Crystal Load Capacitance

Lets say I'm building a micro controller circuit with an oscillator consisting of a crystal and 2 capacitors (one from each crystal lead to GND). Should the value of the capacitors match the load capacitance of the crystal, that is if the crystal is rated at 10MHz and has a load capacitance of 18 pF should the capacitors have a value of 18 pF? If not, what determines the value of the capacitor?

Deleted content.

Make the capacitors about 30pF.

Two 30pF in series gives 15pF, add a few pF for stray capacitance and voila you have 18pF load on the crystal.

Unless you are looking for an EXACT frequency, the capacitance is not critical.
If you are looking for exact frequency, make one of the capacitors variable so that you can adjust the frequency.

JimB

Some xtals need a shunt resistor and some need a series, check the xtal datasheet.
You can if you really want make a fine freq tune under software control, place a varicap diode in one leg of te xtal and control it via a d to a in software.

JimB said:

Unless you are looking for an EXACT frequency, the capacitance is not critical.
If you are looking for exact frequency, make one of the capacitors variable so that you can adjust the frequency.

Click to expand...

Jim, should that be a particular type of trimmer capacitor? I seem to remember something about certain types having a better temperature coefficient than other types, which might affect frequency.

Jim, should that be a particular type of trimmer capacitor?

Click to expand...

A very good question!
I must admit that I had never really give it too much thought.
I would never out of choice use a mica compression trimmer.
Most circuits I have seen have used a ceramic trimmer, which seems to do for most applications.
I suppose the best type would be an air spaced trimmer, either one with rotary vanes, or one of the older phillips type beehive trimmers.

JimB

I just ran into this question myself. Apparently, the capacitance is the "recommended" value" and usually the xtal and resistor are recommended/tested by the part or logic family that uses it. The values are selected to assure a start over the supply voltage range.

I just ran into this question myself. Apparently, the capacitance is the "recommended" value" and usually the xtal. Assure it will start at the lowest voltage. Sometimes the caps are not symmetrical.

There is yet another circuit that's used where another smaller valued resistor is placed in the circuit. This resistor and the cap makes a filter so that the crystal doesn't oscillate at higher harmonics. There is an apparent "polarity" with respect to X1 and X2 and the respective circuit attachment.

You can also, apparently find this "sweet spot" with a network analyzer.

So getting back to my original question, the value of the capacitors isn't critical unless there's a need for an exact frequency? I am referring to connecting a crystal across the microcontroller's osc1 and osc2 terminals and the caps from each terminal to GND.

True. The criticalness is that you need to assure a reliable start of the oscillator and the oscillator may operate at overtones. That's the risk.

When I breadboard projects with a pic I dont bother with any caps just the xtal, it works perfectly without.