Ceramic Resonator Questions

[eev_carl]

Hi,

I'm building a circuit with a ZTT 20.00 MX and had some questions regarding the frequencies I was getting.  The datasheet says that MX's are 13.01-50.00MHz and that the "20.00" indicates a 20.00 MHz frequency.  However, in the circuits I built, the largest frequency I got from the attached schematic was 4.4MHz.  The 4.4MHz was the value of the circuit where the dampening resistor was left off.  I was expecting a higher value.  I'm using 5V to power the CD40106BE Hex Schmitt Trigger IC.

I'm also wondering if the circuit I built with the 50k dampening resistor (shown in the schematic) is correct for the output of my scope (270kHz shown in scope screenshot).

Thanks,
Carl

[RobK_NL]

Standard CMOS logic as very slow, certainly at 5V. The delay from input to output for this part is 140ns typical, which is already almost 3 times longer than the period of a 20MHz signal.

If you replace the 40106 with a 74HC(T)14, you'd get much better results.

[eev_carl]

Thanks! I'm looking for Propagation Delay Time on the datasheets, right?  It's 38ns for the SN74HCT14NE4 and 280 ns for the CD40106BE.

Is the CD40106BE for the circuit in the scheamtic correct?  I have a dampening resistor of 50k which brings the frequency down to 260kHZ (period = 3.8us > 280ns).

[RobK_NL]

Hi,

Yes, Propagation Delay Time is indeed the term you're looking for. That 38ns you mention is probably the maximum over temperature. At 25°C they are normally about twice as fast.

That dampening resistor is way too high in value for the frequency concerned. Somewhere between 250 and 500 Ohms would be much more appropriate. I even doubt if it is necessary at all for use with a resonator. For a quartz crystal yes.

[Benta]

Two comments:

1: DON'T use Schmitt triggers for crystal or resonator oscillators. It's the certain path to failure. The resonator needs a linear inverter (think sine wave).

2: 40106 at 20 MHz?    Even a 74HC04 is marginal at best. Go for a 74AC04.

[eev_carl]

A Schmitt trigger is for slow, noisy edges and not a fast application like this?

[glarsson]

A Schmitt trigger is for slow, noisy edges and not a fast application like this?

Yes, but also that the schmitt will give you a square wave. You don't use them when you need to get a quartz crystal or ceramic resonator to oscillate. Then you want a sine wave. Compare to making a pendulum swing nicely. You push it nicely at the frequency it wants to resonate at, not smash it back and fort with a tennis racket.

[Benta]

Compare to making a pendulum swing nicely. You push it nicely at the frequency it wants to resonate at, not smash it back and fort with a tennis racket.

That's one of the greatest analogies I've read. 

[T3sl4co1l]

Yeah, you want a 74HCU04 in fact.  HC is buffered, meaning it has more delay and very high gain.  That's a good recipe for making an oscillator, but not at a desired frequency...

HCU is unbuffered, available in a few special types like '04, literally just a pair of transistors per section.  The linear range is wide and the gain is modest, making it very good for oscillators.

Tim