3
\$\begingroup\$

TL;DR

Is there a means of measuring the crystal drive level reliably using voltage measurements (even with a differential voltage probe)?


Background:

I am attempting to measure the crystal drive strength in order to ensure that it is not exceeding its rated power of 100uW.

Every application note I have read on the subject says to measure with a current probe and calculate from there.

Measuring crystal drive strength with a current probe

Unfortunately, I don't have a current probe capable of measuring this signal.

Is there any other means of measuring the crystal drive level reliably using voltage measurements (even with a low-capacitance differential voltage probe)? Coarse accuracy (say 25%) is fine - if I'm not that far below the max drive strength, I would concerns about the design in any case.

For example, would it be valid to measure the voltage across a small "R_Q" (e.g. 1R) on the below diagram? What pit-falls are there in using a method like this?

enter image description here

\$\endgroup\$
5
\$\begingroup\$

Next time, I should read those application notes a little more carefully :)

The STM32 Crystal Application Note AN2867 offers an alternate method in Section 3.5.2:

This current can be calculated by measuring the voltage swing at the amplifier input with a low-capacitance oscilloscope probe (no more than 1 pF). The amplifier input current is negligible with respect to the current through C_L1, so we can assume that the current through the crystal is equal to the current flowing through C_L1.

That is, measure the voltage at the non-inverting input to the amplifier (across capacitor C_L1), because the current through the load capacitor is essentially the same as the current through the crystal (since the amplifier is high-impedance).

enter image description here

The drive level can then be estimated as: $$ DL= \frac{ESR \times \left(\pi f C_{tot} \right)^2 \times \left(V_{pp}\right)^2}{2}$$ where $$ C_{tot} = C_{L1} + C_{s}/2 + C_{probe} $$ and ESR is from the crystal, Cs is the board stray capacitance, C_probe is the capacitance of the probe leads, and f is the frequency of operation.

I have not been able to verify this method against the current probe measurement, but it gives sensible values for an NXP S32K development kit (i.e. an 8MHz crystal with a measurement of 0.6Vpp gives a power estimate of 6uW).

\$\endgroup\$
  • \$\begingroup\$ So during the design phase should we just leave a footprint for a resistor for Rext - since we don't know what value it will need to be until we measure this? \$\endgroup\$ – VanGo May 28 at 0:50
1
\$\begingroup\$

This online crystal oscillator simulation may help you estimate the power dissipated in the series (motional) resistance of your crystal by using the (AC) output voltage of the MCU.

The model is from this stack exchange discussion.

The MCU output will be loaded by your scope probe, but if you know the impedance of your probe (and you should), simply add it to the model.

You would have to find estimates for the motional inductance and capacitance from your crystal supplier.

The crystal's series (motional) resistance is normally in the data sheet.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.