Reference 8MHz Crystal Oscillator circuit

Thread Starter

Anonymous User 30

Joined Sep 4, 2017

I'm building an antenna tuning unit for a 1/4 wavelength vertical antenna for the 40m band.

The ATU will require me to use a frequency synthesizer connected to a VCO to generate frequencies between 7.0 and 7.3 MHz. After a frequency sweep is done a directional coupler will determine the lowest VSWR through a ESP8266EX ADC. This will read out VSWR and the resonant frequency of the tuned circuit. I have this design completed, Arduino code written and have ordered parts off Digikey and am now building the first part of the circuit which is a crystal oscillator used as a reference oscillator for the frequency synthesizer.

The 8MHz oscillator is to be fed into a CMOS MB87014A Fujitsu synthesizer. The required output of the oscillator is to be High + 3.5V and Low - 3.5V.

In LTspice I managed to get the simulation oscillating please see below but when constructed on copper board I could not get an oscillation. This is my first time building a crystal oscillator. Here are the multimeter values after probing:


Node 1 (Node after RFC to V+) 1.67v
Q Collector 1.61v
Q Base 660mV
Node 2 (Voltage divider node to crystal) 550mV


THrough collector resistors 3mA
THrough R2 1mA
THrough R9 &R6 1.65mA
THrough L5 & R8 292mA

(Assumed current through crystal 1.65mA - 292mA = 1.43mA
Assumed power through crystal 786 uW - Under 1mW drive level)

Continuity check revealed no connections between grids at the time of board cutting. All capacitors were checked with Ohm meter and respond good.

It appears the RF gain is staying at roughly 1 when probed with oscilloscope.

My assumption is that the toroid core for the impedance matching transformer is unsuitable for a 8MHz frequency. I believe I need a core with a low permeability. The material for this toroid was L8.

Any assistance would be appreciated.


Schematic LTspice.JPGDSC_0005.jpg


Joined Oct 7, 2019
I am using this Nano VNA to test cables and filters and antennas. Mine is -H which goes to 1.5Ghz. It gets a little strange above 900mhz. There are many different versions some that go to 3ghz. Some have three ports but I can not afford that. Price is about $100 usd. It will work very well at 8mhz. It talks to a PC over USB or will work stand alone.

Thread Starter

Anonymous User 30

Joined Sep 4, 2017
Thanks ronsimpson I've heard good things about that unit, i'll definitely order one.

I've made a few discoveries since probing around:

1. The inductances of the transformer windings were way off, thats been rectified and the transformer now has been independently tested and provides the correct voltage ratio of 9 to 1.

2. The crystals I have ordered do not have a 1mW max drive level as two of them have now failed. Probing once current was moving through the 2nd crystal showed that oscillations did not occur perhaps due to the over unity gain of the amplifier. 2 possible solutions change Rc / Re or lower the capacitance of the dc blocking caps c6 and c2 to provide more reactance.

Before I proceed with carrying out 2, I am slightly perplexed with the current readings I am receiving through the circuit (maybe due to my hobbyist background). I anticipated the supply of 3.3v to be a constant voltage supply although my supply switches to constant current and drops the voltage to about 1.7v. I anticipated a max current to move through R6 & R9 to be 3.3 / (390 + 6200) = 500uW, couple this with the voltage division and the drive level of the crystal should be not exceeded. Instead I read 1.65mA through those resistors with 292uW read through the other part of the voltage divider to ground, L5 and R8. This would be leaving 1.3mA through the crystal. I understand this is occuring because the transistor is sinking current through the base which has to go through the transformer, but I though current could be regulated during constant current mode. Any help would be appreciated.


Joined Nov 3, 2020
I noticed that the oscillation on the red trace (Vout.JPG) is around 13MHz, some distance from the 8MHz in post #1 para 2. Taking your crystal simulation (L1= 56uH, C1=7pF, C3=10nF, R5 =60ohm) from the schematic, I tested it standalone in a simulator which gave resonance at near 200MHz. Increasing L1 to 40mH moved the resonance close to 8MHz.

While the crystal is under the microscope, so to speak on a simulator, you could measure the power in R5 at resonance using the anticipated Vout drive amplitude of 3.5V, and including the surrounding passives C2, C7, L2.

I can't help with your real circuit as I don't have the right components to hand. However, I've recently built a 6MHz oscillator* and simulated it with good agreement between the two. It took some care, but it can be done!

( * Fixed frequency, LC, gain-stabilised Collpitts).


Joined Nov 3, 2020
Here is some more information about an 8MHz crystal oscillator. I found a couple of interesting simulation circuits and stitched them together:

These sample simulation traces show: oscillator output V(e), tank output V(out), and power in crystal I(R1)*(V(xtal)-V(r1b)).

Some brief notes
- The Kickstarter is set at -5mA which enables viewing of representative waveforms after about 20uS of simulated running. If the kickstarter current is low (eg -5uA) then oscillator start up time is around 25 to 30mS of simulated running.
- Both the oscillator and the tank incorporate amplitude stabilisation which helps to keep a good signal shape.