Steering a crystal with a better clock source?

[@rt]

Hi
I have a crystal clocked microcontroller with capacitor arrangement much like any other.
I’ve heard years ago that if I have a better clock source (in this case TCXO with TTL output),
it’s possible to attenuate that, and connect it directly to the clock input pin with the crystal still in place,
and that the new signal will correct the crystal oscillator to become a more accurate clock.

If this is true, does it matter that the TCXO is TTL?

With some series resistance, will it drag down the original crystal much when the TCXO circuit is switch off?

I ask the second question because I assume the TCXO is a drain on batteries while it’s operating heating up a resistor oven.
Cheers.

[@rt]

Well Dah! I just don’t see how this happens without dragging down the original crystal (when the TCXO is unpowered).

[Gregory]

Yep, It can work

It's called Injection Locking.

Actually, it is more common to sync a much lower phase noise oscillator to a better one, like a crystal.

You will need to test some resistor values and maybe put a parallel resistor with the crystal.
A parallel resistance will decrease the high crystal Q making it easily steerable.

In a lot of cases a very short pulse is used to lock the oscillator. A short pulse will inject a smaller amount of energy.
Also, a short pulse has more "temporal" information and I think will increase the chance of lock.

A short pulse of energy can be injecting coupling the oscillators with a small capacitor.

[CJay]

AS gregory says, injection locking is a thing.

However, you seem to be confusing TCXO with OCXE, a TCXO monitors temperature and (typically) changes circuit reactance for it to keep the frequency stable.

An OCXO keeps the crystal at a constant temperature  by using heating elements.

So, you may find the TCXO is well within your power budget.