I wonder what you would see for the
risetime on pin 10 over time at startup. The counter has more
inverter stages to help, you can inspect the next one at pin 9.