Can I run fundamental mode crystals in overtone?

[technix]

For example, run a 16MHz crystal in 3rd overtone for 48MHz, a 20MHz crystal in 5th overtone for 100MHz, etc?

[iMo]

You may get an overtone off an fundamental Xtal, it requires your oscillator has got an LC tank tuned to the overtone freq.

PS: mind almost all Xtals >25MHz are 3rd overtones..

[Wolfgang]

Yes, with caution.

Two gotchas:

- The overtone frequencies are NOT exact multiples of the fundamental frequency
- The overtone resonance are much narrower than the fundamental so you have less tuning range.

[Kleinstein]

How well the overtones work depends on the shape of the electrodes - so the coupling efficiency may be as good as with specially made overtone crystals and the tuning range can be even smaller than with specially made overtone crystals.

The extra LC or similar filter to select the overtone is needed, just like with regular overtone crystals.

[DaJMasta]

You could also make an oscillator with the crystal in its standard operating mode with an output using fairly fast logic (hard edge square wave), then use a  bandpass filter to find the odd harmonic you want in the square wave's overtone series and amplify as needed.

It seems like more trouble than it's worth, though, why not just get a 48MHz or 100MHz crystal/oscillator and just divide to get your slower clock?

[BrianHG]

Yes, but your LC circuit will need to be tuned to lock into the third overtone.  It can be a sensitive setup which may loose lock under extreme situations like wide power supply voltage swings, or, extreme temperature change with cheap capacitors that drift in value with temperature.

Using an 74HC/74AC XOR gate, you can also do a 2x frequency doubling trick by feeding the output of any oscillator into input A, tying an inductor form input A to input B and add a varicap at the B input to GND to tune the B input to a  90 degree phase delay.  The output of the XOR gate will be 2x the input frequency.  With a 4x 2 input XOR gate chip, you can use 1 channel for the crystal oscillator, and another for the frequency doubling with 2 spare gates for buffers/inverters.

[Wolfgang]

Some more thoughts about overtone oscillators:

- 3rd overtone is easy, 5th is OK, 7th is very tricky
- tuning ranges become VERY narrow

It could make sense to make a fundamental oscillator (simple to build, good tuning range) and then multiply the output by the factor you need.
- if the factor is even, you could used a balanced pair of transistors or a diode doubler
- if it is odd, a good circuit is the "odd-order diode frequency multiplier" by Charles Wenzel. This one has very low phase noise.

Example:

https://electronicprojectsforfun.wordpress.com/an-noninteger-frequency-divider-with-sinewave-output/

[T3sl4co1l]

Overtones also have different tempcos.

You must use a crystal that is cut for overtone use.

Tim

[G0HZU]

For a one off circuit for home or lashup purposes you can often use a 16MHz fundamental crystal but it probably won't oscillate bang on 48MHz.

The extra LC or similar filter to select the overtone is needed, just like with regular overtone crystals.

It's a good idea to use LC parts in a tuned circuit to select the overtone but you don't have to. For a quick and dirty lashup circuit (minimal parts count?) you can still use a classic BJT circuit with suitable bias resistors plus the 16MHz crystal and two capacitors. You can still set it up to oscillate at 48MHz like this. However, I wouldn't recommend it for volume production