CXO: Low distortion oscillator
2011/8/2
both <<EMRFD>> and <<ARRL hand book 2006>> emphasis one crystal oscillator which get signal direct from the crystal. it's worth to try. i setup a following circuit with NO L1, R1=10k, C1=220pF, C2=470pF. Q1=2sc9018, cystal 7.47Mhz, NO buffer Q2.
Fig1
Fig2 : my bug style build:
then we get a prety good output from crystal, at first glance. after very carefully exam, by align the peak to Y axis, move wave up and down, check the distance between 2 X axis cross point to the zero point, there are about 1ns different between the raise edge and falling edge.
Fig3 crystal output: 20ns/Div, .1V/Div
Such asymmetric due to the even order harmonics. there's another very good experiment done by Jane, http://www.qrp.pops.net/crystal-osc-2010.asp Jane point out if the harmonics more than -50dBc, there are visible asymmetric on Oscope.
another wave, the emitter output is very distortion(below):
Fig4 20ns/Div, .1V/Div
fist of all, check the biasing. Fig1's basing method is not very predicable, with Q1, the 2SC9018. base get only about 1V. so i re-biasing it ( Fig 5. )
Fig5 another populous circuit, good bias
Just change bias didn't change anything. another noticeable configuration in Fig5 is the feed back caps, C1=2.2nF, C1=390pF. use such big C1/C2 value dramatically reduce the positive feedback, get close to unity gain, easy to get more clean signal. I finally use C1=2200pF, C2=56pF to get a pretty good emitter output , Fig 6.
Fig.6 emitter out at C1=2200pF, C2=56pF
What about the Crystal output now? , still not see any observable change, 1nS distortion still there. Call for a SA. low pass filter is mandatory to get more clean clear sine wave.
Fig.7 final Crystal output at C1=2200pF, C2=56pF