Trouble with unwanted B-mode oscillation in OCXO using SC-cut crystal

I'm building an emitter coupled oscillator (schematic: OCXO_56448_II.jpg) using an SC-Cut 3rd overtone crystal at 5.6448 MHz that has the following specs:
ESR:100 ohm
C0:3.72pF
Cl:0.122fF
Q:2.3M

While a simple capacitor (C3 in the schematic) is able to suppress the unwanted B-mode oscillation for all the crystals I have tested from 11 up to 49 MHz, with the above crystal the oscillator starts anyway at around 6.13 MHz instead of 5.6448 MHz.
I suspect that the resistance of the crystal at B-mode is lower than the ESR at C-mode.

I've tried several ways to suppress the B-mode oscillation without success (attached). All the attached schematics seems to work correctly in simulation but don't work in the real world (X2 emulates the B-mode of the crystal).

Any idea about a more efficient way to suppress the unwanted oscillation at B-mode?

Thank you in advance

The best way to compensate b-mode in your case is first to change your oscillator type, from Vackar what it is now, to Colpitts or Pierce.
Then compensate one of the feedback capacitors to be inductive at the frequency of the unwanted mode.

Rather than a Vackar it's a simplified Driscoll oscillator.

Anyway, for some reason, changing the oscillator circuit is not an option.

Any idea about a trap to suppress the unwanted mode?

6.1 MHz LC series circuit across the crystal?