EEVblog Electronics Community Forum

Electronics => Projects, Designs, and Technical Stuff => Topic started by: ChristofferB on March 11, 2018, 11:45:53 pm

Title: Odd (unstable) behavior in xtal CMOS oscillator
Post by: ChristofferB on March 11, 2018, 11:45:53 pm
Hi!

I'm building a little 455KHz square wave oscillator to use as a BFO, using this fairly standard layout: (http://electronbunker.ca/eb/SignalGen455_files/IF%20Aligner%20Basic.jpg)
with another inverter after as buffer, and whatever I do, I can't get it to put out the XTAL frequency!

 I didn't have the 100p trim cap so i had 68p over each leg and then a 40p trim cap paralleled across one of them, and it puts out a VERY high frequency (compared to 455KHz) - 5-15 MHz! on a few rare occasions i've managed to get the desired freq out by tweaking the trim cap, but with a terrible waveform, and so unstable that if i cycle power, it returns to the 5-15 MHz output..

the 455 xtal is really one of those plastic packed ceramic resonators used in radios.

Any clue as to what's going on? I kinda thought xtal oscillators should either resonate at their freq. or not at all.

thanks!
Title: Re: Odd (unstable) behavior in xtal CMOS oscillator
Post by: T3sl4co1l on March 11, 2018, 11:57:50 pm
Try putting a parallel resonant tank on the inverter input. Also, it's an unbuffered type (CD4069UB or 74HCU04), right?

Tim
Title: Re: Odd (unstable) behavior in xtal CMOS oscillator
Post by: BrianHG on March 11, 2018, 11:57:59 pm
Hi!

I'm building a little 455KHz square wave oscillator to use as a BFO, using this fairly standard layout: (http://electronbunker.ca/eb/SignalGen455_files/IF%20Aligner%20Basic.jpg)
with another inverter after as buffer, and whatever I do, I can't get it to put out the XTAL frequency!

 I didn't have the 100p trim cap so i had 68p over each leg and then a 40p trim cap paralleled across one of them, and it puts out a VERY high frequency (compared to 455KHz) - 5-15 MHz! on a few rare occasions i've managed to get the desired freq out by tweaking the trim cap, but with a terrible waveform, and so unstable that if i cycle power, it returns to the 5-15 MHz output..

the 455 xtal is really one of those plastic packed ceramic resonators used in radios.

Any clue as to what's going on? I kinda thought xtal oscillators should either resonate at their freq. or not at all.

thanks!

With a 450Khz ceramic resonator, you should be thinking capacitive loads at 220pf to 470pf at the input side of the inverter, if not ever higher.  You should even add a 100 to 470 ohm series resistor at the output of the inverter before it reaches the resonator & 1 meg resistor.  Note that to drive a digital clock, you still use the direct square wave output from the inverter.

Don't forget a 0.1uf decoupling cap for the inverter IC!!!

Also make sure your inverter is a Schmidt for such a low frequency.  Regular inverters may have their input float around the switch threshold and oscillate like mad ignoring the resonator.
Title: Re: Odd (unstable) behavior in xtal CMOS oscillator
Post by: ChristofferB on March 12, 2018, 12:32:19 am
Thanks! I'll try out with a schmitt trigger, more capacitance and a small-ish series resistance tomorrow!
Title: Re: Odd (unstable) behavior in xtal CMOS oscillator
Post by: w2aew on March 12, 2018, 02:08:43 am
And be SURE to add a power supply decoupling cap at each IC as BrianHG suggested - very important.
Title: Re: Odd (unstable) behavior in xtal CMOS oscillator
Post by: BrianHG on March 12, 2018, 02:28:56 am
And be SURE to add a power supply decoupling cap at each IC as BrianHG suggested - very important.
I also forgot to say that you need to GND the inputs of unused gates.  Otherwise, they also may oscillate randomly.  Especially in a breadboard environment.

Also, the GND side of the caps should go straight to the GND pin row on a bread-board, not the GND rail.  This also removes additional signal bounce introduced by jumper that 1 jumper-wire to the GND rail.
Title: Re: Odd (unstable) behavior in xtal CMOS oscillator
Post by: T3sl4co1l on March 12, 2018, 03:35:46 am
If it's CD4000 family, it can be powered with wet string. Bypass is inconsequential.  74HC is fine if it's within a few inches.  Don't use AC or LVC.

Tim