MSP430F1611: XT2 Drive Strength

Hi Mark,
actually the Drive Level of the crystal is not the parameter to be looked at. This is just defining the limit the crystal is able to cope with. This results from the nature of a crystal, being an electro-mechanical component. Drive levels above the specified value could destroy the crystal, by too large excitation stress.

Best regards
Peter

The parameter to be looked at in terms of startup and oscillation margin is the ESR. This value is basically reflecting the losses of the crystal, which need to compensated by the oscillator, which is used to drive the crystal, means in our case the XT2.
The best document to look at for these topics is the application report www.ti.com/.../slaa225a.pdf
Although mainly addressing 32kHz crystals and oscillators, majority of the aspects and tests described in this application report are applicable also to high frequency crystals.
When looking at the datasheet screen shots of the two crystals you have provided, the ABM3B-8.000MHZ-B2-T has a 50Ohms higher maximum ESR compared to the ABMM2-7.3728MHz-E2-T. Which means it will probably need a 25% higher drive level from the used oscillator, to compensate the higher losses. This is also indicated by the size of the crystal. While not perfectly linear and comparable across different crystal designs, from tendency point of you one can assume, the smaller the crystal, the higher the losses/ESR, the higher requirements on the used oscillator/driver.
Looking at the layout you have provided, the XT2 layout is not optimum. There are two main aspects, one is the close by routing of XIN/XOUT, which increases the shunt capacitance between both. This is lowering the performance. The other point is the placement of the load capacitors, which should be placed close to one another, keeping the oscillation signal loop, consisting if the crystal and load capacitors, as small as possible, and the GND electrodes of the load capacitors connected directly as short as possible to one another and to the MSP430 device GND connection.

To assess the margin of the design with given crystal, the customer would need to perform the safety margin test as described in the application report.
Another aspect, the XT2 high frequency crystal oscillator is a non regulated oscillator. Its drive level is increasing with the device supply voltage level. Thus if not at 3.6V already, the customer could try to increase the supply voltage.
Also lowering the load capacitance tends to ease the required drive level conditions with a given crystal (as you noticed), but there is of course a limit, as the load capacitance value is required for accurate frequency, as the nominal frequency is only achieved with specified load capacitance. Too low capacitance makes the crystal oscillate too fast, vice versa too high capacitance slows it down. This is in small ppm range, so potentially not too critical for given application. The more critical aspect are the resulting phase conditions, where at certain minimum load capacitance for given crystal, the setup might become unstable.

My recommendation would be testing the SFR (see above). If the test results in insufficient margin values, HW change (layout optimization), change of load capacitors, but most probably use of a crystal with lower ESR will be required.

Best regards
Peter

Hi Mark,
did my reply answer your/customer's questions? Anything else we could do for you in this case?
Maybe something to add to put all options on the table.
The MSP430F2xx family, which contains compatible devices to the MSP430F1611 are equipped with a more modern high frequency crystal oscillator, with adjustable drive levels, with definitely better performance compared to F1xx high frequency oscillator.

Best regards
Peter